JP2000041070A - Atm communication system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an ATM communication system capable of providing communication between ATM networks without sacrificing high-speed property, large capacity property and connection oriented property. SOLUTION: This system is composed of plural ATM-LAN 11 and 12 operated in an asynchronous transfer mode, while respectively accommodating plural terminals, and an interworking unit 13 provided between the ATM-LAN 11 and 12 so as to perform internet working between the ATM-LAN 11 and 12. Concerning the packet of at least one part of packets sent from one ATM-LAN connected to the interworking unit to the other ATM-LAN, the interworking unit 13 does not perform processing higher than an ATM adaptation layer but performs at least one of the reference of an ATM cell header and the switching of an ATM cell in ATM layer processing.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to an ATM communication system.

[0002]

2. Description of the Related Art In recent years, demands for various communication such as image communication and high-speed data communication have been increased, and integration of a communication network (B-ISD) has been proposed in order to provide efficient and flexible communication services.
N) is desired. ATM (Asyn
chronous Transfer Mode: Exchange is promising. The ATM switching system is intended to realize a communication service by storing information in a fixed-length packet called a cell irrespective of its attribute and using the cell as a unit of exchange.

The CCITT has defined this ATM switching system as a formal next-generation switching system, and has decided to formally implement B-ISDN using the ATM switching system. Along with this, public networks and corporate networks are constructed based on the ATM switching system, and it is highly likely that this will fulfill the needs of next-generation multimedia communication and broadband communication.

There is a movement to apply this ATM switching system to the field of LAN (local area network). This is to implement a conventional LAN communication system represented by Ethernet by an ATM switching system (hereinafter also referred to as an ATM communication system), and standardization has already begun in the United States and other countries.

When the ATM switching system is applied to a LAN,
The following advantages can be considered.

(1) Realization of Broadband Communication The communication speed of Ethernet, which is a substantial standard of the current LAN, is 10 Mbps. FDDI (100Mbps)
Although higher-speed LANs are also appearing, these are both shared media (a system in which the bandwidth is shared by all terminals). On the other hand, 150Mbps / 620M
The introduction of the ATM switching system using bps as a standard is basically a star type configuration, so that the terminal can exclusively use the band and there is a possibility that the throughput of the LAN will be dramatically improved. .

(2) Suitable for multimedia The current LAN has supported so-called data communication such as file transfer and transaction processing. However, due to the system, support for real-time communication typified by voice and image is supported. Or the overhead is very heavy and the bandwidth supported is very small.

[0008] On the other hand, the ATM switching system uses real-time / continuous communication such as voice and image, and non-continuous communication such as data communication, because of unified processing by hardware and drastic improvement of network reliability. This is a communication method that enables communication fusion, that is, multimedia communication. For this reason, there is a possibility that introduction of the ATM switching system into the LAN may promote the realization of a multimedia environment in the LAN.

(3) Good compatibility with public networks.

As described above, since the target of realizing broadband communication in the public network is determined by the CCITT to be the ATM switching system, the future public network (public communication network) will be constructed as an ATM network. Very likely to be.

[0011] Apart from this, there has recently been an increasing demand to use the LAN environment in a wider area. This means that the same environment as a LAN is made wider, that is, a MAN (Metropolitan Area Network) or a WAN (Wide
Area network). In order to realize this, it is necessary to interconnect terminals or LANs via a public network. Generally, when these terminals or networks are physically separated, they are interconnected via a public network. There is a need.

An ATM switching system is used for a public network, and LA is used.
Assuming that a similar communication method is used in N, it is considered that information / packets can be mutually entered by only a simple protocol conversion at the boundary point,
It shows high affinity for public networks. From these facts, it is considered that the application of the ATM switching system to the LAN can secure a broadband communication line between a long-distance terminal and a network and a real-time property, which were impossible in a conventional public network. , MAN or WAN.

On the other hand, in a conventional LAN environment, for example, Ethernet or the like, a router is arranged between each LAN when connecting LANs, ie, performing internetworking between LANs. This router is layer 3 (network layer) of the OSI protocol layer stack
The main function of this function is to perform routing processing for datagram communication across LANs. That is,
A datagram that spans two LANs is always sent up to Layer 3 by a router, where the destination network layer address is analyzed and delivered to the destination LAN according to the analysis result. This router is often called the “gateway” in the computer communication world, but the term “gateway” is defined as an entity that performs processing up to layer 7 in OSI. Since they are different, they will be referred to as routers hereinafter.

A "bridge" is also known as one similar to a router for realizing the connection between LANs. This is because the router analyzes the destination network layer address and determines the LAN to send, whereas the bridge analyzes the data link layer address (MAC address) and sends the LAN.
Is determined. Specifically, the bridge analyzes the destination MAC address of the received datagram and
When the AC address is not addressed to the own LAN, the datagram is transmitted to the other LAN to realize a connection between LANs. Further, as a similar example, "Bruta" which functions as a router for a predetermined network layer protocol and functions as a bridge for other protocols is known.

[0015] A workstation (WS) has usually been used for these routers, bridges and brouters. That is, the functions of the router, bridge, and brouter are realized by the CPU in the WS analyzing the address and transmitting the address to the assigned physical port.

[0016]

The ATM communication system is A
One of the features is that high speed is achieved by hardware switching of TM cells. That is, A
The TM network is “connection oriented” (hereinafter referred to as CO
A virtual connection (Virtual Connection: VC) or a virtual path (Virtual Path: VP) is established between end-to-end, and these VCs or VPs are label-multiplexed by their identifiers (VCI or VPI). Alternatively, packets called cells are delivered end-to-end in a label-exchanged form.

The information (data) delivered between end-to-end is stored in the payload of the ATM cell, and the ATM cell is exchanged to the destination terminal by hardware switching only along with VC / VP without software intervention.・ Transferred. Hardware switching is based on the VPI / VCI (in some cases, ATM
This is performed by the ATM switch with reference to the values of other areas of the cell header, such as PT, for example.

When this ATM communication system is applied to the field of LAN, communication between terminals in the LAN is performed by the A
It is considered that this can be achieved by communication through TM-VC / ATM-VP, and a dramatic increase in the speed and capacity of communication between terminals can be expected.

However, when communication is performed between LANs (hereinafter referred to as ATM-LAN) to which such an ATM communication system is applied, as described above, the router, bridge or brouter located between the LANs forcibly layer. Termination at layer 3 or layer 2 is performed. It is highly likely that the layer 2 and layer 3 processing after the termination is normally performed by software processing. For this reason, communication over LANs is significantly faster than communication within LANs.
It is highly possible that large capacity is lost. In a conventional system in which a router, a bridge, a brouter, or the like is provided between LANs to perform inter-LAN communication, VP / VC across LANs is basically not established. VP
This is because / VC does not perform the layer processing of the ATM layer or more between the endpoints. This is AT
For communication over M networks, it means that a connection-oriented communication line, which is one of the features of the ATM communication system, cannot be set.

In contrast to the ATM communication system which is a connection-oriented communication system, the communication system used in the conventional data communication is "connectionless" (hereinafter referred to as CL).
). In the connectionless communication system, a connection is not necessarily established between end-to-end, a packet is sent to a network in a form of attaching destination information to a part of the packet, and some node in the network analyzes the destination information and performs routing processing. Then, the packet is transferred to the destination terminal. That is, in connectionless communication, a terminal implements a datagram without performing a connection setup procedure. A packet transmitted to a destination terminal without a connection in this way is called a datagram, and a communication method using this is called a datagram. This is called the transfer method. In other words, connectionless communication is a method for realizing communication in the form of datagram transfer without a terminal performing a connection setup procedure.

Most existing data terminals, such as workstations (WS) and personal computers (PC), apply this datagram transfer method in most cases. This is because most conventional LANs support the datagram transfer method, and software (eg, OS) installed in the data terminal is for datagram transfer. Typical examples thereof include TCP / IP and UDP / IP.

These existing terminals or terminals equipped with an existing protocol, that is, datagrams are generated,
For a terminal sending to the partner terminal / network via the ATM network,
The datagram transfer method is used for communication between terminals.
Therefore, in order to adapt these terminals to the ATM network, it is necessary to (a) replace the current LAN substrate in the terminal, for example, an Ethernet board with an ATM network substrate (ATM board).
Or a function to adapt to the interface with the ATM-LAN by using a terminal adapter (TA), etc .; (b) a function to put a datagram into an ATM cell in a terminal at some form; An improvement such as providing a function of delivering a gram to the destination terminal indicated by the destination address is required on the terminal side and the network.
The term "terminal" here includes a gateway between an existing LAN and an ATM network.

As a function for realizing this, a conventional CLSF
A datagram delivery method using (connectionless service function) has been known. The datagram delivery method using the CLSF is realized as follows.

A CLSF processing unit is arranged in the ATM network, and all datagrams are collected here. That is, all the datagram terminals and the CLSF processing unit are connected by PVC (permanent VC) (or VC, VP, or PVP), and the terminal converts all the datagrams to be transmitted into ATM cells and converts them into ATM cells.
It is put on the PVC heading for the LSF processing unit and sent to the CLSF processing unit. The CLSF processing unit reproduces the received datagram, analyzes the destination address, selects a PVC connected to the destination address, converts the datagram into an ATM cell, and sends it out again. When there is no PVC connected to the destination address, when a plurality of CLSF processing units exist in the network, it is considered that the terminal includes the terminal which is the destination address, or the data is transmitted to the next-stage CLSF processing unit predetermined by the routing rule. The gram is converted into an ATM cell again and transmitted.

In the CLSF processing unit, it is not necessary to analyze the destination address after reproducing the datagram, convert the datagram into an ATM cell again, and transmit the ATM cell. The first cell in which the datagram is converted into the ATM cell contains the destination address. If it is, the destination address in the first cell is analyzed, and the cell is directly forwarded to the destination terminal.
A method of sequentially transmitting the second and subsequent cells obtained by converting the datagram into ATM cells to the destination terminal may be used.

However, in the above-described method using CLSF, all datagrams transmitted from within the network always pass through the CLSF processing unit, and the more datagrams to be transmitted and the more terminals in the network, As the number increases, a higher throughput is required for the CLSF processing unit, and a very high throughput and flexible expandability are required for the CLSF processing unit.

Another method of transmitting a datagram to a destination terminal is to set up an ATM connection, for example, a VC to the destination terminal, and convert the datagram converted into an ATM cell into the V-gram.
This is a method of placing the product on C and delivering it. However, in this method, it is very problematic to which destination terminal the VC is set. In other words, there are practically innumerable terminals capable of sending datagrams, and since datagrams are bursty unlike voice information,
Unnecessarily establishing a connection wastes network resources.

Further, when realizing connectionless communication in a conventional ATM network, an ATM connection is always terminated in the CLSF processing unit, and protocol processing higher than the AAL layer, for example, protocol processing for a connectionless service called CLNAP is performed. Done. In other words, even when datagram communication is performed between very close terminals, the ATM connection is once terminated in the CLSF processing unit. Further, when datagram communication is performed between distant terminals, the datagram passes through a plurality of CLSF processing units, and each CLSF processing unit performs protocol processing on the AAL layer or higher.

Generally, a CLNA higher than the AAL layer
Protocol processing such as P is performed by software processing (AAL layer and lower layers are generally performed by hardware processing), and the processing speed is slow. In addition to communication between terminals belonging to a network supported by the CLSF processing unit, the CLSF processing unit communicates with terminals in a network supported by another CLSF processing unit. For example, it is necessary to analyze network layer address information), and the load of the datagram transfer process is concentrated on the CLSF processing unit. For these reasons, the conventional ATM communication system has a problem that it is difficult to realize high-speed communication in connectionless communication (datagram delivery) between terminals.

An object of the present invention is to provide an ATM system without sacrificing high speed, large capacity, and connection orientation.
An object of the present invention is to provide an ATM communication system capable of realizing communication between networks.

Another object of the present invention is to provide an ATM communication system capable of efficiently performing datagram distribution using an ATM network.

Still another object of the present invention is to provide an ATM communication system capable of performing connectionless communication between terminals connected to an ATM network, that is, high-speed datagram delivery.

[0033]

In order to solve the above problems, an ATM communication system according to the present invention is configured as follows.

(1) A plurality of ATM networks operated in the asynchronous transfer mode, a plurality of terminals respectively accommodated in the plurality of ATM networks, and a plurality of ATM networks provided between the plurality of ATM networks. An internetworking device that performs internetworking of at least a part of packets sent from one ATM network connected to the internetworking device to another ATM network. No processing is performed on the packet of the ATM adaptation layer or higher, and at least the ATM
An ATM communication system characterized by performing at least one of reference to a cell header and switching of an ATM cell.

This network connection device preferably has ATM cell header conversion means.

The network connecting device is preferably C
It has an LSF (connectionless service function) processing means. The CLSF processing unit performs network layer processing (analysis of network layer address, routing processing, and the like) on at least a datagram spanning the internetwork connecting device.

Further, an address resolution protocol in datagram distribution developed in the ATM network (when a network layer address of a datagram destination is known and a physical address to which the datagram is to be transmitted is unknown, A protocol for obtaining a physical address from a network layer address).
Regarding the resolution of addresses located outside the M network, the VPI connected to the CLSF processing unit in the network connection device
/ VCI may be notified.

(2) A plurality of ATM networks operated in the asynchronous transfer mode, a plurality of terminals respectively accommodated in the plurality of ATM networks, and a plurality of ATM networks provided between the plurality of ATM networks. And an internetworking apparatus for performing internetworking, wherein the internetworking apparatus has connectionless service processing means.

The connectionless service processing means performs network layer processing on datagrams delivered at least over the inter-network connecting device having the connectionless service processing means, and performs three or more ATs.
The delivery of datagrams across the M network may be performed by relaying between the connectionless service processing means.

When there are a plurality of these network connection devices, the CLSF processing means in the network connection devices may be connected by a permanent connection.

The connection connecting the CLSF processing units in the network connection device may be a connection that is not managed by the bandwidth management entity.

Further, the inter-network connection device may have a call / connection processing means (a part having a function of setting, disconnecting, changing, and managing the call and / or the connection) internally.

Further, the call / connection processing means in the network connection device may perform at least a process for a call / connection across the network connection device.

Further, regarding an address resolution protocol for detecting a call process associated with a call / connection setup / change / disconnection process developed in the ATM network, the call / connection targeted for the process is executed on the ATM network. The address resolution in the case of an address located outside of the network may be performed using the VPI / VCI addressed to the call processing unit in the network connection device.

The call / connection processing means in the ATM network relays the call / connection processing to the call processing in the inter-network connecting device when the target call / connection is not closed within the ATM network. It may be.

(3) A plurality of ATM networks operated in the asynchronous transfer mode, a plurality of terminals accommodated in each of the plurality of ATM networks, and a plurality of ATM networks provided between the plurality of ATM networks. And an interworking apparatus for performing internetworking, wherein the interworking apparatus has call processing means, and the call processing means processes at least a call / connection spanning the interworking apparatus having the call processing means. Is applied.

In this call processing means, processing of a call / connection over three or more ATM networks may be performed by relaying between the call processing means.

The call processing means in the network connection device may be connected by a permanent connection.

(4) A plurality of ATM-LANs operated in the asynchronous transfer mode, a plurality of terminals respectively accommodated in the plurality of ATM-LANs, and the plurality of ATM-Ls
At least one datagram server provided in the entire AN; connection setting means for setting an end-to-end ATM connection between the plurality of terminals accommodated in the same ATM-LAN; L
The datagram delivery between the plurality of terminals accommodated in the AN is performed through the ATM connection set by the connection setting means,
An ATM communication system comprising: a datagram delivery unit that performs datagram delivery between the plurality of terminals in the LAN through the datagram server.

Here, the datagram server performs the following processing. The datagram (connectionless packet) converted into an ATM cell is terminated once. However, it is not necessary to reassemble the datagram. That is, it is not always necessary to perform network layer termination, and CC
Termination may be performed at the CL layer being discussed at the ITT. Then, after referring to the network layer address once, an appropriate ATM connection (VP / V connected to the node having the network layer address)
C, or V connected to CLSF processing means considered to have a function of delivering the network layer address.
P / VC).

The terminal may set up an end-to-end ATM connection with another terminal, or perform address resolution when the terminal is started up or booted.

The terminal may set up an end-to-end ATM connection with another terminal or perform an address resolution when the terminal logs in.

In the above datagram delivery method, if an ATM connection cannot be established between terminals in the same ATM-LAN, or if address resolution cannot be performed, the terminal transmits the data to be transmitted. The gram may be transferred to a datagram server or may be broadcast in the ATM-LAN.

(5) A plurality of ATM networks operated in the asynchronous transfer mode, a plurality of terminals respectively accommodated in the plurality of ATM networks, and a plurality of data provided in at least one of the plurality of ATM networks. A gram server, and an inter-network connection device provided between the plurality of ATM networks and performing internetworking between the ATM networks.
An ATM connected to the network connecting device;
A for requesting address resolution from the network
If the target node of the address resolution does not exist in the ATM network connected to the network connection device in response to the RP request, a reply to the ATM address of the ATM connection connected to the network connection device is sent to the ARP for the ARP request.
Response means for performing an RP response, and ARP from the response means
The first indicated by the ATM address returned by the response
An ATM communication system, comprising: an ATM connection for connecting the ATM connection and a second ATM connection set between the network connection device and the datagram server.

(6) A plurality of ATM networks operated in the asynchronous transfer mode, a plurality of terminals respectively accommodated in the plurality of ATM networks, and a plurality of data provided in at least one of the plurality of ATM networks. A gram server, and an inter-network connection device provided between the plurality of ATM networks and performing internetworking between the ATM networks.
The network connection device includes a routing protocol terminating unit, and an address resolution device for performing address resolution when receiving an ARP request for requesting address resolution from a first ATM network connected to the network connection device. AR
Notification means for notifying the P server of information relating to routing; and a second address indicated by an ATM address returned by an address resolution response from the address resolution server to the information passed by the notification means. 1 ATM connection and a second AT set between the network connection device and the datagram server.
An ATM communication system comprising: connection means for connecting to an M connection.

(7) Here, the connection connection means in (5) or (6), for example, connects the first ATM connection and the second ATM connection with each other by an ATM cell header converting means and an ATM cell switching means. It is configured to be performed by at least one.

In this case, the number of datagram servers can be increased or decreased or changed by setting at least one of the ATM cell header conversion means and the ATM cell switching means.

(8) AT operated in asynchronous transfer mode
An M network, and a plurality of terminals accommodated in the ATM network, wherein the terminal serves as an ATM communication board for requesting an address resolution addressed to the terminal from an ATM cell from the ATM network. ARP request cell extracting means for extracting an ARP request cell, and an AR for responding to the address resolution request based on the ARP request cell extracted by the ARP request cell.
ARP response cell generation means for generating a P response cell, and multiplexing means for multiplexing the ARP response cell generated by the ARP response cell generation means with an ATM cell stream sent from the terminal device. ATM communication system.

The operation of the ATM communication system according to the present invention will be described below.

(1) In a conventional ATM communication system, a packet input by an inter-network connection device is processed by being raised to an upper layer (a data link layer, a network layer, or a higher layer). On the other hand, according to the present invention, a part of a packet transmitted from one ATM network connected to the network connecting device to another ATM network connected to the network connecting device is not subjected to AAL or higher processing.
At least the reference to the ATM cell header and / or the switching of the ATM cell in the ATM layer processing allows the transmission. Thereby, the AT spanning the network connection device
M connections (VP, VC) can be generated without terminating the connection in the network connection device. Since the ATM connection passes through the inter-network connecting device only by the ATM layer processing without being terminated in the upper layer, it is possible to increase the speed of data and transfer a large amount of data between the respective networks.

In an ATM network, the range covered by the physical address system can be determined as one closed network (ATM-LAN). That is, it can be determined that the range of the entity that determines the value of VPI / VCI, which is the physical address system in the ATM network, is one ATM-LAN.
The entity that determines the value of this VPI / VCI is ATM-LA
There are one or more in N, which operate in coordination. In this way, the VPI / VCI system has an independent ATM
In the network connection device for connecting LANs, these address systems are converted, that is, VPI / VC on one side is converted.
A mechanism for performing address conversion from the I address system to the VPI / VCI address system on the other side is required. By having the ATM cell header conversion means in the network connection device, it is possible to refer to the ATM cell header value including the VPI / VCI value, and it is possible to change the address system by the conversion means. .

(2) In the present invention, CLS is installed in the network connection device.
By providing the F processing means, a datagram distributed across the network connection devices can be processed (terminated) here. When a datagram is delivered using the CLSF processing means, the ATM connection is once terminated by the CLSF processing means, and the datagram converted into an ATM cell is once raised to an upper layer (network layer or CL layer) processing. That is, the network layer address or the CL layer address is referred to. In this case, reassembly of the datagram is not necessary. After this, the AT connected to the referenced address
The M connection is selected and the datagram is sent out through the ATM connection.

By arranging the CLSF processing means in the inter-network connecting device in this manner, datagrams distributed over the inter-network are terminated at the ATM connection here. The CLSF processing means can be directly accessed from the ATM network connected to the interconnection device. This access can be made without using an ATM connection that spans a plurality of networks. (2) The AT connected to the network connection device
With regard to datagram delivery over M networks, by performing the processing here, the conversion of the address system (VPI / VCI value) in each ATM network can be intensively performed here. Routing information (network layer address, CL
It is not necessary to exchange the layer address and the VPI / VCI value between networks).

Further, the AT connected to the above network connection device
In an M-network (ATM-LAN), a terminal desiring to deliver a datagram performs address resolution for detecting where to send the datagram, and the transmission address of the datagram is out of the ATM-LAN. If it is addressed to the located address, CLSF in the network connection device
By notifying the VPI / VCI connected to the processing means by, for example, the CLSF processing means itself or an ARP server to be described later, datagram delivery outside the network can be uniformly performed.

The above configuration can be applied not only to an inter-network connection device between ATM networks but also to an inter-network connection device between an ATM network and a network operated by another system (for example, Ethernet system).

When datagram delivery spans three or more networks, the CLS located in the network connection device
By performing this by relaying between the F means, a datagram delivered outside the own ATM-LAN is transmitted to a CLSF processing means disposed in the network connection device, and the CLSF processing is performed. The datagram destination to be managed by the means can be maintained only in the terminal / node in the network connected to the interworking apparatus except for the other CLSF processing means.
Operation can be planned.

Here, the connection between the CLSF processing means is made by a permanent connection, specifically, a PVP (permanent VP) or a PVC (permanent VC), so that a connection is generated each time a datagram is transferred between the CLSF processing means. Since the setting overhead can be reduced and the delivery of datagrams across three or more networks can be performed via the permanent connection, the traffic used for datagram delivery can be made. Can be easily monitored.

(3) In addition, datagram delivery is usually performed over an unreliable connection. That is, it can be considered that the transmission is performed on the assumption that the transmitted datagram may be lost in the middle, or that the delay time until the transmitted datagram reaches the transmission destination cannot be predicted. For this reason, the connection that connects the CLSF processing means is also excluded from the management of the bandwidth management entity, that is, the connection is not subject to the connection control for the bandwidth management when setting the connection, and the connection having the lower cell delivery priority is set. It can be. As a result, a strategy is taken to preferentially pass the cell of the connection to be subjected to connection control (a connection with a high priority), but the datagram delivery is appropriate in nature.

Further, by providing call / connection processing means in the network connection device, the processing (setting, disconnection, change, management, etc.) of the ATM connection across the network connection devices is processed (terminated) here. be able to. Thus, (1) the processing of the ATM connection across the inter-network connecting device requires information in both connected networks. By arranging the call / connection processing means in the inter-network connecting device located between the respective networks and capable of knowing information in the respective networks, the efficiency of the processing of the ATM connection across the inter-network connecting devices is improved. Can be done (2) From each network connected to the network connection device,
The call / connection processing means can be accessed. This access can be made without using an ATM connection that spans a plurality of networks. (3) The call / connection processing means in the network connection device can specialize in processing of an ATM connection across the network connection device. (4) For an ATM connection spanning between networks, it is necessary for the network connection device to convert each ATM address system (VPI / VCI system). By providing call / connection processing means in the network connection device, the VPI / V
It is possible to enjoy such an advantage that elements indispensable for the conversion of the address system such as the setting means of the conversion table of the CI system can be included in the call / connection processing means or can be tightly coupled to the processing means.

Further, A connected to the above network connection device
In a TM network (ATM-LAN), when a terminal that wants a call / connection processing request (setting, disconnection, change request, etc.) performs address resolution for detecting where to send the processing request, If the connection destination address of the call / connection is an address located outside the ATM-LAN, the VPI / VCI connected to the call / connection processing means in the inter-network connecting device is notified, so that the outside of the network is notified. Call / connection processing can be performed in a unified manner.

When call / connection processing means is provided in the ATM-LAN, if the processing for the requested call / connection is closed within the ATM-LAN, The processing is performed in the call / connection processing means, and when the processing is a processing request with an entity outside the ATM-LAN, the call in the inter-network connection device that manages the processing of the call / connection outside the ATM-LAN. /
By relaying the processing to the connection processing means, processing relating to the call / connection with the outside of the network can be uniformly performed.

The above configuration can be applied not only to an inter-network connection device between ATM networks but also to an inter-network connection device between an ATM network and a network operated by another system (for example, Ethernet system).

When the above-mentioned call / connection processing spans three or more networks, the call / connection processing is performed by relaying between call / connection processing means arranged in the inter-network connecting apparatus, so that the own ATM-connection processing is performed. The method of sending an ATM connection outside the LAN to the call / connection processing means arranged in the network connection device and the entity (terminal, node, etc.) managed by the call / connection processing means are other Except for the call / connection processing means, it is possible to maintain a method of using only terminals / nodes in the network connected to the inter-network connection device, and to efficiently operate the call / connection processing.

The connection between the call / connection processing means is defined as a permanent connection (PVP or PV
C), it is possible to reduce the connection setup overhead that occurs each time the processing information is transferred between the call / connection processing means, and to perform the call / connection processing over three or more networks by using the permanent It is possible to always perform the relaying between the call / connection processing means in the network connection device via the connection. Also, it is possible to easily monitor traffic used for call / connection processing.

(4) As described above, unlike the wide area network environment, the application program is considered to require a sufficiently small latency and a short call setup time for the LAN in the LAN environment. Since the datagram does not require call setup, the latter requirement can be satisfied. However, when a datagram server is used in an ATM environment, the residence time of the datagram in the datagram server generally becomes a problem. , Small latency cannot be secured. An environment that particularly demands low latency is
Since it is considered as an application in the LAN (a physical distance exists in a WAN environment, a propagation delay is unavoidable), terminals in the ATM-LAN are connected to the same AT.
By providing an ATM connection end-to-end with another terminal in the M-LAN and performing datagram delivery, both of the above requirements can be satisfied. In addition, since the traffic of datagrams in a LAN is generally considered to greatly exceed the traffic of datagrams across LANs, the exchange of completed datagrams in the LAN is performed via a datagram server. Doing so directly can also prevent the datagram server from requiring excessive throughput. In this case, for a datagram extending over a plurality of networks, the datagram may be delivered through a datagram server having datagram delivery means.

When a datagram to be transmitted is generated, the terminal device transmitting the datagram waits for the datagram on an ATM communication board or an arbitrary memory to set an ATM connection or address resolution. To the physical address (VPI in case of ATM network)
It is not always necessary to take out a datagram from the ATM communication board or an arbitrary memory, convert the datagram into an ATM cell, attach the physical address, and transmit the datagram after the / VCI value or other ATM cell header value is determined. . This is because the time during which the ARP is being performed is a state in which the datagram is kept in a waiting state, and is a useless time from the viewpoint of transmitting the datagram. It is considered that efficient datagram delivery can be performed by setting an ATM connection or performing address resolution in advance for a destination terminal that may transmit a datagram. The setting or address resolution of the end-to-end ATM connection may be performed at the time of starting (booting) or logging in the terminal device.

If it is not possible to establish an ATM connection between terminals in the same ATM-LAN,
Alternatively, when address resolution cannot be performed, the terminal can transmit datagrams to be transmitted to a datagram server, thereby enabling datagram communication. This is because the datagram server is generally considered to have the ability to deliver the datagram to the destination terminal of the datagram (because the datagram server has the ATM-LAN
), Since it recognizes the existence of all network layer addresses of the node / IWU / terminal connected to the same ATM-
Datagram delivery into a LAN is possible because the datagram server can be used even if there is no end-to-end ATM connection or the datagram sending terminal cannot recognize the ATM connection.

When an ATM connection cannot be established between terminals in the same ATM-LAN,
Alternatively, when the address resolution cannot be performed, the terminal transmits the datagram to be transmitted to the ATM-L.
If broadcast in the AN, the information to be broadcast is
Since all terminals in the M-LAN are receiving and analyzing, information can be transmitted.

(5) Next, the network connection device according to the present invention will be described. Generally, in a network connection device,
Information on a network connected to the device (information on what network layer address or other node having an address is arranged, how to perform routing, and the like; routing information) can be directly collected. Because of its position, the main operating point of the routing protocol (in terms of the exchange of routing information in the backbone network, and in terms of the concentration of routing information in the branch network)
It has become. Conventional existing network connection devices (so-called routers in LANs) analyze the meaning of “the end point of the routing protocol” and the network layer address of the datagram input to the network connection device. However, it has two functions in the sense that "routing of the input datagram" is performed based on routing information obtained through the above routing protocol or the like. On the other hand, in the ATM network, the latter function, that is, a function of "analyzing a network layer address of an input datagram and performing routing of the datagram based on routing information obtained through the above routing protocol". Is the function of the datagram server,
Generally requires high cost and complicated equipment. Providing a large number of datagram servers as described above despite the fact that the total amount of datagrams actually communicated is not so large is considered to waste communication resources.

(6) Therefore, the routing protocol terminating means and the datagram service processing means are separated as in the present invention, and the AT of the address resolution result is further separated.
When the M address is required (when the address resolution destination node is not connected by the end-to-end ATM connection, that is, when it does not exist in the same ATM-LAN), the ATM address to be led to the datagram server By doing so, it is possible to provide a datagram service at an appropriate cost and scale without using more than the required number of datagram servers.

(7) In the above-mentioned interworking apparatus, the ATM address indicated by the ARP response indicates
An ATM connection connecting to the network connection device and an ATM connecting the datagram server from the network connection device.
The connection with the connection is made by setting the ATM cell header conversion means and / or the ATM cell switching means in the network connection device, so that the datagram transmitting terminal and the datagram server are directly connected by the ATM connection, that is, the ATM connection. They can be combined only by layer processing.

In an ATM network for transferring datagrams between an arbitrary terminal and a datagram server in the above-described manner, when the number of datagram servers is increased or decreased, or when the number of datagram servers is increased or decreased, the ATM in the network connection device is used. By doing this by changing the settings in the cell header conversion means and / or the ATM cell switching means, setting changes such as when adding or removing datagram servers or when changing datagrams are only simple table changes in the network connection device. In addition, it can be easily performed in a very short time.

(8) By using a terminal (ATM communication board) as in the present invention, generally, a broadcast channel (from any transmitting terminal to all nodes / IWUs / terminals belonging to the ATM-LAN, An ARP request cell (address resolution request cell) which is considered to be performed via an ATM connection capable of broadcasting the transmitted cell is quickly and rapidly handled in a higher-order (inside a terminal device) processing device (CPU). Etc.) without consuming processing time.

As described above, the entire network of the ATM communication system is constituted by a plurality of ATM networks.
In other words, the subnetting clarifies the distinction between the address analysis protocol and the routing protocol, and can flexibly cope with network expansion.

[0085]

Embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) FIG. 1 shows an ATM network according to the first embodiment. As shown in FIG. 1, the ATM network according to the present embodiment includes a first ATM-LAN 11, a second ATM-LAN 12, an inter-network connection device (hereinafter, IWU).
13).

First ATM-LAN 11, Second ATM
-The LAN 12 is a local area network operated by the ATM system. Each ATM-LA
Within N, the addressing scheme is independent. That is, the VPI / VCI value used inside each ATM-LAN has the authority to be determined by the VPI / VCI value determination function existing inside the ATM-LAN, and this authority is assigned to each ATM-LA.
N is independent. The terminal device and the node in the ATM-LAN, when the information to be transmitted exists (whether the destination is in the ATM-LAN or not), stores the information in the ATM cell, Appropriate ATM
A cell header is added and the packet is transmitted into the ATM-LAN.

FIG. 2 is a diagram showing the internal structure of the network connection device 13. The network connection device 13 includes an add / drop processing unit 21, a multiplexer / demultiplexer (hereinafter, referred to as a multiplexer / demultiplexer).
MUX / DEMUX) 22, CLSF processing unit 2
3, a call processing unit 24, an IWU management unit 25, and an ATM cell header conversion unit 26. The network connection device 13 is located between two ATM-LANs and has a function of managing internetworking (connection between LANs) of the two ATM-LANs.

The add / drop processing unit 21 refers to the header portion of the input ATM cell and, if the cell has an appropriate header value (that is, the cell is In the case where the cell is to be terminated, the processing of dropping the cell to the DEMUX 22 and the processing of inserting (adding) the cell from the MUX 22 into the cell transmission path are performed. Here, the add / drop processing unit 21 has a function of dropping cells from either the left or right on the cell transmission path, and a function of inserting cells into the cell transmission path in either the left or right direction.

In the add / drop processing unit 21, information on which cell is to be inserted on the cell transmission line in the left or right direction is specified by each of a CLSF processing unit 23, a call processing unit 24, and an IWU management unit 25, which will be described later. Shall be performed. That is, for example, the inserted cell is passed to the add / drop processing unit 21 together with a value of 0 if inserted into the cell transmission line in the left direction and 1 if it is inserted in the right direction. Decide the direction. The cell may be inserted at a location where an empty cell passes over a cell slot on a cell transmission line, replacing the empty cell.

When the cell is dropped from the add / drop processing unit 21 to the MUX / DEMUX 22 side,
A drop table as shown in FIG. 3 is referred to. This drop table is provided in each of the cell transmission lines in both the left and right directions, and the drop destination of the cell (CLSF processing unit 23, call processing unit 24, IWU management unit 25) is determined according to the value of this table. As shown in FIG. 3, the drop table has a cell header value as an entry, and a value indicating a drop destination is referred to. As a value indicating this drop destination,
For example, values are assigned such as 1 for the CLSF processing unit 23, 2 for the call processing unit 24, 3 for the IWU management unit 25, and so on. The cell to be dropped by the add / drop processing unit 21 includes a value indicating this drop destination,
The value is sent to the MUX / DEMUX 22 side together with a value indicating whether the cell is a drop cell from the left or right cell transmission path (for example, 0 if it is from the left, 1 if it is on the right).
The information on this drop table is stored in the IW
The U management unit 25 performs initialization, addition, change, and the like.

The MUX / DEMUX 22 refers to a cell (drop cell) sent from the add / drop processing unit and refers to information sent in parallel to any module connected to the lower part (in this embodiment, CLSF processing). Part 2
3. A function (DEMUX function) to be transmitted to the call processing unit 24 and the IWU management unit 25) and a cell (add cell, inserted cell) to be inserted into the cell transmission path transmitted from the module group connected to the lower part are multiplexed. , And a function (MUX function) for passing the information to the add / drop processing unit 21 together with information on which direction to insert.

The CLSF processing section 23 performs processing of a connectionless service function. As will be described in detail later, the CLSF processing unit 23 terminates the datagram (connectionless packet) distributed across the network connection device 13 once (note that it is not necessary to reassemble the datagram). That is, it is not always necessary to terminate the network layer, and it is also possible to terminate at the CL layer discussed in CCITT).
It has a function of sending to a connection (a VP / VC connected to a terminal / node having the network layer address or a VP / VC connected to a CLSF processing unit considered to have a function of delivering the network layer address). It should be noted that the ATM connection for datagram delivery is once terminated in the CLSF processing unit 23.

The call processing unit 24 basically includes the inter-network connection device 1
Setting, disconnection, change,
It has a function to perform management and the like. Further, it may have a function of managing the band on the ATM connection or the ATM transmission line in the network connection device. Details will be described later.

The IWU management unit 25 has a function of managing and controlling the network connection device 13.

The header conversion unit 26 (from both left and right directions)
It has a function of referring to the header value of the input cell and, if the header value is an appropriate value, rewriting this to another header value. The header conversion unit 26 basically has a function of referring to a cell header value (input cell header value) and rewriting the cell header value to a certain cell header value (output cell header value). Has a correspondence table inside. Initialization, addition, and change of the correspondence table are performed by the IWU management unit 2
5 is performed. This correspondence table matches the drop table in FIG. 3 in that the entry value is the input cell header value, so that the table can be easily integrated with the drop table.

It is assumed that each module in the network connection device 13 is connected to a bus (not shown), and control such as a change of a set value of each module from the IWU management unit 25 is performed through this bus. Shall be

Instead of exchanging information between the modules via the bus as described above, the information may be put on ATM cells and the cells may be exchanged between the modules.

In this embodiment, various processes (CLSF processing, call processing, etc.) on the cell transmission path in both the left and right directions are performed by the same module on the left and right sides. It is possible.

In the present embodiment, various processes in the IWU are performed separately (that is, the CLSF processing unit is performed by the CLSF processing module, and the call processing is performed by the call processing module. All, or some of these processes are performed by the same CPU / MPU
It is also possible to perform at.

Also, the case where there is only one header conversion unit 26 in the network connection apparatus of FIG. 2 of the present embodiment is described. And / or after which a header conversion unit is provided,
Each of them performs a header conversion of an ATM cell of a unidirectional cell flow (FIG. 38), or an add / drop processing unit is provided before and after a header conversion unit as shown in FIG. A method of adding and dropping ATM cells is also conceivable. By doing so, the function of the network connection device 13 can be symmetrically viewed from both the left and right directions.

Next, FIG. 4 shows the first ATM-LAN 11,
One embodiment of a node / terminal configuration inside the second ATM-LAN 12 is shown. Thus, the first ATM-LAN
Is 11 with three switch nodes 41, 42, 43,
It comprises terminals 4A, 4B, 4C and 4D connected to these. In addition, the second ATM-LAN 12 includes two switch nodes 44 and 45 and terminals 4E connected thereto,
4F, 4G.

An ATM switch is mounted in the switch node, and connection between switch nodes, connection with an inter-network connection device, and connection with a terminal device can be performed. In this case, a star type (or a tree type) is used. ). The interface speed between these switch nodes and the terminal device / switch node / inter-network connection device is 10M,
Various values such as 20M, 155M, and 622M can be selected.

Next, FIG. 5 shows an embodiment of an ATM connection connection state between terminals in both ATM-LANs and a CLSF processing unit in the network connection device 13. Here, wiring between the terminal and the switch node and between the switch node and the network connection device are omitted for simplicity. Other components such as a call processing unit and a header conversion unit in the network connection device are also omitted in the figure.

As described above, each terminal device and the network connection device 1
Each of the CLSF processing units 23 in 3 is connected by an ATM connection. This ATM connection is
It may be P or VC. This ATM connection is a connection that is not subject to band management.
That is, when physically establishing an ATM connection between the terminal device and the CLSF processing unit 23 in the inter-network connecting device 13, the band is not reserved (that is, the band management for the band management is performed in the call admission control). (Without going through an evaluation function). AT that does not perform such band management
It is considered that the M connection has a lower priority than the ATM connection that performs band management. That is, regarding parameters such as a cell delivery delay or a cell loss rate, when compared with an ATM connection that performs band management,
ATM connections that do not perform bandwidth management will show bad values. The ATM connection between the terminal device and the CLSF processing unit is used for datagram delivery, but the delivery of datagrams originally requires a large delay time and unreliable physical connection. Because it was considered in such an ATM
It is reasonable to use a connection.

As described above, in the present embodiment, the ATM
An ATM connection in the LAN (A
The ATM connection that extends over the TM-LAN includes an ATM connection that performs bandwidth management and an ATM connection that does not perform bandwidth management. ATM connection for bandwidth management has constant communication quality (QOS)
A connection that can perform communication while maintaining communication (a connection that can expect a cell loss rate equal to or higher than a certain value and a delay time equal to or less than a certain value). An ATM connection that does not perform bandwidth management is a connection (cell with no guarantee of communication quality) There are no restrictions on the discard rate and delay time).

An ATM connection that performs bandwidth management is considered to be a high-priority connection, and an ATM connection that does not perform bandwidth management is considered to be a low-priority connection. For example, only when there is no cell belonging to a high-priority connection, priority is given. This can be realized by performing priority control such as permitting passage of a cell belonging to a low-degree connection.

An ATM connection for performing band management is a connection whose setting is permitted only when a cell transmission path through which the connection passes and communication resources in the switching node are secured. The ATM connection which does not perform the call is out of the band management target.
As long as the number is less than or equal to the maximum number of accepted connections (of each connection management entity), it is unconditionally accepted.

The ATM connection between the terminal device and the CLSF processing unit 23 in the network connection device 13 is as follows.
It should be noted that it does not extend over the network connection device, that is, does not extend over the ATM-LAN. ATM-L
When setting up an ATM connection spanning between ANs, generally, negotiations between both ATM-LANs, address conversion, protocol conversion, etc. are generally required as described later, and a CLSF processing unit is provided in the network connection device. This eliminates the need to provide such an ATM connection across the ATM-LAN for datagram delivery.

The terminal device delivers datagrams via the inter-network connection device 13, that is, datagrams outside the ATM-LAN, via the CLSF processing unit 23 in the inter-network connection device 13. . That is, although the datagram converted into an ATM cell is sent to the CLSF processing unit 23, there are several approaches until the terminal device knows that the datagram should be sent to the CLSF processing unit 23. The method is considered.

The terminal apparatus determines the destination address of the datagram to be transmitted (a network layer address or a mail address including a domain name).
The datagram is sent to any ATM address (ie, which VP
(I / VCI value) must be resolved. As described above, the datagram / ATM cell added to the datagram / ATM cell transmitted from the destination address
The work of resolving the PI / VCI value here is based on ARP (Address Resolution Pr
otocol (address resolution protocol).

There are several methods for implementing this ARP as follows.

(Method 1): VP addressed to the CLSF processing unit in advance
When I / VCI is assigned.

In this case, the ATM of the CLSF processing unit
The address (VPI / VCI) is uniquely assigned to the LAN, and the datagram can be sent unconditionally with the VPI / VCI attached.

In the case of such a system, the terminal uses its own ATM-
Since it is not possible to distinguish between datagram delivery within the LAN and datagram delivery to another LAN via the network connection device, the CLSF processing unit 23 in the network connection device 13
However, it is necessary to have a datagram delivery function closed within the ATM-LAN. In addition, as shown in FIG.
A CLSF processing unit is separately provided in the TM-LAN. For datagram delivery across the network connection device, a datagram of a format in which the datagram is relayed to the CLSF processing unit in the network connection device is provided. The delivery method is included in this method (ATM-
VPI / VCI addressed to the CLSF processing unit in the LAN is defined).

(Method 2): Method Using ARP Server There is an ARP server in the ATM-LAN (not shown), and the terminal device sends the ARP server a datagram of the destination address. This is a method for asking whether to send out with VPI / VCI attached.

The ARP server has a table inside, which stores a destination address, a VPI /
And the VCI value. A who received the inquiry
The RP server determines the VPI / VC corresponding to the destination address.
The I value is analyzed with reference to an internal table, and the analysis result (VPI / VCI value) is returned to the terminal device of the inquiry source. At this time, if an ATM connection is not established between the two terminal devices / nodes, the fact may be notified, or the ARP server sets up an ATM connection between the two terminal / nodes and thereafter sets the ATM connection. The VPI / VCI value of the ATM connection may be returned. Thus A
RP is performed.

[0118] Here, the terminal device is the AT of the ARP server.
M address (VPI / address from the terminal device to the ARP server)
(VCI value) or not.

If the user knows, there is no problem if an inquiry is made using the VPI / VCI value.

If the user does not know the ARP server's own AT
Perform the resolution (ARP) of the M address or use the broadcast channel to
A method may be considered in which the ARP server broadcasts the message on the TM-LAN and the broadcasted query is recognized by the ARP server as a query for itself, and performs subsequent processing. When this broadcast channel is used, its own address (network layer address, mail address if necessary, ATM address VPI / VCI value, server name (AR
P server), the physical address assigned to the communication board, etc.), the destination address (the network layer address of the entity from which the broadcast cell is to be received,
If necessary, it is necessary to include at least information of a mail address, an ATM address, a server name, a function name, and the like, and a broadcast cell type (what kind of broadcast cell is meant) (FIG. 7). Also, the address of the protocol (address of the transmission source) and the address of the destination (reception side address)
(For example, an IP address or an E.164 address) may be included. Also, when the same source makes multiple requests to the same destination at the same time,
A random number, a discriminating number, or the like may be added to the cell in order to indicate which request the response is to.

When this broadcast cell is used for ARP, for example, the address (network layer address, mail address if necessary) of the terminal / node which is the target of address resolution is entered in the other information field of FIG. Etc.) at least. In this case, the receiving side address in the broadcast cell format may be undefined (for example, all 1s). The reason for making the receiving address indefinite is that the receiving address in the broadcast cell is the address of the terminal to be received in the broadcast cell. This is because, when a broadcast cell is used for ARP, it is not considered that the terminal device targeted for address resolution always responds to the ARP, that is, it is not considered that the terminal device receiving the ARP receives and processes the cell. .

When performing the broadcast, for example, VPI = all 1 is used. VCI may be all 1s, or the value of VCI may include broadcast cell type information.

As an example of performing ARP using broadcast, for example, an appropriate ARP server that has received a broadcast cell (ARP cell) refers to the broadcast cell type and determines that this is a datagram transmission request ARP. Knowing that the ARP request is to be processed via the CLSF processing unit in the network connection device with reference to the address resolution target address, ATM address (VP
(I / VCI value). FIG. 8 shows an example of the format of a broadcast cell when such ARP is performed using a broadcast cell.

A request for address resolution is sent to the ARP server without using broadcast (end-
(Through an end ATM connection)
Needless to say, the inquiry needs to include the target address of the address resolution.

When the ARP server recognizes that the transmission destination of the datagram is the destination of the inter-network connecting device (that is, outside of the own ATM-LAN), the ARP server outputs the resolution as a result of the resolution in the inter-network connecting device. The ATM address (VPI / VCI value) is returned to the CLSF processing unit.

In order to recognize that [the transmission destination of the datagram is the destination of the inter-network connecting device], the destination address of the datagram is determined by using, for example, a subnet mask.
A method of examining whether or not it has an M-LAN sub-address is conceivable.

When the ARP server replies the analysis result to the inquiry source, several methods can be considered.

First, there is a method of replying the inquiry result using a broadcast channel. In this case, the response to the ARP is entered in the broadcast cell type in FIG. 8, and the ATM address (VPI / VCI), which is the result of the address resolution, replaces the address resolution target address or continues. .

The details will be described later.
1267, the scheme (here, this is V
PRP routing method), that is, a method in which one VPI is allocated to each terminal / node, and routing in the ATM-LAN is performed using the VPI. ARP is performed as the "transmission request ARP".

That is, the ARP is not always returned with the "datagram transmission request ARP" (the VPI / VCI of the ATM connection directly connected end-to-end with the destination. For example, the ATM connection to the CLSF processing unit) Of the connection connection request AR).
It should be noted that there is "P" (VPI / VCI of the ATM connection directly connected end-to-end with the partner). .

(Method 3) Method via Call Processing Server A terminal device that wants ARP requests a call processing server (not shown) existing in the ATM-LAN to set up an ATM connection with the CLSF processing unit. It is a method to do.
When the call processing server recognizes that the delivery destination of the datagram is beyond the network connection device, the call processing server may establish an ATM connection with the CLSF processing unit in the network connection device.

(Method 4) Direct ARP between Terminals / Nodes One VPI is allocated to each terminal / node in the ATM-LAN, and routing in the ATM-LAN is performed using the VPI (VP routing). Method), A
The RP is connected to the datagram transmitting terminal and the CLSF processing unit 23 in the network connection device 13 without going through any server.
Can be done directly between. That is, the terminal device issues an ARP request using the broadcast channel.

The contents of the request cell are as shown in FIG. The source address contains the address of the terminal device itself (the VPI value or the VPI / VCI value may be used), the receiving address contains the address of the other party, or an indefinite value, and the ARP destination address contains the address of the other party. In this case, since there are two areas where the destination address is located and it is considered redundant, the format of the ARP cell (broadcast cell) for direct ARP can be simplified as shown in FIG. It is possible.

The network connection device 13 that has received this (if necessary, the CLSF processing unit 23 of the network connection device 13 may be replaced by a broadcast cell type in a drop table, and the ARP If it is determined that this is a datagram that is delivered across the own network connection device, the ATM address addressed to itself is set to the address. Therefore, the resolution result is returned to the inquiry source terminal device with the VPI of the inter-network connecting device 13 and the required VCI value. This is because the inquiring terminal sends the cell with the VPI / VCI value added thereto, so that the cell reaches the inter-network connecting device according to the VPI value and is dropped to the CLSF processing unit according to the VCI value. become. When returning the resolution result, a broadcast channel may be used, or this may be performed via a VP addressed to the inquiry source terminal device.

Note that ATM-LAN does not always need to apply VPI routing, and ARP is directly applied to the general case where the receiving terminal knows the ATM address of the ATM connection with the transmitting terminal. It is possible.

As described above, the ARP for the delivery of the datagram across the network connection device is performed, and the terminal device that sends out the datagram thereafter converts the datagram into an ATM cell, and then converts the datagram into an ATM cell. If the destination is a destination that straddles the inter-network connection device, the previously resolved ATM address (VPI / VCI) is used.
This is sent to the CLSF processing unit in the network connection device. For an address once resolved, it will be used hereinafter. The CLSF processing unit in the network connection apparatus once converts this into a network layer or a CL.
A that terminates at the layer, analyzes the destination address, and then connects to the destination (after conversion to an ATM cell if necessary)
The TMgram will be appropriately selected and the datagram will be delivered through it.

As described above, by arranging the CLSF processing unit in the network connection apparatus, (1) the network connected to the network connection apparatus
The processing unit can be directly accessed. Also,
This access can be made without using an ATM connection spanning a plurality of networks.

(2) ATM connected to the network connection device
By performing processing here for datagram delivery across networks, the address system (V
Here, the conversion of the PI / VCI value can be performed intensively.

(3) ATM connected to the network connection device
Routing information on the network (for example, information on the relationship between the network layer address, the CL layer address and the VPI / VCI value and existence information) can be terminated / intensively managed by the network connection device / CLSF processing unit. There is no need to change over.

The advantages described above can be enjoyed.

Note that the ARP method (method 1 to method 4) described in the present embodiment does not necessarily require the CLSF processing unit to be present in the IWU, and the CLSF processing unit may be present at an arbitrary position in the network. Conceivable.

As described in detail in (method 2), "datagram transmission request ARP" (VPI / VCI of the ATM connection directly connected to the destination end-to-end is not necessarily returned to the ARP) Not limited, for example, V of ATM connection to CLSF processing unit
The PI / VCI value may possibly be resolved) and the "connection connection request ARP" (the VPI / VCI of the ATM connection directly connected end-to-end with the partner is returned). Attention should be paid to this point.

Next, FIG. 9 shows an embodiment of an ATM connection connection state between terminals in both ATM-LANs and a call processing unit 24 in the inter-network connecting device 13. Here, for the sake of simplicity, wiring between terminals and switch nodes and between switch nodes and networks are omitted. Other components such as a CLSF processing unit and a header conversion unit in the network connection apparatus are also omitted in the figure. In addition, both ATM-LAN
, Call processing units 91 and 92 are added.

As described above, each terminal device and the call processing unit 24 in the network connection device 13 are connected by the ATM connection. This ATM connection may be a VP or a VC. This terminal device,
It should be noted that the ATM connection between the call processing units 24 in the network connection device 13 also extends over the network connection device, that is, does not extend over the ATM-LAN, like the CLSF processing unit. This ATM connection is a connection for signaling.

In this case, it should be noted that only the call processing unit in the network connection device needs to have the configuration information of both ATM-LANs.

If the terminal device / node in the ATM-LAN wants to establish an ATM connection across the inter-network connecting device 13, the call processing unit 24 is used. The correspondence is slightly different depending on whether the ATM connection that straddles the network connection device 13 is a connection for performing bandwidth management or a simple connection connection that does not require bandwidth management. Hereinafter, description will be made in order.

First, a description will be given of a case where only a simple connection connection (bandwidth management is unnecessary) is required.

(Method 1) A case where the terminal issues a connection connection request to the call processing unit in the network connection device.

In this case, “connection setting request AR
P ". As described above, in the "connection setting request ARP", the ATM directly connected to the partner is used.
Since the VPI / VCI of the connection is resolved, it corresponds to general signaling discussed in CCITT.

First, a terminal device / node (referred to as a transmitting terminal) that wants to establish an ATM connection across an internetwork connecting device is called by a call processing unit 24 in the internetwork connecting device 13.
Sends a connection request to

At this time, if the transmitting terminal does not know to which call processing unit the connection connection request should be issued, or if the signaling channel of which ATM address (VPI / VCI) is used, the connection setting across the networks is made. There may be cases where it is not known whether the request can be issued. In this case, a "call processing request ARP" asking which call processing unit should issue a connection connection request.
Will be used. That is, although the details follow the case of the CLSF processing unit, the destination address of the receiving terminal is set as the address resolution target address and the “call processing request AR
P ”. When transmitting the "call processing request ARP" through the broadcast channel, the fact is written in the broadcast cell type area of the broadcast cell format. On the other hand, if the destination address is on the other side via the network connection device, the ATM address (VPI / VCI) to the call processing unit 24 in the network connection device 13 is returned. . This response is sent, for example, even if the
It may be performed by the RP server. Upon receipt of the request, the transmitting terminal again sends a connection connection request (connection setting request ARP) to the call processing unit 24 in the network connection device 13.
Will be issued.

Incidentally, without issuing the “call processing request ARP”,
A method of directly broadcasting the “connection setting request ARP” through a broadcast channel is also conceivable.

In this way, the call processing unit 24 in the inter-network connecting device 13 that has received the connection connection request makes the call processing unit 24 between the transmitting terminal and the inter-network connecting device and the inter-network connecting device in each ATM-LAN. By establishing an ATM connection between the receiving terminals and setting the header conversion unit 26 appropriately, the two ATM connections are connected, and finally an ATM connection between the two terminals is established. here,
When establishing an ATM connection in each ATM-LAN, the call processing units 91 and 9 in both ATM-LANs
2 may be used.

(Method 2) A method in which the network connection device relays ARP In this method, when the transmitting terminal device issues a “connection setting request ARP”, the network connection device that has received the “connection setting request ARP” Relay the setting request ARP. That is, the transmitting side terminal device indicates the address of the receiving side terminal (such as a network layer address or a mail address) and the fact that the connection setting request has been made (for example, when this ARP is performed by broadcast, the broadcast cell type indicates this). Include) (connection setting request) A
Perform RP. Upon receiving this, the call processing unit in the network connection device recognizes that the connection setting request ARP is a connection setting request ARP that spans the network connection device, and passes this ARP over the network connection device. Next A
Relay to TM-LAN. This is, for example, the A
ARP request cell for RP request is transferred to the next ATM-LAN
This is done by sending (relaying) to In parallel with or before or after the ARP relaying, an ATM connection between the transmitting terminal and the network connection device is set / established.

During this time, the current ARP
You may tell them that you are inside.

When the ARP is completed, that is, when the ATM connection between the network connecting device and the receiving terminal is established, the ATM connection (the ATM connection between the network connecting device and the receiving terminal) and the transmission are performed. The ATM connection established between the terminal device on the side and the network connection device is connected. In this case, this can be performed by appropriately setting the header conversion unit in the network connection device. In this way, for the end-to-end ATM connection between the transmitting terminal and the receiving terminal, the ATM connection
The connection connection is completed by notifying the transmitting terminal device of the VPI / VCI value of the connection.

As described above, the ARP relaying is used for the ATM.
When a connection is established, it is necessary to note that no special call processing unit is required in the inter-network connecting device, and it is sufficient that the inter-network connection device simply has an ARP relaying function. Therefore, this method 2 can be considered to be a special case of the method 1 above.

Next, as a specific example of the method 2, both ATM-
An example of a flow of a connection setting request across an inter-network connecting device when a VP routing method is used in a LAN will be outlined.

For example, an example will be described in which a terminal A of the first ATM-LAN requests the setting of an ATM connection (end-to-end ATM connection) with the terminal B of the second ATM-LAN.

The terminal A of the first ATM-LAN specifies the address of the terminal B (for example, a network layer address) as an address of the address resolution, and sends out a connection setting request ARP. Upon receiving this, the interworking apparatus identifies, for example, in the call processing unit that the terminal B targeted for the address resolution of the connection setting request ARP exists in the second ATM-LAN, or Select the second ATM-LAN that is the ARP destination specified in
Relays the connection setting request ARP to the second ATM-LAN (that is, sends the connection setting request ARP toward the second ATM-LAN). At this time, the source address of the ARP cell may be rewritten and used as the address of the call processing unit in the network connection device.

In parallel with or before or after this, the call processing unit of the internetwork connecting device sets the A between the transmitting terminal A and the internetwork connecting device.
Secure TM connection. Specifically, the first AT
In the VPI value (VPI value = # P) assigned to itself (inter-network connection device) on the M-LAN side, an unused VCI value is appropriately selected (selected VCI value = # P).
#Q), and this is defined as an ATM connection between the transmitting terminal A and the network connection device. During this time, the first A
It should be noted that no setting is made in the (switching table of) the switch node in the TM-LAN.

In the second ATM-LAN, the ATM connection between the receiving terminal B and the network connection device is ARP
(I.e., the network connection device obtains an ATM address from the network connection device to the receiving terminal B. This ATM address value is
PI value = # R, VCI value = # S), and an ATM connection between the network connection device and the receiving terminal B is determined.

The call processing unit in the inter-network connecting device converts the ATM connection between the transmitting terminal A and the inter-network connecting device and the ATM connection between the inter-network connecting device and the receiving terminal B into a header conversion function. By properly setting the table,
That is, (VPI, VCI) = (# P, #Q) and (VPI,
VCI) = (# R, #S), the two ATM connections are combined, and the transmitting terminal A
An end-to-end ATM connection is established between the terminal and the receiving terminal B.

Here, the call processing unit in the network connection device is an AR
It should be noted that the relaying of P is performed, and that only the setting of the table in the network connection device is performed. That is, it is not always assumed that a special call processing unit exists in both ATM-LANs.

In the end-to-end ATM connection established in this way, as a ARP response, in the interworking apparatus, the call processing unit (VPI, VCI) =
(#P, #Q) is returned to the transmitting terminal A as the resolution result. The transmitting terminal A (VPI, V
CI) = (# P, #Q), the cell is transmitted at the ATM address, and the end-to-end communication is performed only with the receiving terminal B and the ATM layer processing without receiving the termination at the AAL or more on the way. This means that the end-to-end ATM connection has been established.

[0166] The information of the attribute of the communication to be connected (such as UPC parameters and QOS) may be added to the information section of the cell of the connection connection request ARP.

The above is a description of the case where A is transmitted from the transmitting terminal to the receiving terminal.
The flow up to the establishment of the TM connection has been described, but it is of course possible to easily establish an ATM connection from the receiving terminal to the transmitting terminal in parallel with this to enable two-way communication. In this case, the connection setting request ARP is loaded with the VCI value to be used in the reverse ATM connection (for example, from the network connection device to the transmitting terminal A or from the reception terminal B to the network connection device). (This V is used for the reverse ATM connection.)
Use CI value).

Here, the first and second ATM-LAs
N has been described on the premise that the call processing units 91 and 92 exist, however, it is not always necessary that one or more call processing units exist in the ATM-LAN.
A configuration in which call processing is performed using a call processing unit in the M-LAN is also conceivable.

Next, a description will be given of a case where bandwidth management is performed for an ATM connection straddling the network connection device, that is, a case where an appropriate QOS is obtained for the ATM connection.

In the case of the above (Method 1), each ATM-
When setting up the ATM connection in the LAN, the entity (which may be in the call processing unit) which performs the bandwidth management and the bandwidth management of the ATM transmission path in the inter-network connecting device is provided in the inter-network connecting device. Bandwidth management is performed, and A in both ATM-LANs
Only when all of the band management units of the TM connection (which may be in the call processing units 91 and 92) and the band management unit of the network connection device determine that the ATM connection can be established, The ATM connections may be connected by appropriately defining a header conversion function in the inter-network connecting device, and the ATM connection may be established.

In the case of the above (method 2), after or when an ATM connection spanning the inter-network connecting device is established, the band management units in both the ATM-LAN and the inter-network connecting device are provided with the ATM connection of the ATM connection. The validity of the bandwidth management is inquired, and if permitted, the use of the ATM connection via the bandwidth management (that is, maintaining a certain or more QOS) is permitted. Here, the bandwidth management unit determines that Q
If the use of the ATM connection while maintaining the OS is not permitted, the transmission side terminal (and the reception side terminal if necessary) is notified that the band management cannot be secured.
In this case, it should be noted that an ATM connection without bandwidth management has been established. Here, if the terminal side does not mind the connection without the bandwidth management, the terminal starts communication with the ATM connection. If it is determined that the communication is not possible with the ATM connection without the bandwidth management, The communication is abandoned. At that time, the ATM connection disconnection request may be issued.

The above-described processes are not limited to connection setting requests, but are performed in substantially the same manner when connection setting / disconnection / change requests are made.

By providing the call processing function in the network connection device, the following advantages can be obtained.

(1) The processing of an ATM connection across an inter-network connecting device requires information in both connected networks. By arranging a call / connection processing unit in an inter-network connection device located between each network and capable of knowing information in each network, the efficiency of processing an ATM connection across the inter-network connection device is improved. Can be done

(2) Each of the networks connected to the network connection device can access the call / connection processing unit. This access can be made without using an ATM connection that spans a plurality of networks.

(3) The call / connection processing unit in the network connection device can specialize in processing of an ATM connection across the network connection device.

(4) For an ATM connection spanning between networks, it is necessary for the network connection device to convert each ATM address system (VPI / VCI system). By providing the call / connection processing unit in the network connection device, elements indispensable for the conversion of the address system, such as the setting unit of the conversion table of the VPI / VCI system, are included in the call / connection processing unit, and the same processing is performed. It is possible to tightly couple with the part.

In the call processing method described in the present embodiment, the call processing unit does not necessarily need to be present in the IWU, but may be arranged at an arbitrary position in the network. However, in this case, since the call processing unit needs to set a table in the IWU, it is desired that an ATM connection directly connected to the IWU exists.

Next, FIG. 10 shows another embodiment of the connection state of the ATM connection between the terminals in both ATM-LANs and the call processing unit 24 in the network connection device 13. In the present embodiment, the call processing units 101 and 102 in each ATM-LAN
And the call processing unit 24 in the network connection device 13 are connected by ATM connections. This ATM connection may be a VP or a VC.

In this case, the ATM connection connection request which straddles the network connection device 13 is first terminated in the call processing units 101 and 102 in the ATM-LAN, and the ATM connection requested here is terminated. Is recognized by the call processing units 101 and 102, the call processing is relayed to the call processing unit 24 in the network connection device 13, and the call processing is further relayed in the network connection device 13. Is relayed from the call processing unit 24 to the call processing unit in the ATM-LAN on the opposite side. In this process, both AT
ATM connections within the M-LAN are each ATM-L
Both ATMs are established by the call processing unit in the AN, and by appropriately setting the header conversion unit in the network connection device.
The connection is established, and as a result, an ATM connection is established across the network connection device.

Here, in the call processing system shown in FIG. 10, a connection setting request from a terminal in the ATM-LAN is transmitted between the networks regardless of whether the request is closed in the own ATM-LAN. This is a method that is always output to the call processing units 101 and 102 in the ATM-LAN, regardless of whether the connection device is straddled. In this system, the ATM-LA
N, and a call processing unit 24 in the network connection device 13.
It should be noted that the ATM connection between them also crosses the CLSF processing unit or the network connection device as in the previous embodiment, that is, does not cross the ATM-LAN.

Further, by arranging the call processing units in such a manner, the call processing units 101, 10 and 10 in the ATM-LAN are provided.
It should be noted that only the configuration information in the own ATM-LAN needs to be arranged in 2 and that only the call processing unit in the network connection device needs to have the configuration information on both ATM-LANs. is there. Here, the call processing unit 1 in the ATM-LAN
The rules 01 and 102 can be set so that when the arrived ATM connection setting request is not closed to the own ATM-LAN, relaying to the call processing unit in the network connection device is sufficient.

Next, FIG. 11 shows an example in which a call processing unit does not exist in the network connection apparatus. The setting of the ATM connection over a plurality of ATM-LANs in such a form is performed by the call processing unit 11A in the first ATM-LAN 111.
Is performed in cooperation with the call processing unit 11B in the second ATM-LAN 112. That is, for example, a plurality of ATM-LAs from a terminal in the first ATM-LAN
A request for setting up an ATM connection spanning N is first passed to the call processing unit 11A in the first ATM-LAN,
The call processing unit 11A recognizes that the request is over a plurality of ATM-LANs, and
Relaying is performed to the call processing unit 11B in the TM-LAN 112.
Establish an M connection. Here, it should be noted that establishment of the ATM connection in the network connection device, that is, setting of the header conversion unit in the network connection device is also responsible for one of the call processing units. A permanent ATM connection between the two call processing units (either VP or VC)
ATM that connects the call processing units
It should be noted that a connection spans between both ATM-LANs. In this case, all call processing units basically need to have information on the configuration of the adjacent ATM-LAN inside.

As described in detail above (except for the example in FIG. 11), the cells of the C plane relating to the connection spanning a plurality of networks are basically a call processing unit for connection between LANs (the network in this embodiment). It is terminated at the call processing unit in the interconnection device. Here, it is also possible to adopt a configuration in which the cells of the M plane are terminated in the network connection device. In this case, the network management unit is arranged in the network connection device. The network management unit disposed in the network connection device can collectively manage, hold, or easily prepare the management information of the ATM-LAN from the network management unit. It is possible to greatly reduce the addition when exchanging or acquiring network management information when the LAN is configured in a hierarchical manner (it is possible to terminate before and after the network connection device).

[0185] Also, the inter-network connecting device is an ATM of another vendor.
If it is located between LANs, a protocol conversion unit (that is, a conversion unit of an ATM-LAN protocol for each vendor) may be included therein.

Also, a so-called broadcast channel broadcast in the ATM-LAN is basically terminated in the network connection device. This can be easily realized by stipulating that a certain broadcast cell is not transmitted to the next ATM-LAN but discarded in the header conversion unit in the header conversion unit in the network connection device. It is.

Note that VPI = all 1 is a broadcast cell,
Assuming that the attribute of the broadcast cell is determined by the value of the VCI, a certain broadcast cell, that is, a broadcast cell having a certain VCI value is discarded in the header conversion unit, and another broadcast cell (another broadcast cell). With respect to a broadcast cell having a certain VCI value), the broadcast cell can be easily relayed by sending the broadcast cell to the next LAN broadcast channel after subjecting the broadcast cell to header conversion as needed. is there. Also, the cells received via the broadcast channel can be easily broadcast via the broadcast channel in the next ATM-LAN by appropriately setting the header conversion unit. FIG. 12 is a diagram summarizing these.

(Second Embodiment) Next, FIG. 13 shows an ATM network according to a second embodiment of the present invention. As shown in the figure, the ATM network of the present embodiment is a first ATM-
LAN 131, second ATM-LAN 132, third A
It comprises a TM-LAN 133 and an inter-network connection device (IWU) 134.

First ATM-LAN 131, Second AT
The M-LAN 132 and the third ATM-LAN 133 are local area networks operated by the ATM system, respectively, as in the first embodiment. In addition, each A
The address system is independent within the TM-LAN. AT
If there is information to be transmitted, the terminal device or node in the M-LAN stores the information in an ATM cell regardless of whether the transmission destination is in the ATM-LAN or not. To the ATM-
Send out to LAN.

FIG. 14 is a diagram showing the internal structure of the network connection device 134. In this way, the network connection device 134
M switch 141, CLSF processing unit 142, call processing unit 1
43, IWU management unit 144, input processing units 14A, 14
, B,..., Output processing units 14X, 14Y,.

The network connection device 134 is located between two or more ATM-LANs, and the plurality of ATM-LANs.
It has a function of controlling LAN internetworking (connection between LANs).

The input processing sections 14A,...
The header value (for example, VPI / VCI value) is analyzed (and converted if necessary) for the cell, and a routing tag for routing the cell in the ATM switch is newly added to the input cell. Has functions.

The ATM switch 141 is an ATM switch having N inputs and M outputs (N and M are positive numbers, for example, N = M = 8). The cell is routed according to the routing tag assigned to the cell. A broadcast function and a multicast function may be provided inside.

The output processing units 14X,... Have a function of deleting a routing tag from ATM cells arriving via an ATM switch, and a function of converting the value of an ATM cell header if necessary.

The conversion function of the ATM cell header is a function necessary for either the input processing unit or the output processing unit. The input processing unit and the output processing unit are in charge of connection with a switch node in a LAN adjacent to the terminal unit, a terminal device, or, if necessary, another network connection device.

The function of the CLSF processing unit 142 is a connectionless service function (ConnectionLess Service Functi
on). This function is basically the same as the CL of the first embodiment.
Since it is based on the SF processing unit 23, detailed description is omitted.

The call processing unit 143 basically sets, disconnects, and sets up an ATM connection across the network connection unit 134.
It has functions to make changes, management, etc.
The function is the same as that of the call processing unit 24 of the embodiment. Therefore, detailed description is omitted.

The IWU management unit 144 controls the main network connection device 1
34 has a function of performing management and control. Network connection device 1
It is assumed that each module in 34 is connected to a bus (not shown), and control such as a change of a set value of each module from the IWU management unit 144 is performed through this bus.

Instead of exchanging information between the modules via the bus as described above, the information may be put on ATM cells and the cells may be exchanged between the modules.

In the present embodiment, various processes in the IWU are performed separately (that is, the CLSF
In the SF processing module, the call processing is performed by the call processing module, respectively), all or some of these processing can be performed by the same CPU / MPU.

Next, FIG. 15 shows the first ATM-LAN 13
1, second ATM-LAN 132, third ATM-LA
5 shows an embodiment of a node / terminal configuration inside N133. The first ATM-LAN 131 includes three switch nodes 151, 152, 153 and terminals 15A, 15B, 15C, 15D connected thereto. The second ATM-LAN 132 has two switch nodes 15
4, 155 and the terminals 15E, 15 connected thereto.
F, 15G. The third ATM-LAN 133 is
It consists of two switch nodes 156 and 157 and terminals 15H, 15I and 15J connected to them. The function of the switch node is the same as in the first embodiment.

Next, FIG. 16 shows an example in which the terminal in each ATM-LAN and the CLSF processing unit in
4 shows an embodiment of a TM connection connection state. Here, for the sake of simplicity, wiring between terminals and switch nodes and between switch nodes and networks are omitted. Further, other components such as a call processing unit and a header conversion unit in the inter-network connecting device and a switch node in the ATM-LAN are also omitted in the figure.

As described above, each terminal device and the network connection device 1
The ATM connection (VP or VC) is connected between the CLSF processing units 142 in the 34 respectively. This A
The TM connection is a connection that is not subject to band management, as in the first embodiment. A connecting the terminal device to the CLSF processing unit 142 in the network connection device 134
It should be noted that the TM connection does not span the inter-network connection device, that is, does not extend between the ATM-LAN, as in the first embodiment.

When the terminal device wants to deliver a datagram via the network connection device 134, that is, outside the ATM-LAN, this is performed in substantially the same process as in the first embodiment. The following is a brief description.

Regarding the delivery of datagrams across the network connection device 134, the CLSF in the network connection device 134
It will be delivered via the processing unit 142. That is, AT
The datagram converted into M cells is sent to the CLSF processing unit 142. Here, the terminal / node in the ATM-LAN transmits the datagram to the CLSF processing unit 142.
Since the approach (ARP) until the terminal device knows that it should be sent to the terminal device is the same as in the first embodiment, the details are omitted. As in the first embodiment, ARP for datagram delivery across the network connection device is performed, and a terminal device that sends out the datagram thereafter converts the datagram into an ATM cell, and then converts the datagram into an ATM cell. If the destination is a destination that straddles the network connection device, the previously resolved ATM address (V
PI / VCI), and connect this to the CL in the network connection device.
Send to SF processing section. The CLSF processing unit in the network connection apparatus once converts this into a network layer or a CL.
A that terminates at the layer, analyzes the destination address, and then connects to the destination (after conversion to an ATM cell if necessary)
The TMgram will be appropriately selected and the datagram will be delivered through it.

Here, even with this method, the same advantage as the advantage of the CLSF processing unit in the network connection apparatus in the first embodiment can be enjoyed.

Next, FIG. 17 shows the AT between the terminal in each ATM-LAN and the call processing unit 143 in the network connection device 134.
4 illustrates one embodiment of an M connection connection state. here,
For simplicity, wiring between terminals and switch nodes and between switch nodes and inter-network connection devices are omitted. Other components such as a CLSF processing unit processing unit and a header conversion unit in the network connection apparatus are also omitted in the figure. Also, both A
Call processing units 171, 172, and 173 are added to the TM-LAN.

As described above, each terminal device and the call processing unit 143 in the network connection device 134 are connected by an ATM connection (VP or VC). It should be noted that the ATM connection between the terminal device and the call processing unit 143 in the network connection device 134 also extends over the network connection device, that is, does not extend over the ATM-LAN, like the CLSF processing unit. is necessary.

In this case, it should be noted that only the call processing unit in the network connection device needs to have all the configuration information of each ATM-LAN.

When a terminal device / node in the ATM-LAN wants to establish an ATM connection across the network connection device 134, the call processing unit 143 is used. Basically, it is the same as the first embodiment, but there are some differences, so that a brief description will be given.

First, a description will be given of a case where only a simple connection is required.

(Method 1): The case where the terminal issues a connection connection request to the call processing unit in the network connection device.

[0213] This case is based on the first embodiment. Therefore, detailed description is omitted.

(Method 2): A method in which the network connection device relays ARP.

Also in this method, as in the first embodiment,
When the transmitting terminal issues a “connection setting request ARP”, the inter-network connecting device that has received the “ARP” relays the ARP, but the target address of the ARP is a plurality of ATM-connections connected to the inter-network connecting device 134. The first embodiment is different from the first embodiment in that it analyzes which ATM-LAN belongs to the LAN, and then relays ARP only for the ATM-LAN. In other words, the call processing unit has an internal database (a correspondence table between a connected LAN and a network address or a domain name, etc.).
(Resolution at the network level), and for the ATM-LAN identified there, the terminal /
The resolution will be applied at the node level.

Of course, ARP may be relayed to all ATM-LANs connected to the network connection device without performing network-level resolution in the network connection device. ATM-LAN connected to
ARP may be performed until ARP is sequentially completed (for example, in ascending order of switch port numbers).

When bandwidth management is performed for an ATM connection spanning the network connection device,
Is required for the ATM connection according to the first embodiment, and a detailed description thereof will be omitted.

The above-described processes are not limited to connection setting requests, and are substantially the same in connection setting / disconnection / change requests.

Here, even with this method, the advantage of the call processing unit in the network connection device in the first embodiment can be enjoyed.

[0220] The call processing unit 143 is not necessarily provided by the IWU1.
34, it may be at any position in the network.

Next, FIG. 18 shows an example of the communication between the terminal in each ATM-LAN and the call processing unit 143 in the network connection device 134.
9 shows another embodiment of the connection state of the TM connection. Call processing units 181, 182, 183 in each ATM-LAN
And the call processing unit 143 in the network connection device 134 are connected by an ATM connection (VP or VC). Also in this case, similarly to FIG. 10 of the first embodiment,
An ATM connection connection request across the network connection device 134 is first sent to the call processing units 181 and 18 in the ATM-LAN.
ATM terminated at 2,183 and required here
When it is recognized by the call processing units 181, 182, 183 that the connection spans the network connection device,
The call processing unit is relayed to the call processing unit 143 in the network connection device 134, and further, the call processing unit
A call processing unit 143 determines which ATM-LAN is a connection / connection request with a terminal / node in the ATM-LAN. Based on the determination result, the corresponding ATM-LAN is determined.
Is relayed to the call processing unit. In this process,
ATM connection within each ATM-LAN
The ATM connection is established by the call processing unit in the M-LAN, and by appropriately setting the input processing unit in the network connection device or the header conversion unit in the output processing unit, the two ATM connections are connected. AT spanning connecting devices
An M connection is established.

When the call processing unit 143 in the interworking apparatus 134 relays the arrived connection connection request across the interworking apparatus to the call processing unit in the ATM-LAN, as described above. Instead of analyzing which ATM-LAN call processing unit is to be relayed by the call processing unit in the network connection device, (if necessary,
All connected ATM-Ls except for the call processing unit in the ATM-LAN to which the connection setting request was issued)
It is also possible to adopt a method of broadcasting this to the call processing unit in the AN or a method of sequentially sending inquiries to them.

Here, similarly to the first embodiment, in the call processing system shown in FIG. 18, a connection setting request from a terminal in the ATM-LAN is transmitted to the own ATM-L.
The call processing unit 1 in the ATM-LAN, regardless of whether it is closed in the AN or spans the inter-network connecting device.
81, 182 and 183. In this system, the ATM connection between the call processing unit in the ATM-LAN and the call processing unit 143 in the network connection device 134 is also used as the CLSF processing unit or the network connection device as in the previous embodiment. It should be noted that it does not span ATM-LAN.

Further, by arranging the call processing units in such a manner, the call processing units 181 and 18 in the ATM-LAN are provided.
It should be noted that only the configuration information in the own ATM-LAN may be arranged in 2,183, and that only the call processing unit in the network connection device needs to have the configuration information on both ATM-LANs. is necessary. Here, the call processing units 181, 182, and 183 in the ATM-LAN relay to the call processing unit in the network connection device when the arrived ATM connection setting request is not closed to the own ATM-LAN. It is also possible to make a rule that it is good.

Next, FIG. 19 shows an example in which a call processing unit does not exist in the network connection apparatus. Also in this case, similarly to the first embodiment, a plurality of ATM-LAs in such a form are used.
The setting of the ATM connection over N is performed by the first AT
A call processing unit 19A in the M-LAN 191 and a second ATM
A call processing unit 19B in the LAN 192 and a third ATM-L
This is performed by cooperative distribution with the call processing unit 19C in the AN 193. That is, for example, from a terminal in the first ATM-LAN to a terminal in the second ATM-LAN,
A request for setting up an ATM connection spanning a plurality of ATM-LANs is first sent to the call processing unit 1 in the first ATM-LAN.
9A, the call processing unit 19A recognizes that the request is over a plurality of ATM-LANs,
This request is sent to the call processing unit 1 in the second ATM-LAN 192.
9B, and the plurality of call processing units cooperate and distribute to establish an ATM connection. Here, it should be noted that establishment of the ATM connection in the network connection device, that is, setting of the header conversion unit in the network connection device is also responsible for any of the call processing units. The call processing units are connected by, for example, a permanent ATM connection (either a VP or a VC).
It should be noted that it spans between LANs. In this case, all the call processing units basically need to have information on the configuration of the adjacent ATM-LAN inside. In addition, a plurality of A
Under the situation where the TM-LAN hangs, each ATM-LA
Permanent connections between the call processing units in N must be basically formed in a mesh form.

As described in detail above, in the second embodiment as well as in the first embodiment (except for the example in FIG. 19), the cells of the C plane relating to the connection spanning a plurality of networks are basically the same. It is terminated in a call processing unit for connection between LANs (in this embodiment, a call processing unit in the network connection device). Also, as in the first embodiment, an M plane cell can be terminated in the network connection device. When the network connection device is located between ATM-LANs of other vendors, a protocol conversion unit (that is, a conversion unit of an ATM-LAN protocol for each vendor) may be included therein. . Also, the broadcast channel in the ATM-LAN is basically terminated in the network connection device as in the first embodiment.

(Third Embodiment) Next, FIG. 20 shows an example of a configuration method of a large-scale ATM network as a third embodiment. This large-scale ATM network includes an ATM backbone network 201, a first ATM-LAN 202, a second ATM-LAN 203, and a third ATM-LAN 20.
4. Inter-network connection devices 20A, 20B and 20C between the ATM backbone network and each ATM-LAN. These ATM-LANs can have a hierarchical network structure via the ATM backbone network 201.

The ATM backbone network 201 is provided for each ATM.
-A network that connects LANs to each other and interfaces with a public network (not shown). Although the internal architecture is not particularly limited, it is a highly flexible, scalable and highly reliable network that can adopt various architectures such as ring / star / bus / tree / mixture. ATM-LAN 202, 203, 20
4 is the same as in the first and second embodiments.

The network connection devices 20A, 20B, 20C
It is almost the same as the network connection device 13 in FIG. 2 except that it manages internetworking between the ATM backbone network and the ATM-LAN. However, in the ATM backbone network, from the viewpoint of ensuring the reliability, the ATM-LAN Since stricter traffic management is performed, a policing mechanism is attached to the output interface of the inter-network connecting device to the backbone network side, so that the determined traffic conditions are observed. Other small ATMs
A main difference from the interworking apparatus 13 of FIG. 2 is that a protocol conversion mechanism between the LAN and the ATM backbone network is provided.

[0230] Here, for example, the ATM backbone network is a network that is managed by the management department of the entire company or university, or the management department of the whole business office. LAN installed in each room
It is. Compared with the present situation, it can be considered that the ATM backbone network corresponds to the current telephone network (PBX network), and the ATM-LAN corresponds to the current computer communication LAN. The large-scale network according to the present embodiment uses a telephone network and a computer network as AT hierarchical networks.
It is considered a network integrated by the M method. Further, the ATM backbone network may be assumed to be a network that is deployed on a nationwide scale via a dedicated line of a public network.

FIG. 21 shows the internal structure of the ATM-LAN in this large-scale network. First ATM-LA
N202 includes three switch nodes 211, 212, 2
13 and the terminals 21A, 21B, 21 connected thereto.
C, 21D. Also, the second ATM-LAN 20
3 is composed of two switch nodes 214 and 215 and terminals 21E, 21F and 21G connected thereto. Further, the third ATM-LAN 204 includes two switch nodes 216 and 217 and a terminal 21 connected thereto.
H, 21I and 21J. The function of the switch node is the same as in the first embodiment.

Next, the details of the datagram delivery method in the ATM-LAN in the large-scale network of this embodiment will be described. First to third ATM-LAN2
02 to 204, the datagram delivery method is the same, so the first ATM-LAN 202
An example will be described.

In the first ATM-LAN 202,
A unique VPI value is assigned to each of the nodes, IWUs, and terminal devices in the LAN. That is, if a cell originating from any other node / IWU / terminal has the VPI value in the VPI field of its ATM cell header, the cell is always routed to the corresponding node / IWU / terminal. You. Here, it should be noted that a unique VPI value is assigned to the switch node, the network connection device, and the like in the LAN.

For example, as shown in FIG.
When one VPI value is assigned to each terminal, "VPI value = # A" is set, and cells transmitted from any terminal / node are always transmitted to terminal 21A regardless of the number of VCI values. Routing, ie delivered. In order for the cell to be routed in this manner, A
The routing table of the switch node in the TM-LAN is set.

By setting in this way, the ATM
-Nodes / IWUs / terminals belonging to the LAN are equivalent to an ATM connection previously established in a mesh form (end-to-end). That is, any terminal / I
From a WU / node (called a transmitting terminal), any terminal /
When performing communication with an IWU / node (called a receiving terminal) (when transmitting a cell), the cell is transmitted to the receiving terminal by using the VPI value (ATM cell header value) assigned to the receiving terminal. Routed. This means that the ATM connection from any transmitting terminal to any receiving terminal is set up in a mesh form (although there is no guarantee of QOS) (VP routing method).

The format of the ATM cell header in the ATM-LAN of the present embodiment conforms to the format of a UNI (user / network interface) cell in the CCITT recommendation. ATM-L VPI value
Since the elements (nodes / IWUs / terminals) constituting the ATM-LAN are allocated so as to be unique in the AN, the upper limit of the total number of nodes / IWUs / terminals in the ATM-LAN is 256. Limited. This is because the VPI area has only 8 bits. In the present embodiment, the upper limit of the total number of nodes / IWUs / terminals in the ATM-LAN is further smaller as described later.

Here, a switch is set so as to perform the VP routing method in the switch node. That is, if an appropriate VP is set in the ATM cell header and the cell is transmitted, the cell is routed to the intended receiving terminal. Despite this, the transmitting terminal may not recognize the ATM address (VPI value) of the receiving terminal. In this case, it is assumed that the transmitting terminal has recognized the network layer address of the intended receiving terminal (or may be the physical address value of the communication board, etc.).

In such a situation, in the existing LAN (for example, Ethernet), the physical address (MAC address, for example, Ethernet address) is known even though the network layer address (for example, IP address) of the receiving terminal is known. Respond if you do not know. In such an existing LAN, in such a case, a protocol for resolving a physical address from a network layer address (address resolution protocol, AR
P) will work.

Similarly, the ATM-LA of the present embodiment is
N, from the network layer address of the receiving terminal, the ATM address (V
ARP to obtain (PI value)
To do.

The ARP method in the case where the receiving terminal exists in the same ATM-LAN as the transmitting terminal will be described below. First, the following two methods can be considered as this method.

(Method 1): Direct ARP of Sender Terminal-Receiver Terminal The sender terminal knows the network layer address (eg, IP address or E.164 address) of the receiver terminal, but this address is It is not known which ATM address (specifically, VPI / VCI value) corresponds. In this case, the transmitting side terminal determines in advance that the ATM-
ARP is performed through a broadcast channel defined in the LAN. Although the details will be described later, there are several types of ARPs, and this ARP is a “datagram transmission request ARP” therein.

Here, the broadcast channel is
This is an ATM connection that can broadcast the transmitted cell from any transmitting terminal to all nodes / IWUs / terminals belonging to the ATM-LAN. In this ATM-LAN, for example, "VPI value = all" When the cell "1" is transmitted, the cell is recognized by the network as a broadcast cell (broadcast cell), and the cell is transmitted to all nodes / IWUs belonging to the ATM-LAN.
/ Delivered to the terminal. Here, the cell transmitted as a broadcast cell may or may not be delivered to the terminal that transmitted the cell.

FIG. 23 shows an example of the format of a cell for performing a datagram transmission request ARP. As described above, the datagram transmission request ARP cell includes the transmission source address (Sou
rceAddress), Destination Address
, Broadcast cell type, and ARP type.
The source address includes the network layer address type, the network layer address of the transmitting terminal, the AT
In the M-LAN, the VP assigned to the transmitting terminal
I value is included.

Here, the network layer address type indicates which network layer address (which network layer protocol address) is the network layer address of the transmitting terminal that is subsequently included in this area. Area. For example,
As shown in FIG. 24, LLC + (SNAP (SubNetwork Attatc
hment Point) or NLPID (Network Layer Protocol
ID))). In addition, as the network layer address type,
Request for Comments (RFC) 1134 P
The same protocol type as PP (Point to Point Protocol) may be used.

The VPI assigned to the transmitting terminal
As described above, if the value is used, cells can be delivered from all the nodes / IWUs / terminals belonging to the ATM-LAN to the “transmission side terminal” (VP routing method). That is, this value indicates its own ATM address in the ATM-LAN.

The destination address includes the network layer address type and the network layer address of the receiving terminal. The network layer address type and the network layer address of the receiving terminal are the same as the relationship between the same type of source address and the same address, and thus the details are omitted.

[0247] The broadcast cell type is a field that describes what meaning the broadcast cell has. Specific "meaning of a broadcast cell" includes, for example, "datagram transmission request ARP", "connection connection request A".
RP "(meaning will be described later)," broadcast "(a cell carrying information required by all nodes / IWUs / terminals in the ATM-LAN, such as routing information, etc. This type of cell is basically all cells Node / IWU / terminal performs reception and processing). 6 for broadcast cell type identification
As a bit coding method, for example, “0000”
00; datagram transmission request ARP "," 00000
1: connection connection request ARP ","0000010;
Broadcast "and the like.

The ARP type means that the broadcast cell is an ARP
(“Datagram transmission request ARP” or “connection setting request ARP”), whether the ARP is an “ARP request” (for example, ARP type value =
0) or “ARP response” (same = 1).

As a 2-bit coding method for identifying the ARP type, for example, “00; ARP request”,
"01; ARP response", "10; RARP request", "1"
1: RARP response ". here,
The “ARP request” in ARP is an AR used for inquiring the ATM address (VPI value, etc.) of the partner.
The type of the PCell, “ARP response”, is an ARP that returns an ATM address as a response to the “ARP request”.
This is the cell type.

Note that an error correction code such as parity or CRC may be inserted in the ARP cell. This error correction code is desirably inserted at the end of the cell to facilitate pipeline processing of the cell.

It is desirable that the present ARP cell be completed in one cell without extending over a plurality of cells. This is because, when the ARP cell extends over a plurality of cells, reassembly of the information over the plurality of cells is required, which complicates processing. That is, an information packet is converted into an ATM cell,
When transferring this over a broadcast channel, the receiving terminal first refers to the destination address included in the cell, recognizes that it is a broadcast cell addressed to itself, and then refers to the source address to determine the source cell. And an information packet is obtained by performing cell reassembly for each source address. Therefore, when the broadcast cell is multiplexed and received from a plurality of transmitter terminals, the receiver terminal individually reassembles the information packet for each pair of the destination address and the source address. You need a function to do it. This means that a large cost is required for realizing the analysis and processing function of the broadcast cell, especially for the hardware. On the other hand, if the broadcast cell is completed in one cell, reassembly is not required, and the processing for each cell may be performed sequentially, which is easy to realize. This is a situation common not only to ARP cells but also to all broadcast cells.

Now, when the transmitting terminal performs ARP,
The broadcast cell is transmitted to the ATM-LAN using a "datagram transmission request ARP (ARP request)". At that time, the meaningless data is contained in the field of the ATM address of the address of the other party. The protocol may be defined in this way. All nodes / IWUs / terminals belonging to the ATM-LAN receive the broadcast cell, identify the broadcast cell type, identify that the broadcast cell is a "datagram transmission request ARP (ARP request)", Furthermore,
Whether this is an ARP addressed to you (the ARP
It is determined whether the own address is included in the destination address of the cell). If the cell is a “datagram transmission request ARP (ARP request)” addressed to itself, the ATM terminal (VPI assigned to itself) is sent to the transmitting terminal requesting the ARP. ), A response cell (ARP response) including its own ATM address (VPI value) is transmitted. This reply may be made using a broadcast channel from the receiving terminal. However, in the present embodiment, in consideration of the communication resources of the network, the transmission included in the “datagram transmission request ARP (ARP request)” cell is considered. The ATM address (VPI value) of the terminal on the side is replaced with the ATM address of the reply cell (ARP response cell).
The response is performed using the address (VPI value). In these cases, the address of the transmitting terminal (terminal that has made an ARP response), the address of the receiving terminal (terminal that has made an ARP inquiry), the broadcast cell type (using a broadcast cell, AR)
The P type is entered. Thus, the ARP inquiry (AR
P request), the transmission source address and the reception side address are exchanged.

Here, when the ARP response is made without using the broadcast channel, that is, when the ARP response is
Point to the sending terminal that made the inquiry
A case of using a point ATM connection (ATM connection based on a VP routing method using a VPI value of a transmitting terminal) will be described.

A VCI value to be used when replying to (the ARP of) the broadcast channel, such as an ARP response, is determined in advance, for example, "VCI value = 0". It is assumed that the payload (48 octets) of the cell having the “VCI value = 0” has the same format as the broadcast cell (see FIG. 25). By doing so, the broadcast cell type,
Since the information such as the ARP type, the source address, and the address of the other party is included, the “transmitting terminal (the terminal that made the ARP request)” that has received the ARP response, Is ".

As described above, the node /
FIG. 26 summarizes the flow of ARP between IWUs / terminals.
As shown in this figure, the node / IWU / terminal that has received the ARP request not addressed to itself is the cell (ARP request cell)
Should be discarded.

The other terminals sequentially update and learn the address table (correspondence table between L3 address and VPI) in the own terminal with reference to these ARP cells which are not necessarily related to the own terminal. Is also good.

(Method 2): When an ARP Server is Used The network layer address in the ATM-LAN and A
At least one server (ARP server) that manages and recognizes the correspondence with the TM address (VPI value) exists in the ATM-LAN, or the ATM-LAN does not exist.
All nodes / IWUs / terminals belonging to the LAN
When the access method to the server is recognized, an arbitrary terminal makes an inquiry to the ARP server in order to perform the resolution from the network layer address to the ATM address (VPI). However, the ARP server is not shown in FIG.

A transmitting terminal that makes an ARP request sends an ARP request cell through a broadcast channel or
The VPI value addressed to the RP server is set in the ATM cell header (at this time, for example, VCI = 0. The reason is (method 1).
And sends the ARP request cell to the ARP server.

It should be noted that if the ATM address (VPI value) of the ARP server is unknown, it is necessary to resolve this address. Here, when the resolution of the address of the ARP server is performed, the ARP
Address resolution from the network layer address of the server to the ATM address may be performed.
If the network layer address of the RP server is unknown, the AR indicates that the type of broadcast cell is an ARP request cell.
The P server may autonomously recognize and respond to this. This may be performed using a process server or the like.

The ATM address of the ARP server is
Even if it is predetermined by default (ATM-L
AN is uniquely determined).

The ARP server that has received the ARP request (datagram transmission request ARP) in this manner
When the resolution destination of the P request is a node / IWU / terminal in the ATM-LAN, a VPI value corresponding to the resolution destination is returned as an ARP response. This ARP response may be made through a broadcast channel, as in (Method 1), or may be made using a point-to-point ATM connection to the transmitting terminal that made the ARP request.

As described above, the ARP server has one A
Not only one form in the TM-LAN but also a plurality in the same ATM-LAN, or only one in a plurality of ATM-LANs and a plurality of ATMs
-It may be configured to be accessible from the LAN.

The above embodiment has described the address resolution method in the case where the receiving terminal exists in the same ATM-LAN as the transmitting terminal. Therefore, an address resolution method in the case where the receiving terminal is in an ATM-LAN different from the transmitting terminal will be described.

Basically, A in the embodiments described so far.
In a TM network, for a datagram delivered over a plurality of ATM-LANs, that is, a datagram delivered over an inter-network connecting device (IWU), an ATM connection is terminated once by a CLSF processing unit, and the datagram is sent to the ATM network. Receives network layer address processing (network layer processing or CL layer processing).

Therefore, when the address resolution destination (the receiving terminal having the network layer address) belongs to a different ATM-LAN from the viewpoint of the transmitting terminal, it is a "datagram transmission request ARP". In the ARP response to the ARP request, the resolution of the address is
VPI / VCI of ATM connection to LSF processing unit
It must be a value, that is, an ATM address.

Here, “datagram transmission request ARP”
And the distinction between “connection connection request ARP” will be described. In the “datagram transmission request ARP”, when there is a datagram to be transmitted (in the form of an ATM cell) (when only the network layer address of the datagram to be transmitted is known and the ATM address is unknown), “ARP Request this resolution as "request". On the other hand, a "datagram transmission request ARP" replies as a resolution result which ATM address is used to deliver the datagram to a desired receiving terminal. Here, the delivery destination delivered with the ATM address value returned as the resolution result is not always the receiving terminal.
For example, the destination delivered by the ATM address value is CLS.
This is the case for the F processing unit.

On the other hand, “connection setting request AR
"P", an AR requesting an ATM connection directly connected to the receiving terminal as an address resolution result
P. That is, the receiving terminal is the same ATM as the transmitting terminal.
Using the ATM address value as the resolution result, whether in the LAN or not, the ATM connection indicated by the ATM address is halfway (for example, IW
U or a CLSF processing unit, etc.), and is an ARP requesting a state in which the terminal on the receiving side is connected to an end-to-end ATM connection. Such ARP is similar to the connection setup method defined by the signaling procedure discussed in CCITT. However, at the time of the connection setting, in the present embodiment, bandwidth management (securing of bandwidth)
Overhead is not always necessary, just end-
The setting of the end ATM connection is performed (in the sense that the table in the switch node has been set), and the use of the ATM address only guarantees that the end-to-end communication can be performed through the ATM connection. It should be noted that QOS and the like are not necessarily guaranteed. If this ARP (connection setting request ARP) is used, an ATM connection coupled end-to-end can be obtained, so that QOS (Quali
ty of service), negotiation of communication attributes, and the like.

Next, if the receiving terminal is different from the transmitting terminal A
An address resolution method (“datagram transmission request ARP”) in the case where it exists in the TM-LAN will be specifically described. Also in this case, the following two methods can be considered.

(Method I): Transmission side terminal—CLS in IWU
Direct ARP of F processing unit Basically, the receiving terminal is the same ATM-L as the transmitting terminal.
According to (method 1) in the case where it exists in the AN. In other words, if the network layer address of the receiving terminal (or the physical address of the communication board, etc.) is known, but it is not known which ATM address corresponds to this address, the transmitting terminal preliminarily checks the ATM-address. ARP through broadcast channel defined in LAN
It is a method of performing.

The difference from the above (Method 1) is that the ARP
If the receiving terminal that performs the (datagram transmission request ARP) belongs to a different ATM-LAN from the transmitting terminal (if the datagram cannot be delivered unless it straddles the network connection device), the ARP The point that the response is returned from the network connection device.

[0271] The inter-network connecting device is an AT to which the transmitting terminal belongs.
The network layer addresses of all nodes / IWUs / terminals in the M-LAN are known, and it is possible to recognize that the receiving address included in the ARP request does not exist in the ATM-LAN. In this case, the network connection device sets its ATM address value (V
(PI value) as an ARP response.

Also, the ARP response may be made after confirming the existence of the receiving terminal having the receiving address included in the ARP request on the opposite side straddling the network connection device. .

[0273] Also, some routing protocol is I
It operates between WUs or CLSF processing units, etc.
The routing information may be exchanged, and an ARP response may be made based on the routing information.

The other points are almost the same as (method 1) in the case where the receiving terminal exists in the same ATM-LAN as the transmitting terminal. In this case, if the received broadcast cell is a datagram transmission request ARP, the add / drop function in the internetwork connecting device can perform an internal ARP process on the received broadcast cell (in this embodiment, the LSF processing unit). ) Needs to be dropped. For example, if the received broadcast cell is a “connection setting request ARP”, it is dropped to the internal call processing unit.

Also, the CLSF processing unit grasps the network layer address of the node / IWU / terminal belonging to each ATM-LAN connected to the inter-network connection device as described above, and the receiving side address included in the ARP request indicates the
It is necessary to analyze whether it is in the same ATM-LAN as the transmitting terminal that issued the P request, and if not, generate an ARP response and send an ARP response cell to the transmitting terminal. It is.

It is also conceivable to consider whether or not the request is addressed to the own subnet by using a subnet mask or the like.

As described above, the CLSF processing section has a datagram relaying function, an address resolution function, and a routing information processing function. Note that an address resolution function and / or a processing function of routing information may be provided separately from the CLSF processing unit.

(Method II): When an ARP Server is Used Basically, the receiving terminal is the same ATM-L as the transmitting terminal.
According to (method 2) in the case where it exists in the AN. That is, a network layer address in the ATM-LAN and A
There is a server (ARP server) that manages and recognizes the correspondence with the TM address (VPI value). To perform the resolution from the network layer address to the ATM address (VPI), the ARP server And make an inquiry.

The difference from the above (Method 2) is that the ARP
If the receiving terminal that performs the (datagram transmission request ARP) belongs to an ATM-LAN different from the transmitting terminal, that is, if the datagram cannot be delivered without crossing the inter-network connecting device, the ARP The point is that an ATM address to (the CLSF processing unit of) the network connection device is returned as a response.

From this, the ARP server uses the ATM-L
Table for each AN (network layer address and AT
M address (VPI value), see FIG. 27). In addition to the network layer addresses of the nodes / IWUs / terminals in the ATM-LAN to be processed, the ATM address of the inter-network connecting device is included. It is noted. In this way, as the ARP response to the receiving terminal belonging to the ATM-LAN different from the transmitting terminal, the network connection device (the CLSF processing unit therein) is selected.

The setting of the table in the ARP server is as follows.
It may be done manually by hand. Further, the table setting may be automatically performed by an appropriate routing protocol so that the ARP server obtains information on the routing.

The other points are almost the same as (method 2) when the receiving terminal is in the same ATM-LAN as the transmitting terminal.

In the above (method 2) and (method II), if the ATM-LAN is a network that is not routed by the "VP routing method", for example, an ATM connection is established by a call processing server. In the case of a network, the ARP server that has received the ARP request transmits a call processing server (not shown) between the transmitting terminal that has issued the ARP request and the receiving terminal that has been requested by the ARP request.
Setting, disconnection, and setting of ATM connection in ATM-LAN
An ATM connection is set up (end-to-end in the ATM-LAN, and between the CLSF processing units in the inter-network connection device outside the ATM-LAN) by using a server performing change, management, and the like. ATM connection VPI / V
A method of notifying an ATM cell header value such as a CI value to a transmitting terminal and, if necessary, a receiving terminal may be used.

[0284] In the above (method 2) and (method II), the ARP server has a function of authentication (permission of connection setting, etc.), that is, a function of performing address resolution only for a specific terminal. May be.

(Method I) or (Method II) described above
As described above, the processing for the ARP request relating to the transmission of datagrams over a plurality of ATM-LANs is performed. (Address Resolution Protocol for Datagram) If ARP fails in any way, the broadcast cell type is set to "broadcast" through the broadcast channel, and the destination network layer address is written in the destination address field. Thus, the minimum communication within the ATM-LAN can be performed.

If the ARP fails, the datagram to be transmitted to the CLSF processing unit within the inter-network connecting device (or inside or outside) may be transmitted by (CLSF
Since the processing unit recognizes the existence of all network layer addresses of the node / IWU / terminal connected to the network connection device), by transmitting the datagram to the CLSF processing unit, the data into the same ATM-LAN is transmitted. Note that gram delivery is also possible.

Next, the rules for attaching an ATM cell header relating to datagram delivery in this embodiment will be described.

The ATM cells in the ATM-LAN according to the present embodiment include UN in the CCITT recommendation.
Since the I cell is used, the value of VPI can be used from 0 to 255, and the value of VCI can be used from 0 to 65535.

As described above, the rules for attaching the VPI are as follows: each node / IWU /
At least one VPI value is assigned to each terminal, and an appropriate VPI value is set to AT
If the packet is transmitted after being attached to the M cell header, the packet is routed to the corresponding receiving terminal (the receiving terminal given the VPI value) (VP routing method). In the present embodiment, the presence of a VPI value having a special meaning is considered (for example, for meta-signaling, network management, emergency, etc.), and the ATM-LA is used.
VPI assigned to each node / IWU / terminal in N
The value is a value between 16 and 254. That is, the same AT
Node / IWU / which can exist in M-LAN
The total number of terminals is up to 239 in the present embodiment. As an exception to this VPI value assignment method, “VP
I = all 1 (255) ". In the case of this VPI value, an ATM cell having the VPI value is broadcast as a broadcast channel in the ATM-LAN.

It should be noted that the reserved VPI value (VCI
Value) may have the same meaning as the reserved VPI value (VCI value) in B-ISDN under discussion at CCITT.

Next, an example of a rule for attaching a VCI in this embodiment will be described. As described above, “VC
"I = 0" is used for the response of the broadcast cell. Also, VPI
As in the case of (1), reserved bits are set for "VCI = 1 to 15". If "VCI = 16-254", this shall be used for datagram delivery. In other words, when an arbitrary transmitting terminal sends out a datagram, the destination terminal (destination VPI) and “VC
A cell having a value of "I = 16 to 254" is used. At this time, the VPI value assigned to the
A value shall be entered. That is, “VPI value = # X”
When a node / IWU / terminal to which is assigned a datagram transmits a datagram, the VPI area of the ATM cell header obtained by converting the datagram into an ATM cell has a VPI value (data The VCI value assigned to itself (that is, “VCI value = own VPI value”) is transmitted in the same VCI area.

The node / IWU / terminal that has received this cell recognizes that the cell has been delivered to itself based on the VPI value, and sets the VCI value to “16 ≦ VCI value ≦ 254”.
Therefore, it can be recognized that this is a datagram. In order to determine the transmission source, the receiving terminal needs to use a different transmission terminal unless the ID that can determine the transmission source is in a layer higher than the ATM layer such as AAL type 3/4 for each cell. Or CLSF
It is necessary that a different ATM cell header be added to a datagram issued from a processing unit or the like. This is made possible by using the ATM cell header value assignment method as in the present embodiment.

Further, if the VPI value is addressed to itself and the cell having a VCI value of 16 to 254 is received, the inter-network connection device determines whether the cell is the inter-network connection device or the inter-network connection device. , And the cell can be routed to an internal CLSF processing unit.

As described above, when the VCI value is 16 to 254, the receiving terminal can recognize that the cell is a datagram, and furthermore, this is assigned a corresponding VCI value for each source terminal. And the datagram is forwarded, the same source ATM-
Datagram reassembly is possible even for AAL (ATM Adaptation Layer) in a mode in which packets from the SAP must be sent continuously (each transmitting terminal has a different VPI / VCI value. It should be noted that it can be used. That is,
Datagram delivery can be performed without negotiating the VCI value used between the transmitting terminal and the receiving terminal.

Note that the VCI value = 256 to 65535 can be appropriately used by, for example, negotiating how to use it between end and end.

Next, the timing of performing the datagram transmission request ARP will be described. It is not always necessary to perform the datagram transmission request ARP when the datagram falls from the OS (operating system) in the ATM board (ATM communication board). That is, when a datagram to be transmitted is generated, the datagram is made to wait on an ATM board or an arbitrary memory, a datagram transmission request ARP is performed, an address resolution is performed, and an ATM address is determined. It is not always necessary to take out a datagram from the ATM board or an arbitrary memory, convert the datagram into an ATM cell, add a resolved ATM address, and transmit the ATM cell. This is because the datagram is kept in a waiting state during the ARP, so that it is useless from the viewpoint of sending the datagram. It is considered that efficient datagram delivery can be performed by performing address resolution in advance for a destination terminal that may transmit a datagram.

The timing at which the datagram transmission request ARP is performed is determined when the node / IWU / terminal is booted (a network layer address at which the datagram transmission request ARP is to be specified is specified in the boot program, and The datagram transmission request ARP is performed by, for example, batch processing or the like).

In addition, at the time of user login of the node / IWU / terminal (login program, that is, a program started at the time of login, a network layer address at which a datagram transmission request ARP is to be performed is specified. , A datagram transmission request ARP is performed by, for example, a batch process).

For example, in the former method, any node /
For the IWU / terminal, the address of an ATM connection for a datagram with an indispensable destination (for example, a NIS server, a file server, a network management server, a network server, etc.) is resolved, and in the latter method, the user is Applications such as resolving the address of an ATM connection for datagrams with a frequent datagram communication partner can be considered.

Here, the address resolution of the ATM connection at the time of booting or logging in as described above is not limited to the datagram, that is, not limited to the datagram delivery request ARP, but the connection setting request ARP.
It is important to note that you can also do

Also, in the present embodiment, from the description that “ARP is performed at boot or login”, the setting of the switch node (setting of the routing table for VP routing) is completed before performing the ARP. It is assumed that the terminal is ready to transmit the cell to the desired destination terminal if the terminal transmits an ATM cell with an appropriate ATM address (VPI value). For example, at the time of booting the terminal / IWU / node, the setting of the switch node is performed first, and then the ARP is performed.

However, setting of the switch node in advance (setting of a routing table for VP routing)
It should be noted that the setting of the ATM connection is not completed, and "the ATM connection is set using a call setting server or the like at the time of booting or logging in" may be considered.

It should be noted that, as will be described later, when an ATM connection is set in advance at the time of booting or at the time of login, it does not particularly squeeze communication resources. That is, A
The TM connection is obtained by simply setting A in the table in the switch node without going through connection admission control or the like.
Since it is only a TM connection, it can be considered as a connection that is not subject to band management, that is, a connection with a low priority, and does not squeeze communication resources.

The reason will be described below. In the present embodiment, an ATM connection within an ATM-LAN (in some cases, an A-
ATM that includes bandwidth management)
There are connections and ATM connections that do not perform bandwidth management. An ATM connection that performs bandwidth management is a connection that can perform communication while maintaining a certain communication quality (QOS) (a connection that can expect a cell loss rate of a certain level or more and a delay time of a certain level or less), and does not perform bandwidth management. The ATM connection is a connection without any guarantee in communication quality (there is no restriction on the cell discard rate and the delay time). An ATM connection that performs bandwidth management is a connection with high priority, and an A that does not perform bandwidth management.
The TM connection is considered to be a low-priority connection. For example, only when there is no cell belonging to a high-priority connection, priority control such as allowing a cell belonging to a low-priority connection to pass is performed. Can be realized. For an ATM connection that performs bandwidth management, the connection is permitted to be set only when the cell transmission path through which the connection passes and the communication resources in the switching node are secured.
ATM connections that do not perform bandwidth management are not subject to bandwidth management, and are therefore unconditionally accepted during call / connection admission control as long as the number is not more than the maximum number of receptions of each connection management entity. The maximum number of receptions is defined by, for example, the capacity of a table in the switch node. As for the ATM connection for performing the bandwidth management, a "bandwidth management server" is particularly provided (not shown).
This centrally manages the communication resources in the ATM-LAN, executes the calculation of the evaluation function at the time of admission control, and the like. Only the ATM connection authorized by the bandwidth management server can be used as an ATM connection for performing bandwidth management.

Next, FIG. 28 shows an embodiment of an ATM connection connection state between CLSF processing units in an inter-network connection device in an ATM backbone network. Here, for the sake of simplicity, each physical wiring and the like are omitted. In addition, other functions such as a call processing function and a header conversion function in the network connection device, a switch node in the ATM-LAN, and a component in the ATM backbone network are omitted in the figure (actually, these are omitted). An ATM connection may be established from the switch node in the ATM-LAN to the CLSF processing unit in the network connection device.

As described above, in the ATM backbone network, the CLSF processing units in the inter-network connecting device are connected by, for example, a permanent connection (VP or VC). This permanent connection is also a connection that is not subject to bandwidth management. For datagrams that span the ATM backbone network (or to terminals / nodes that reside within the ATM backbone network), these C
This is performed by the relaying of the LSF processing unit. That is, the CLSF processing unit in the network connection device terminates the packet once in the network layer or the CL layer, analyzes the destination address, and (if necessary, converts it into an ATM cell again) to the destination. The connected ATM connection is appropriately selected and the datagram is delivered through this. As shown in FIG. 28, the CL in the network connection device
Since the SF processing units are connected by a permanent ATM connection, the CLSF processing unit converts the datagram (transformed into an ATM cell) into a C in the next inter-network connection device through these permanent ATM connections.
Relay to the LSF processing unit.

Now, as described above, the CL is connected in the network connection device.
By arranging an SF processing unit and distributing datagrams across networks via the CLSF processing unit in the network connection apparatus, the following advantages can be enjoyed.

(1) From the network connected to the network connection device,
The CLSF processing unit can be directly accessed. This access is performed over A
This can be performed without using a TM connection.

(2) ATM connected to the network connection device
By performing processing here for datagram delivery across networks, the address system (V
Here, the conversion of the PI / VCI value can be performed intensively.

(3) Routing information (network layer address, C
Information on L layer address, and the address and VP
Since information related to the I / VCI value is normally terminated and exchanged in the network connection device, a CLSF processing unit that analyzes and processes datagrams using the routing information is arranged in the network connection device. This is appropriate for the reason that the relationship between the routing protocol and the CLSF processing unit can be made dense and the routing table is shared.

However, of the functions of the CLSF processing unit, after terminating a datagram (connectionless packet) and once referring to the network layer address, an appropriate ATM connection (VP connected to the terminal / node having the network layer address) / VC or a function to send to the network layer address (VP / VC) connected to the CLSF processing unit, which is considered to have a function to deliver to the network layer address, requires a large device as a function in the network. It is considered that providing the function of the unit in the network leads to an increase in cost (over specification). Therefore, a large-scale ATM as shown in FIG.
In a network, the provision of a CLSF processing unit for each inter-network connecting device or for each ATM-LAN is sufficient when the amount of datagram communication between each ATM-LAN or with an ATM backbone network or an external network is sufficient. It is considered to be the final form when it becomes larger.

[0312] Therefore, an example of a development of a method of arranging CLSF processing units in a large-scale ATM network will be described below. As described above, the arrangement of the CLSF processing unit in each network connection device as shown in FIG. 28 is considered to be the final form of this development. Hereinafter, the development of the arrangement method of the CLSF processing units will be described in three phases.

FIG. 29 shows a diagram at the time of initial introduction as phase 1. As described above, only one CLSF processing unit is provided in the ATM backbone network. Note that the CLSF processing unit may be located in any of the inter-network connection devices or in any ATM-LAN.
It is assumed that the CLSF processing unit has a sufficient throughput to process datagrams belonging to the large-scale ATM network (between networks) during the phase 1. Here, as described above, the ATM-
It should be noted that datagram communication between terminals in a LAN may be performed by an end-to-end ATM connection.

[0314] In this embodiment, the CLSF processing unit does not exist in all inter-network connecting devices. However, in the network connection device, a routing protocol termination processing unit (hereinafter, referred to as a routing processing unit) operating in the large-scale ATM network. The network connection device in this phase 1 corresponds to the CLSF processing unit in FIG. Note that a routing processing unit is provided instead, or a routing processing unit is provided instead of the CLSF processing unit and the call processing unit in FIG. 2).

Normally, the network connection device is in a position to obtain routing information on a network connected to the network connection device (or another network located further ahead of a network directly connected to the network connection device). . For this reason, a routing protocol in a large-scale network may be operating in such a manner that routing information is exchanged between the network connecting devices. Even in this large-scale ATM network, the routing protocol as described above is working.
Each network connection device collects or exchanges routing information. In the present embodiment, the routing processing unit terminates the routing protocol.

Of these, the exchange of routing information or the exchange of location information (where a terminal having an address is located) (operation of a routing protocol) may be performed by delivery of datagrams between network interconnecting apparatuses. good.
ATM for exchanging routing information between internetworking devices
A connection may be provided, and if necessary, an ATM connection may be separately provided for each routing protocol to exchange routing / position information. In this way, the network connection device can obtain information on routing and position.

In this phase 1, since the CLSF processing unit does not exist in the network connection apparatus, the address resolution method when the receiving terminal is in an ATM-LAN different from the transmitting terminal is firstly used. This is slightly different from the cases of (Method I) and (Method II) described above. An embodiment will be described below.

(Method I '): Direct ARP between the transmitting terminal and the routing function in the IWU Except for the fact that the routing function in the internetwork connecting device returns a response to the ARP request from the transmitting terminal. Basically, it conforms to (Method I).

Since the routing processing section in the network connection device has routing information on the network to which the transmitting terminal belongs, it determines whether or not the ARP request is closed to the ATM-LAN. be able to. Therefore, the AR
If the P request is not closed in the ATM-LAN, that is, if it is recognized that the address to be addressed does not exist in the ATM-LAN to which the transmitting terminal belongs, the routing function sends an ARP response. Do. Here, as the ARP response, the VPI value assigned to the own network connection device is returned as in the case of the above (method I). Therefore, from the perspective of the transmitting terminal, CLS
F processing unit is inside the network connection device (and, consequently, own ATM-LAN
It cannot be determined whether or not it is located in (in).

(Method II '): When an ARP Server is Used According to (Method II). However, the ARP server may receive information on routing from a routing processing unit in the network connection device.

However, (Method I ′) and (Method I ′)
In either case I ′), there is no CLSF processing unit in the network-to-network connection device, so the routing processing unit needs to transfer the transmitted datagram to the CLSF processing unit. Hereinafter, this embodiment will be described.

This is a method performed by appropriately setting the header conversion function in the network connection device. That is, the datagram (actually, the AT
When an M cell is sent (as described above, this can be determined by the arrival of an ATM cell having a VPI value assigned to the own network connecting device and a VCI value of 8 to 254). ) Converts this to an appropriate header (described later) in the header conversion table, and performs ATM termination without terminating by the network layer or CL layer.
The cell is transferred to the CLSF processing unit only by the processing. The setting of the header conversion table (header conversion unit) may be performed when ARP is performed for the first time, or may be performed in advance.

In this case, since the CLSF processing unit needs to uniquely identify the transmitting terminal for each of the received cells ("identification" here refers to which terminal the transmitting terminal is, for example, It does not mean identifying the network layer address, etc., but the meaning of identifying "cells originating from different terminals (cells belonging to different connectionless packets)". Note that different ATM cell header values need to be provided.

The following is an example of an ATM cell header value assignment method by which the CLSF processing unit can uniquely identify a transmitting terminal, in a case where the "VP routing method" is also applied to the ATM backbone network. Will be explained.

In the ATM backbone network, "VP
Since the “routing method” is performed, at least one VPI value is also assigned to the CLSF processing unit in the ATM backbone network, and datagrams (transformed into ATM cells) going to the CLSF processing unit are assigned to the CLSF processing unit. , The V
A PI value is assigned (registered in the header conversion function). Therefore, the CLSF processing unit identifies the transmitting terminal based on the VCI value.

As the VCI value, for example, as shown in FIG. 30, for the first eight digits of the VCI value, the ATM-
As for the VPI value in the LAN and the subsequent lower eight digits, the VPI value on the ATM backbone network side assigned to the network connection device is used. By doing this,
The CLSF processing unit can uniquely identify the transmitting terminal from the ATM cell header of the received cell (since different VCI values are always used if the transmitting terminal is different). By registering such a VCI value in the header conversion function, the network connection device can transfer the datagram delivered across the network connection device to the CLSF processing unit using only the ATM layer processing. Becomes By taking such a method, 64k (=
The CLSF processing unit can support terminals up to 2 ^ 16).

This rule can be applied not only to a two-layer network as in the present embodiment but also to a multi-layer network by appropriately devising the VCI field hierarchy.

Here, in the ATM-LAN, there is a possibility that the datagram destined for the interconnection device itself may be transferred to the CLSF processing unit according to the above-described rules.
A plurality of VPI values, for example, two of #X and #Y, are given to the network connection device, and when the resolution of the address of the network connection device itself is obtained by ARP, "VPI value = # X"
And the address resolution of the receiving terminal that straddles the network connection device is determined as “VPI value = #
Y "as ARP responses, respectively.
When "VPI value = # X", the datagram is taken into the connection device between own networks, and when "VPI value = # Y", it is determined that the datagram is not a datagram addressed to the connection device between own networks, and the header is determined. After converting the header by the conversion function, the cell can be transferred to the CLSF processing unit.

On the other hand, when a cell is received from the CLSF processing unit, the reverse of the above processing is performed, and the AT
It should be noted that the datagram may be sent to a terminal in the M-LAN.

A specific example of the above exchange is shown in FIG.
Put together. Descriptions of switch nodes and the like are omitted. ATM-LAN 311 and ATM backbone network 3
No. 12 is operated by the VP routing method, and the ATM-
The terminals 31A, 31B, 31C and the network connection device 31P in the LAN 311 have VPs #A, #B, #C, and #P, respectively.
An I value has been assigned. ATM backbone network 31
In the CLSF processing unit 31X and the network connection device in
X and #Y VPI values are assigned.

A terminal in the ATM-LAN 311 transmits a datagram (A) distributed over the internetwork connecting device 31P.
As shown in the figure, "VPI
Value = # P ”is transmitted. As mentioned above, VCI
The value is transmitted with the VPI value assigned to itself (for example, “VCI value = # A” in the terminal 31A).

The header conversion table (the side going from the ATM-LAN 311 to the ATM backbone network) in the network connection device 31C is set as shown in the figure, and the datagram is automatically converted into an ATM cell. Can be transferred to the CLSF processing unit only by the ATM layer processing.

For example, the above datagram is
To go to the processing unit, the VPI value is rewritten to #X,
Further, the VCI value is as follows: (1) The last eight digits of the VCI value are
The CLSF processing unit, which has received the cellized data, determines the VPI value (in this case, = # Y in this case) of the internetwork in order to determine from which internetwork the datagram was transferred. ) Enters.

(2) The first eight digits of the VCI value are the VPI of the transmitting terminal in order to identify the transmitting terminal for each network connection device.
A value (in this case, = # A) is entered.

The operation is performed as follows.

In this example, the cell reaching the CLSF processing unit has a VCI value other than 16 to 254 although it is a datagram.
Since basically only datagrams are transferred to the processing unit, this is valid as an ATM cell header value assignment rule for the CLSF processing unit (that is, in this example, the ATM-L is used in the CLSF processing unit in the ATM backbone network).
This means that a VCI value different from the VCI value assignment rule in the AN is assigned). Note that a datagram directed to the CLSF processing unit itself can be recognized as “a datagram addressed to itself when the first eight digits of the VCI value are 0”.

On the other hand, a datagram (which is converted into an ATM cell) from the CLSF processing unit to the network connection device is:
The VPI value is the VPI value of the network connection device (in this case, = #
Y), and the VCI value is the VPI value (= # X in this case) assigned to the CLSF processing unit in the lower 8 digits, and the VPI value of the receiving terminal (the receiving terminal) in the upper 8 digits. Is the terminal 31B, == B) is entered.

As shown in the figure, the header conversion unit (on the side from the ATM backbone network 312 to the ATM-LAN 311) in the network connection device has an ATM cell header value sent from the CLSF processing unit side, that is, the VCI value. Referring to the eight digits, it recognizes that this is a datagram sent from the CLSF processing unit, and
With reference to the digit, the value of V of the cell to be transmitted to the receiving terminal is
Used as PI value. Here, as described above, the VPI value assigned in the ATM-LAN 311 is entered as the VCI value of the cell transmitted to the receiving terminal.

The ATM backbone network 312 has 20
If zero or more interworking devices (or ATM-LANs) do not belong, the lower eight digits of the VCI value are further hierarchized to, for example, four digits or four digits, and the multistage configuration of the interworking devices is changed. It is also possible to take.

In the present embodiment, the number of ATM-LANs connected to the network connection device is one. However, when the number of ATM-LANs connected to the network connection device is plural, the ATM Each ATM in the backbone network
-One VPI value may be assigned to each LAN.
In this case, in the ATM backbone network, the network connecting device has a plurality of VPI values. Also, a plurality of ATM-LAs connected to the network
When the VPI value (equal to the number of N) cannot be assigned, the datagram (transformed into ATM cells) is transmitted between the CLSF processing unit and the header conversion unit of the internetworking device to which terminal of which ATM-LAN. Since it is necessary to decide whether to allocate the VCI values, it is necessary to negotiate a logical allocation method of the VCI values.

Next, FIG.
The figure when adding a processing part is shown. Thus, CLS
A plurality of F processing units are provided in the ATM backbone network. Some of the CLSF processing units may be located in any of the inter-network connection devices or in any ATM-LAN. In such a case, for example, the datagram traffic on the network increases, and
It is considered that the throughput of the CLSF processing unit set in the above becomes insufficient and the CLSF processing unit is added.

It should be noted that also in this embodiment, the routing processing unit exists in the network connection device. In this phase 2, the address resolution method when the receiving terminal is in a different ATM-LAN from the transmitting terminal is the same as in phase 1. However, since the CLSF processing unit has been added, the transfer destination of the datagram (which is converted to an ATM cell) is added to the CL in the header conversion function in some network connection devices.
In order to change to the SF processing unit, the header rewriting table is rewritten so as to be converted into an ATM connection header (ATM cell header) connected to the additional CLSF processing unit. It should be noted that the CLSF processing unit can be added (deleted or changed in some cases) only by this rewriting, so that the CLSF processing unit can be added or removed very easily.

In this case, it is necessary for the transmitting side terminal (receiving side terminal) in the ATM-LAN to completely recognize a change in the inter-network connecting apparatus due to the addition of the CLSF processing unit (a change in the setting of the header rewriting table, etc.). It should be noted that there is no. The transmitting terminal can perform communication using the address resolution and the datagram in exactly the same manner as in phase 1. This is the same not only in phase 2 but also in phase 3 described later. By repeatedly performing the addition (or note that the CLSF processing unit may be switched or changed) as in the phase 2, FIG.
8, it is possible to change to a mode (phase 3) in which the CLSF processing units are arranged in all the network connection devices. Here, when a CLSF processing unit is added in the network connection device, a CLSF processing unit may be added instead of the routing processing unit, or the CLSF processing unit may be arranged in a state where the routing processing unit is arranged. The part may be added. This is done only by changing the settings of the header conversion table (header conversion unit) and the add / drop processing unit (or the switching function when the distribution of cells in the network connection apparatus is performed by the ATM switch function). It should be noted that the number of CLSF processing units can be increased and decreased very easily.

Here, the ATM connection between the inter-network connecting device and the CLSF processing unit and the ATM connection between the CLSF processing unit and each terminal are determined by the AT which does not perform band management.
It is necessary to pay attention to the fact that M connection is sufficient.

Next, returning to FIG. 28, “CLSF processing section A
RP "will be described. Referring to FIG. 28, when the CLSF processing unit delivers a datagram, a permanent ATM connection (PVC, PV) to the CLSF processing unit in the inter-network connecting device extended in the ATM backbone network.
In P), it may be necessary to select which PVC / PVP should be sent. That is, the destination address (network layer address,
Alternatively, a mail address or the like may be stored in a table in the CLSF processing unit (in the case of this destination address, the P
If it is not registered in the table in which the instruction to send to the VC / PVP is written, the
Resolution for determining whether to send the datagram to C / PVP is performed. This is called "CL
ARP between SF processing units ". There are several methods for performing the “ARP between CLSF processing units” as follows.

(Method 1): Method Using ARP Server This is a method similar to (Method 2) of ARP for datagram distribution in the ATM-LAN of the first embodiment.
An ARP server is placed in the ATM backbone network (not shown). This ARP server has a table inside, and this table contains a destination address, VPI / VCI
Values correspond. This VPI / VCI value is stored in a CLSF processing unit (in some cases, AT
CLSF processing unit in the M backbone network. This is responsible for the terminal device in the ATM backbone network).
It is the VCI / VPI value of VC / PVP. FIG. 33 shows one embodiment of the outline of the table. As described above, there are as many tables as the number of CLSF processing units, and each table corresponds to an inquiry from one CLSF processing unit. The ARP server having received the inquiry analyzes the VPI / VCI value corresponding to the destination address with reference to the table, and returns the analysis result (VPI / VCI value) to the CLSF processing unit of the inquiry source.

Here, in a system in which the ATM backbone network allocates one VPI to terminals / nodes inside the ATM backbone network and performs routing of cells in the ATM backbone network using the VPI, the VPI to be assigned by the destination address is used.
Since the value is uniquely determined, there is no need to prepare a table for each destination address, and only one table need be prepared, so that a large amount of tables can be reduced. In this case, when sending a datagram or ARP, for example, the first eight digits of the VCI value are all 0 (this is used as a datagram delivery mark.
The cell in which the first eight digits of the value are all 0 is transferred to the CLSF processing unit), and the last eight digits are filled in with the VPI value of the interworking apparatus to which the source CLSF processing unit belongs. The unit can know the source of the cell and can continue the datagram delivery process. When returning the ARP, if the VPI of the transmission source CLSF processing unit included in the VCI area is used,
RP reply can be made without using broadcast. If the CLSF processing unit does not need to know the source CLSF processing unit,
It is not always necessary to put the VPI of the interworking apparatus to which the source CLSF processing unit belongs in the last eight digits.

[0348] Others are the same as those of the first embodiment.
This is almost the same as the method 2 described for ARP to the CLSF processing unit in the LAN.

(Method 2): A method in which ARP is performed directly between CLSF processing units in an ATM backbone network In an ATM backbone network, the ALS is connected to another CLSF processing unit in an inter-network connection device or in an ATM backbone network. This is a method for performing ARP directly. In other words, the CLSF processing unit on the side that performs ARP (that is, the side that asks for the VPI / VCI value) broadcasts an ARP request to other CLSF processing units, or a method similar to the broadcast (for example, listening in turn). This is a method in which the corresponding CLSF processing unit notifies the VPI / VCI value addressed to itself by responding to this.

In this system, one VPI is allocated to each node / inter-network connection device in the ATM backbone network, and in a system in which routing in the backbone network is performed using the VPI (VP routing system), the VPI of the own device is used. This can be done by notifying the value and the appropriate VCI value.

When the CLSF processing unit that has received the ARP determines that this is the destination address to which it is responsible, the CLSF processing unit and the transmission source CLSF
VPI / VC of PVC / PVP tied between processing units
This can be performed by notifying the I value, and it is not always necessary to use the VP routing method.

(Method 3): A method in which a table relating to all addresses is given to each CLSF processing unit in advance without using ARP When the overall network scale is not so large,
A table relating to all addresses can be given in advance to each CLSF processing unit. The loading of this table may be performed, for example, when the CLSF processing unit is started.

In this case, for a destination address not in the table, a destination CLSF processing unit is determined by default, and the destination CLSF processing unit may be transferred to this destination address. It is also possible to adopt a construction method such as having

After such “ARP between CLSF processing units”, the datagram is the VPI / VC which is the result of ARP.
Transfer to the CLSF processing unit indicated by I (relaying)
The destination address of the datagram delivered to the CLSF processing unit is analyzed again, and is relayed to an ATM connection connected to an appropriate destination. The set of the destination address and the VPI / VCI value resolved by the “ARP between the CLSF processing units” may be stored in a table in the CLSF processing unit thereafter.

With respect to the ATM backbone network, in order to realize efficient routing, sufficient consideration is required for addressing. For example, in order to facilitate creation of a subnet mask, a network address having a large value overlap is given to a LAN that is physically close or a LAN that is connected to a switch node in the same network connection device / ATM backbone network. It is.

Also in this method, the advantages of the CLSF processing unit in the network connection device in the first and second embodiments can be enjoyed.

Next, FIG. 34 shows the connection state of the ATM connection between the call processing unit in the network connection unit and the terminals in the ATM-LAN, and the ATM connection between the call processing units in the network connection unit in the ATM backbone network. 4 shows an embodiment of a connection state. Here, each physical wiring is omitted for simplicity. Further, other components such as a CLSF processing unit and a header conversion unit in the network connection device, a switch node in the ATM-LAN, and a component in the ATM backbone network are omitted in the figure. Actually, the switch node in the ATM-LAN also sends A to the call processing unit in the network connection device.
A TM connection may be established.

As described above, each terminal device in the ATM-LAN and the call processing unit in the network connection device are connected by an ATM connection (VP or VC). This A
It should be noted that the ATM connection between the terminal device in the TM-LAN and the call processing unit in the inter-network connecting device is not a connection over the inter-network connecting device as in the first and second embodiments. It is.

In the ATM backbone network, the call processing units in the inter-network connecting device are connected by, for example, a permanent connection (VP or VC). It should be noted that this permanent connection is also an ATM connection closed in the ATM backbone network, and is not a connection that spans an inter-network connecting device.

[0360] Further, by making the connection between the call processing units by a permanent ATM connection, it is possible to reduce the connection setting overhead generated every time processing information is transferred between the call processing units, and to reduce the number of connections between three or more networks. The processing of a call / connection spanning the network can be always performed by relaying between the call processing units in the network-to-network connection device via the permanent ATM connection. Also, it is possible to easily monitor traffic used for call / connection processing.

Hereinafter, a terminal / node in the ATM-LAN (hereinafter, referred to as a transmission side terminal) transmits another ATM-LAN.
AT to terminal / node (hereinafter referred to as receiving terminal) within
A process for setting the M connection will be described.

Now, when a terminal device / node in the ATM-LAN wants to establish an ATM connection over an inter-network connection device and an ATM backbone network, the call processing unit in the inter-network connection device is used. become. Basically, it is the same as the first and second embodiments, but there are some differences.

First, a description will be given of a case where only a simple connection connection (bandwidth management is not required) is required.

First, the point that a terminal / node in the ATM-LAN (referred to as a transmission side terminal) which issues an ATM connection setting request acts on the call processing unit in the network connection apparatus is different from the first and second embodiments. It is performed in almost the same process.
Detailed description is omitted.

Next, the call processing unit in the inter-network connecting device that has received the request searches the inter-network connecting device for which the target address of the “connection setting request ARP” is responsible for the call processing unit. The ARP is relayed to a call processing unit in the network connection device. As shown in FIG. 34, since the call processing units in the network connection apparatus are connected by permanent ATM connections, the call processing units receive the connection setting request via these permanent ATM connections. Relaying is performed to the call processing unit in the network connection device.

[0366] Here, the call processing unit may send the call to any PVC / PVP among the permanent ATM connections (PVC, PVP) of the call processing unit in the inter-network connecting device set up in the ATM backbone network. Sometimes you need to make a choice. That is, the destination address of the analyzed connection connection request (which may be a network layer address or a mail address) is stored in a table inside the call processing unit (in the case of this destination address, this PVC / PV
P, or relaying to this call processing unit) is not registered in the table), the resolution to which PVC / PVP the connection request should be transmitted from the destination address. Will be held.
This is referred to herein as “ARP between call processing units”.

There are several methods for performing the "ARP between call processing units" as follows.

(Method 1): Method Using ARP Server This is almost the same as the case of the ARP server in the CLSF processing section in FIG. 28, except that the VPI / VCI values in the table inside the ARP server are different from those of other networks. VCI of PVC / PVP leading to the call processing unit in the interconnection device (possibly, the call processing unit in the ATM backbone network, which is responsible for the terminal device in the ATM backbone network)
/ VPI value. This table can easily be shared with the table used in the CLSF processing unit.

(Method 2): A method in which ARP is performed directly between call processing units in an ATM backbone network. In an ATM backbone network, a call is sent to another call processing unit in an inter-network connection device or in an ATM backbone network. This is a method for performing ARP directly. Since it can be realized by a method substantially similar to (method 2) in the CLSF processing unit shown in FIG. 28, the details are omitted.

(Method 3): A method in which a table relating to all addresses is given to each call processing unit in advance without using ARP. Similar to the case of the CLSF processing unit, if the entire network scale is not so large, all the addresses are set. Can be given to each call processing unit in advance. The loading of this table may be performed, for example, when the call processing unit is started. Also, the table is set to CLSF
It can be easily shared with the processing unit. See CL for details
Since it is almost the same as that of the SF processing unit, the description is omitted.

In this case, for a destination address not in the table, a transfer destination call processing unit may be determined by default, and the destination call processing unit may be transferred here. This default transfer destination call processing unit can also adopt a configuration method such as having all the complete tables.

[0372] After such "ARP between call processing units", the connection setting request includes a VP which is a result of the ARP.
In the connection setting request transferred (relayed) to the call processing unit indicated by the I / VCI and delivered to the call processing unit, the destination address is analyzed again, and the receiving terminal in the first and second embodiments is analyzed. Performs the same operation as that of the call processing unit in charge of. The set of the destination address and the VPI / VCI value resolved by the “ARP between call processing units” is thereafter stored in a table in the call processing unit.

The above-described processes are not limited to connection setting requests, but are performed in substantially the same manner when connection setting / disconnection / change requests are made.

Note that a permanent ATM connection is not established between the call processing units in the network connection apparatus, and a permanent ATM connection / ATM connection is established between the call processing units each time a connection setup / disconnection / change request is made. There is also a conceivable method of extending ARP and exercising ARP over the entire ATM backbone network.

When bandwidth management is performed for an ATM connection over an ATM backbone network, that is, when an appropriate QOS is requested for the ATM connection, the network connection device conforms to the first and second embodiments. Is omitted.

Also in this method, the advantages of the call processing unit in the network connection device in the first and second embodiments can be enjoyed.

It should be noted that in the case of the call processing unit, similarly to the case of the CLSF, the call processing unit can be gradually increased / decreased / changed as shown in FIGS. 29, 32 and 28. is there. In the same manner, such a change can be dealt with only by a simple change of the table in the IWU. This is C
This is a general point not only for the LSF and the call processing unit but also for any server.

Next, FIG. 35 shows another embodiment of a method for arranging the call processing units in a large-scale network. In the example of FIG. 35, a call processing unit 351 is arranged in the ATM backbone network. The number of call processing units in the ATM backbone network is not necessarily one, but may be distributed (load distribution or function distribution). When the call processing unit in the network connection device desires an ATM connection across the ATM backbone network, the call processing unit 351 relays the connection connection request. Here, as shown in FIG. 35, the call processing unit in the network connection device and the call processing unit 351 in the ATM backbone network are connected by a permanent ATM connection. It should be noted that this permanent ATM connection is not a connection over an inter-network connecting device, but is closed within an ATM backbone network. Call processing unit 3
Reference numeral 51 designates collectively setting, disconnecting, and managing the ATM connection in the ATM backbone network, and this call processing unit sets / disconnects / changes the ATM connection traversing the ATM backbone network between the inter-network connecting devices. I do. A call processing unit in the inter-network connecting device couples the ATM connection in the ATM-LAN and the ATM connection between the ATM backbone networks by appropriately setting a header conversion unit, and controls the ATM connection across the ATM backbone network. I do.

A similar operation is performed for a terminal /
It is also used when controlling an ATM connection between a node and a terminal / node in the ATM backbone network.

By arranging the call processing units in such a manner, the call processing units 351 in the ATM backbone network are provided.
Only within the ATM that connects to the ATM backbone network
It should be noted that the configuration information about the LAN may be held. Here, the call processing unit in the network connection device includes:
It can be said that it has a multiplexing function to the call processing unit 351 in the ATM backbone network.

Now, the ARP as described above
(From network layer address to ATM address = V
When performing address resolution to the PI value), the A
The entity that performs RP is the sending terminal (node / IWU / terminal)
There are also several possible cases regarding the location of the location. Hereinafter, each case will be described. Here, the term ATM board is used,
This is a communication expansion board for a terminal device that performs communication using the ATM communication system, and is a board (substrate) having an ATM interface and a terminal internal bus interface.

First, the ARP request will be described.

(Case 1): The ATM board autonomously sets the AR
When making a P request.

In this case, a datagram to be transferred is received from an upper level (for example, OS) via the terminal internal bus interface. This datagram is a network layer datagram (for example, an IP datagram).
Therefore, the network layer address of the transfer destination is notified to the ATM board from the upper level (in a form stored in an appropriate area of the network layer datagram).

[0385] In parallel with this, information relating to a MAC address different from that of the ATM system may be transferred to the ATM board. This is because existing terminals (for example, UNI
This situation is conceivable when, for example, an ATM board is used as a communication board without making any changes to the settings of an OS or a device driver in place of the Ethernet board of (X / TCP / IP + Ethernet). In other words, considering the previous example, the OS is that the communication board is an ATM.
In this case, the user does not know that the board is a board, but recognizes that it is an Ethernet board.

In this example, the function of the ARP request is A
Since the TM board has a network layer address and an ATM address (in this embodiment, VPI
The correspondence table (ARP table) with the values (in general, ATM cell header values including VPI / VCI values, etc.) is A
It will be on the TM board. If the resolution of the ATM address is not completed for the destination address of the datagram received from the upper
The board will autonomously make an ARP request.

The ARP may be such that the ARP request cell is generated by a processor or the like mounted on the ATM board in terms of software / firmware, or may be generated by dedicated hardware on the ATM board. ARP generated
The request cell is sent to the ATM-LAN in the manner described above. During the ARP, the datagram waits in the memory on the ATM board. During this time, the ATM board follows the datagram,
The processing relating to the datagram sent from the host may be performed, or the datagram may be converted into an ATM cell.

When an ARP response is returned, an ATM cell header is generated according to the rules described above using the ATM address (VPI value) described in the response as an ATM cell header, and the datagram (transformed into an ATM cell) is generated. A
Send it out to TM-LAN.

In addition, the ATM address whose address has been resolved in parallel with the
Register in the P table. By this registration, the next datagram to the same destination is transmitted from the network layer address to the A without the address resolution.
Conversion to a TM address can be performed.

The conversion from the network layer address to the ATM address may be performed by a processor or the like mounted on the ATM board in terms of software / firmware, or dedicated hardware on the ATM board may be used. You may go.

Note that the conversion table between the network layer address and the ATM address registered in the ARP table may be deleted if it has not been used for a certain period of time. This is a method that can be applied when deciding data to be deleted when there is data to be newly registered because the size of the ARP table is limited.

(Case 2): When an ARP request program is running as an OS program in the terminal or a device driver.

In this case, it is the program in the terminal that makes the ARP request. The activation of the ARP is as follows: (1) There is a correspondence table between the network layer address and the ATM address in the ATM board. When starting (2) There is a correspondence table between the network layer address and the ATM address on the OS or device driver side, and the emergence of a network layer address that does not exist in the correspondence table allows the OS or device driver to autonomously determine whether an ARP request is necessary. Recognition and activation of the ARP request processing program of the OS or the ARP request processing program of the device driver can be considered. In the case of (2), a correspondence table between the network layer address and the ATM address may exist in the ATM board.

The ARP processing program may be operated in several cases. (A) The ARP request processing program includes an ARP request cell or a "payload of the cell (that is, the content indicating that the cell is an ARP request) + Autonomously generates "information that informs the ATM board that the information is ARP", and notifies the ATM board. The ATM board sends the above information as an ARP to the ATM-LAN. That is, the ARP processing program has the ability to generate ARP information.

(B) The ARP request processing program issues an instruction to the ATM board that “perform an ARP request for the network layer address”. The ATM board generates an ARP request cell and sends it out to the ATM-LAN.

The following two cases can be considered.

[0397] Further, the AT when the ARP response is returned
There are several cases of correspondence of the M board. (I) There is a correspondence table between the network layer address and the ATM address in the ATM board on the ATM board, and the ATM address of the resolution result is registered in the correspondence table. .

(Ii) There is a correspondence table between the network layer address and the ATM address on the OS or device driver side, and the ATM address of the resolution result is registered in the correspondence table.

(Iii) Both the ATM board and the OS or device driver have a correspondence table between the network layer address and the ATM address in the ATM board, and register the ATM address of the resolution result in both correspondence tables.

Here, in (iii), the correspondence table on the ATM board may be a cache of the correspondence table of the OS or the device driver. That is, only a part of the data in the correspondence table held by the OS or the device driver is described. Various methods such as a FIFO method and a round robin method can be considered as a method of selecting a part to be described.

Next, the ARP response will be described.

(Case A): The ATM board autonomously sets the AR
When making a P response.

[0403] This is because the ATM interface
The ATM board that receives the P request autonomously sends the A
RP request is detected, and the ARP request addressed to the
This is a case where a P response is made.

FIG. 36 shows an example of an ATM board for performing such processing. The ATM board includes an ATM interface 361, an ARP filter unit 362, an ARP processing unit 363, an insertion unit 364, a bus interface unit 365.
Consists of The ATM interface 361 has a function of interfacing the ATM board with the ATM-LAN.

The ARP filter section 362 extracts an ARP cell from the cells flowing on the input transmission line, and
It has a function of dropping a cell to the ARP processing unit 363. In addition, a function of inserting an empty cell into a cell slot that has become empty after dropping may be further provided.

[0406] The ARP processing section 363 analyzes the ARP cell sent from the ARP filter section 362, and
It checks whether the cell is an ARP cell addressed to itself by analyzing a destination address, discards the cell if it is not an ARP cell addressed to itself, and generates an ARP response cell if it is an ARP cell addressed to itself. , The ARP response cell into the insertion unit 3
64 is provided. The processing of the ARP processing unit may be performed by hardware logic, or may be processed by software / firmware by a processor or the like.

[0407] The insertion unit 364 appropriately multiplexes the ARP response cell sent from the ARP processing unit 363 and the cell sent from the bus interface unit 365, and
The TM-LAN side has a function of transmitting the cell.

[0408] The bus interface unit 365 has a function of interfacing the ATM board with the bus inside the terminal.

Here, in FIG. 36, only a typical ATM cell transfer path is indicated by an arrow. Actually, a host CPU or the like accesses each other via a terminal internal bus with each module, such as notifying the ARP processing unit of the network layer address of the terminal device or exchanging an error notification. There is a control line for realizing the above function.

[0410] The ARP request cell arriving at the ATM board is sent to the ARP processing section 363 by the ARP filter section 362.
Side, and the ARP cell is addressed to itself (for the terminal on which the ATM board is mounted), the ARP processing unit 363 generates an ARP response cell, and the insertion unit 3
The data is transmitted to the ATM-LAN side via the H.64. that time,
The network layer address and the AT in the ATM board
If there is a correspondence table with the M address, the information contained in the ARP request cell (specifically, the transmission source address) is used, and the information contained in the transmission source address (the network of the transmission source terminal) is used in the correspondence table. (Corresponding information between a layer address and an ATM address) may be registered.

If the OS or device driver has a correspondence table between the network layer address and the ATM address, the above information (correspondence information between the network layer address of the transmission source terminal and the ATM address)
May be notified to the OS or the device driver side in order to register the information.

As can be seen from the above description, the ARP request and response are both performed by the ATM board or software (ie, by the OS or device driver).
In addition to the form of performing the ARP request, "ARP request is performed by software (that is, the OS or the device driver
It should be noted that a form such as "ARP response is performed by the ATM board" is also possible.

(Case B): When the OS program or device driver in the terminal makes an ARP response.

In this case, it is the program in the terminal that makes the ARP response. The activation of the ARP response program is as follows: (1) The case where the ATM board recognizes the arrival of the ARP request cell addressed to itself and notifies the OS or the device driver side.

(2) The ATM board transfers the ARP request cell and other cells to the OS or the device driver without distinction, and the OS or the device driver recognizes the arrival of the ARP request cell addressed to itself, and To start the ARP response processing program of the OS or the ARP response processing device driver.

There are two possibilities.

[0417] Several cases are considered as the operation of the ARP response processing program. (A) The ARP response processing program includes an ARP response cell or "payload of the cell (that is, the content of the ARP response)".
+ Information for telling the ATM board that the information is an ARP response "is generated autonomously and notified to the ATM board.
The ATM board uses the above information as an ARP response cell
Send it out to TM-LAN. That is, the ARP processing program has the ability to generate ARP response information.

[0418] (B) The ARP response processing program issues an instruction to the ATM board that "ARP response can be made for the network layer address (and ATM address)". The ATM board generates an ARP response cell,
And sending the packet to the TM-LAN.

[0419] In the third embodiment, the AT
Although the example using the "VP routing method" as the cell routing and addressing method in the M-LAN or ATM backbone network has been described, the various methods in the present invention are not limited to the VP routing method. It should be noted that the present invention can be applied to general ATM communication systems.

[0420] For example, instead of the "VP routing method", a processing unit such as a call processing server sets an end-end ATM connection every time a connection connection request is issued (the ATM cell header is rewritten during cell transfer. May be used), and a method of performing end-to-end communication (hereinafter, referred to as
ATM-LAN or ATM backbone network), or "ATM-LAN is a VP
The ATM backbone network is operated by a call processing server system, and the ATM backbone network is operated by a VP routing system, and the ATM backbone network is operated by a call processing server system. The present invention is also applicable to an ATM network in which various systems are mixed, such as a VP routing system for some ATM-LANs and a call processing server system for some ATM-LANs. Are easily applicable.

Also, in the present embodiment, the ATM-L
The datagram delivery method in the AN or ATM backbone network has been described.
It should be noted that the present invention is not limited to a LAN or an ATM backbone network, but can be easily applied to a subnetted ATM-LAN or a subnetted ATM backbone network.

Also, the network connection device in the first, second, and third embodiments described so far is an optional ATM-communication device.
It can be considered as belonging to the LAN or the ATM backbone network. It is also possible to consider that the network connection device belongs to all connected networks.

The network connection device in the embodiments described above is a computer (workstation or the like).
Further, an extension unit / extension board may be added to the configuration, or a structure integrated with a computer may be adopted. In this case, the CLSF processing unit, the call processing unit, and the IWU management unit may be configured to be realized by the CPU of the computer.

(Fourth Embodiment) Next, an embodiment relating to datagram communication within a subnetwork will be described. Here, the case of a general ATM network, Ezaki, Tsuda, and Natsahori: "Method for Realizing Data Transfer in ATM-LAN", as described in the Institute of Electronics, Information and Communication Engineers (Information Network Workshop, March, 1993) Na V
A case where the PI routing method is applied will be described.

(Embodiment 4-1) A case where datagram communication inside a subnetwork is applied to a general ATM network.

FIG. 40 shows the embodiment, in which the terminal (T
E) When a datagram transmission request is issued to the terminal 411, the terminal 411 uses the ATM connection 41A to send an address resolution server (ARS) 413.
A to transfer the datagram to the target terminal
A TM address acquisition request (AR request) is performed. This address acquisition request is always made when a datagram transfer request occurs, or only when there is no address information of the destination terminal in the address resolution table stored in the terminal 411 (see FIG. 42). is there. In the latter method, if an address already exists in the AR table, there is no need to execute the AR request and AR response procedure.

[0427] The ARS 413 stores VCI / VPI information (the VC to be added by the terminal 411) which is the identifier of the ATM connection for transferring the datagram to the destination terminal 412.
I / VPI information) to the terminal 411 using the ATM connection 41B (referred to as an AR response). The terminal 411 that has received the AR response adds the notified VCI / VPI and inputs the datagram to the network. The datagram is sent to the terminal 4 through the ATM connection 41C.
12 is delivered directly. In the case of this embodiment, it is necessary that an ATM connection be set between all terminals in the sub-network in a full mesh state.

FIG. 42 shows a transmitting terminal (terminal 4 in this example).
11 is a flowchart illustrating an example of a protocol executed in 11). This is an example where the transmitting terminal can cache the address information of the destination. In step 432, a cache table for address resolution in the terminal is searched. If there is no information of the destination terminal 412 in the cache entry, the AR response
This is performed for RS413.

FIG. 41 shows terminals 411, 412 and AR
It shows the exchange of data between S413. this is,
This is an example in the case where the address resolution cache entry of the terminal 411 has no information of the destination terminal 412. The AR request 41A and the AR response 41B can be realized by a point-to-point ATM connection, but can also be realized by using a broadcast channel.

FIG. 43 shows another embodiment. Causes CLSF 414 to perform the actual datagram transfer. When a datagram transfer request occurs at the terminal 411, the terminal 41
1 performs resolution of the ATM connection information for transferring the datagram to the target terminal 412. If the terminal 411 does not have ATM connection information for transferring the datagram to the destination terminal 412, that is, if there is no entry in the cache table, the ATM
Address resolution is performed using the connection 41A (AR request).

[0431] When the destination terminal is a terminal in its own subnetwork, the ARS 413
A reply to the terminal 411 is ATM layer address information for transferring the datagram to the terminal 414 (AR response). In the case where the terminal to which the datagram is to be transferred belongs to a network other than the own subnetwork,
This will be explained in the following subsections.

The terminal 411 that has completed the address resolution adds the appropriate VCI / VPI (cache table or VCI / VPI information obtained by the AR response) to the cell for transferring the datagram, and adds it to the network. throw into. The VCI / VPI is an identifier of the ATM connection 441, and the cell is transferred on the ATM connection 441 and reaches the CLSF 414. C
The LSF 414 analyzes the address information of the received datagram, adds a VCI / VPI for transferring the datagram to the destination terminal 412, and inputs the cell to the network. The input cell is the ATM connection 442
To reach the terminal 412. In the case of the present embodiment, a star-shaped ATM connection needs to be set between the CSLF 414 and each terminal. Note that the AR request and the AR response can be performed using a broadcast channel instead of a point-to-point ATM connection.

The address resolution procedure performed by the ARS 413 may be a search on network address space information (address mask). As shown below, the ARS 413 may analyze the external network to determine in which network address space the terminal is located.
There is no need to resolve VCI / VPI information for accessing by connection.

FIG. 44 shows a procedure for data transfer. Here, regarding the transfer of the datagram from the terminal 411 to the CLSF 414, once the entire datagram is once taken into the CLSF 414, the data
12, the transfer of datagrams can be performed in a pipelined manner.

(Embodiment 4-2) A case in which VPI routing is applied to datagram communication inside a subnetwork.

As shown in FIG. 45, it is assumed that a VPI is assigned to each network element. Each terminal /
The server is equivalent to setting up a multipoint-to-point ATM connection from every UNI point in the subnetwork. That is, the terminal is the V of the destination terminal.
When a cell with a PI is entered into the network, the cell is transferred to the target UNI point. For example, ARS4
In order to transfer a cell to the V.13, VPI ₄₁₃ may be added as VPI.

The method described with reference to FIG.
6A, 46B and 46C are used. The terminal 411 is a VPI
An AR request cell having ₄₁₃ as VPI information is transferred to the ARS ₄₁₃ . At this time, the ARS 413 can recognize that the received cell is a cell transmitted from the terminal 411 using the VCI information or the identifier of the upper layer. For example, if VPI _411, which is the VPI information of the terminal 411, is written in the VCI field,
The cell received by referring to the header is
You can recognize that it is from. The VCI information can also be an identification number that explicitly indicates that the cell is an AR request cell for datagram communication.

The ARS 413 that has performed the address resolution of the terminal 412 stores VPI = VPI in the AR response cell.
₄₁₁ is written and the terminal 411 enters the network. Similarly to the ARS 413, the terminal 411 recognizes that the received cell is an AR response using at least one of the VCI information and the header information of the upper layer. In the AR response cell, at least VPI ₄₁₂ which is access address information of the terminal ₄₁₂ is written.

The identification information for identifying that the cell for datagram transfer received by terminal 412 has been transferred from terminal 411 is VCI information or header information of an upper layer. ARS413
Can be notified to the terminal 411 as information written in the AR response. The easiest way to identify the source and the cell type by the receiver is to use the above-mentioned sequence of procedures, and as the information of the VCI field, 8 bits are the source VPI information and the last 8 bits are the cell type. A method of indicating is possible.

Terminal 411 V of cell to be transferred to terminal 412
As a coding method of the CI field (16 bits), there is the following method.

Example 1: 8 bits are the access address of the own terminal (the same number as the VPI number), and the remaining 8 bits are an identification number indicating connectionless communication. Example 2: 8 bits are the access address of the own terminal. The remaining 8 bits are your network identification number. However, at this time, each terminal needs to acquire a VPI for connectionless communication separately from a VPI for connection-oriented communication (a different VPI is used for CL and CO).

Example 3: 16 bits are set when the terminal is booted. It is also possible for the receiving terminal to make an appropriate assignment.

The method described with reference to FIG.
6A, 46B, 46D and 46E are used. Terminal 411
Specifies an AR request cell having VPI ₄₁₃ as VPI information.
It transfers to S413. The ARS 413 that has performed the address resolution of the terminal 412 stores the VPI in the AR response cell.
= Write VPI ₄₁₄ and put the cell (AR response) into the network. Some AR response cells have CLSF4
At least VPI _414, which is the access address information of No. 14, is written. The information for identifying that the cell for datagram transfer received by the CLSF 414 has been transferred from the terminal 411 is VCI information or header information of an upper layer.
The RS 413 can also notify the terminal 411 as information written in the AR response. CLS 414 analyzes the address information of the received cell and transmits VPI = VPI ₄₁₂ to transfer the datagram cell to terminal _412.
Is added to the cell and the cell is input to the network. The cell is transferred to the terminal 412 through the ATM connection 46E.

[0444] Five specific examples of this embodiment are shown below.

(1) The CLSF 414 resolves the access address of the receiving terminal. The CLSF 414 analyzes the address of the destination terminal based on the network layer address written in the received datagram (or an address of another layer is also possible). At this time, the datagram is once reassembled by the CLSF 414.

(2) The CLSF 414 performs resolution of the access address of the receiving terminal. The CLSF 414 analyzes the address of the destination terminal based on the network layer address written in the received datagram (or an address of another layer is also possible). At this time, the datagram is not once reassembled by the CLSF 414, and the cells are relayed by pipeline processing. That is, the address information written in the first cell of the datagram is analyzed to analyze the access address of the destination terminal, and cells subsequent to the first cell are rewritten with VPI / VCI based on the VCI information and transmitted to the destination terminal. Relaying. The CLSF 414 needs to assign a different VCI for each datagram.

(3) The transmitting terminal resolves the access address of the destination terminal and writes this in the payload of the first cell. Upon receiving the first cell, the CLSF 414 reads the information of the payload section of the first cell and uses this to transfer the cell to the destination terminal. At this time, the datagram is once reassembled by the CLSF 414.

(4) The transmitting terminal resolves the access address of the destination terminal and writes this in the payload of the first cell. Upon receiving the head cell, the CLSF 414 reads the information of the payload part of the head cell and uses this to transfer the cell to the destination terminal. At this time, the datagram is not once reassembled by the CLSF 414, and the cells are relayed by pipeline processing. That is, by analyzing the address information written in the first cell of the datagram, the access address of the destination terminal is analyzed,
The cells after the first cell are VPI / VPI based on the VCI information.
The VCI is rewritten and relayed to the destination terminal. CL
The SF 414 needs to assign a different VCI for each datagram.

(5) The transmitting terminal resolves the access address of the destination terminal, and transfers this to the CLSF 414 using the VCI field. That is, for example, VC
The 8 bits of I are the access address of the destination terminal, and the remaining 8 bits are the access address of the source terminal. CLSF
414 relays the cell to the destination terminal by copying the address of the destination terminal in the VCI field. Relaying includes a method of once reassembling a datagram and a method of relaying cells in a pipeline.

(Fifth Embodiment) Next, as an embodiment relating to datagram transfer to an external network, a case where a two-layer network is used will be described first.

(Embodiment 5-1) Case of Using General ATM Network—Part 1

FIG. 46 is a schematic diagram of the network architecture of the present embodiment. This network is 2
It consists of a hierarchical network. Each of the networks 471 to 475 is internetworked via an IWU (Internetworking Unit) 476 to 479. Network 471 and public network 475
Are connected via the IWU 479. IWU47
6 to 479 can realize the relaying of the ATM cell without terminating the ATM connection. That is, it has a function of converting the VCI / VPI of the received cell into the VCI / VPI assigned to the corresponding ATM connection in the adjacent network.

FIGS. 47 and 51 show the setting of the ATM connection related to the address resolution.
Each of the ARSs 481 to 484 manages at least address information of terminals or networks accommodated in the networks 471 to 474 to which each ARS belongs. FIG.
8, from the network 472 to another network 47
ATM connections 496, 497, 49 required to transfer datagrams to 1,473,474,475
8 and 49B are shown. IWU476 to CLSF4
ATM connections (one-way ATM connections) to 91, 493, and 494 are set. An ATM connection 49 to the CLSF in the public network 475 is also provided.
B (bidirectional ATM connection is normally defined) is set. Note that the same connection is set for other sub-networks. The location where the CLSF and ARS are located can also be located at the location of the IWU. Further, it is possible that the CLSF and the ARS exist at the same position.

A cell (connectionless communication cell) to be transferred from the public network 475 to a terminal existing in the network being discussed here is temporarily stored in the network 471.
Connectionless communication (ATM)
Server that terminates the connection). Public network 4
75, this server is defined as an access point for connectionless communication. This server terminates the ATM connection and forwards the datagram to the destination terminal. The method of transferring datagrams from the server is realized in the same manner as the transfer of datagrams from each terminal to another terminal.

The protocol of the terminal in this embodiment, that is, the procedure for the terminal to transmit a connectionless communication datagram to the target terminal is as follows.

(1) The terminal issues an address resolution request (AR request). This is performed when the terminal cannot resolve the address by its own capability, for example, when there is no entry in the address resolution cache table, or always.

(2) The terminal obtains, from the address resolution server, VCI / VPI information which is an identifier of an ATM connection prepared for accessing the corresponding terminal.

(3) The terminal attaches the obtained VCI / VPI and inserts a cell into the network, thereby completing the datagram transfer.

Note that the terminal does not need to perform the connection setting procedure defined in the ATM network when transferring the datagram.

[0460] There are two types of protocols for the address resolution server in the present embodiment, "backbone ARS-led" and "front-end ARS-manual".
Each is as follows.

(I) Backbone ARS Initiative ARS 482 existing in each of the networks 472 to 474
-484 and ARS481 existing in the network 471
A star-like ATM connection (two-way communication channel) is formed between the two. For example, the terminal 47A is the terminal 47
To obtain the VCI / VPI to transfer the datagram to D, the terminal 47A transfers the address resolution request cell having the address information of the terminal 47D to the ARS 482, which is the ARS of its own network 472. The ARS 482 that has received the request cell analyzes that the address of the target terminal written in the received cell is not the own network 472, and
AT already set to ARS481 existing in
The relaying of the address resolution request cell is performed using the M connection 485. Alternatively, A
It is also possible for the RS 482 to cache in advance the network address information of the external network that has been acquired in the past through the ATM connection 485 and the VCI / VPI information for transferring the cell to each corresponding CLSF.

The ARS 481 uses the VCI / VP for transferring the datagram to the CLSF (existing in the network 474) to which the datagram is to be transferred from the address information of the destination terminal 47D held by the received cell.
It analyzes the I information (VCI / VPI information will be described later) and transfers the VCI / VPI information to the ARS 482. The ARS 482 receiving the VCI / VPI information,
Based on the VCI / VPI information, V
The CI / VPI information is returned (AR response).

The resolution of the address in the ARS 481 is performed as follows. The ARS 481 has the address information (address space information) of the terminal accommodated in the own network 471 and the information of the address space of the sub-networks 472 to 474. The ARS 481 compares the address written in the received address resolution request cell with the address space information of each subnetwork, and analyzes the corresponding destination network. Here, it should be noted that the ARS does not necessarily analyze up to the host address of the destination address, but only analyzes up to the network address when performing the comparison.

Note that there are the following two methods for identifying a datagram destined for the public network 475.

(A) Indicates whether the address information written in the address resolution cell is a datagram intended for the public network 475 or a datagram not intended for the public network. In other words, the terminal knows whether it is for the public network or not at the time of the address resolution request, and the terminal resolves the address resolution in a way that can explicitly identify whether the ARS is for the public or not. Transfer the cell to the ARS. If the address information is not an address for the public network 475, the ARS
When the address does not exist in the address entry 481, the information that the address does not exist is indicated by the ARS.
482.

(B) When the address of the received address resolution request cell does not exist in the address entry of the ARS 481, the address is set to the public network 4
75.

As described above, the ARS 481 is a terminal and sub-network 47 belonging to the network 471.
It has 2,473,474 addresses and address space information and performs address resolution.

FIGS. 49 and 50 show views of the sub-network space in the present embodiment from the viewpoint of ARS. FIG. 49 is an address space view of an adjacent network as viewed from the ARS 481. FIG. 50 is an address space view of an adjacent network viewed from the ARS 482. From the ARS 482, the network 511 connected by the IWU 476 is not initially visible as a plurality of sub-networks. From ARS482, AR
Only after exchanging information with S481, the address space of each sub-network is resolved.

The address information (ATM layer address information) that the ARS 482 receives from the ARS 481 is the IWU
A from 476 to CLSF existing in the target subnet
This is the information of the identifier of the TM connection (VCI / VPI information, and in some cases, also includes identification information of an upper layer). For example, from terminal 47A to terminal 47D
When transferring the datagram to the terminal 47F, the identifier of the ATM connection 495, 497 for accessing the CLSF 494 is notified (AR response). When transferring the datagram to the terminal 47F, the ATM connection 495, 496 for accessing the CLSF491 is notified. Of the identifier. The ARS 482 determines that the ATM connection is IW based on the VCI / VPI information received from the ARS 481.
VCI / V as relayed normally in U476
The PI information is notified to the terminal 47A. The VCI / VPI number notified to terminal 47A is rewritten by IWU 476 to another VCI / VPI.

(Ii) Front-end ARS Initiative ARS 482 existing in each of the networks 472 to 474
-484 and ARS481 existing in the network 471
Between them, a star-shaped ATM connection shown in FIG. 47 or a mesh-shaped ATM connection shown in FIG. 51 is formed. Each ARS acquires the address space information and the ATM connection information (VCI / VPI) of the external subnetwork as viewed from its own subnetwork, using the ATM connection defined in FIG. 47 or 51, respectively. FIG. 47 shows a configuration in which the ARS 481 is like a master ARS, and FIG. 51 shows a distributed configuration in which each ARS operates independently. If the network does not have three or more layers, it is more appropriate to use the backbone network as the master as shown in FIG.

On the other hand, when there is no limitation on the hierarchy of the network, which of the forms shown in FIGS. 47 and 51 is selected depends on the form or management form of the network and the number of sub-networks existing in the network. . For example, the terminal 47A may acquire the VCI / VPI to transfer the datagram to the terminal 47D.
A is ARS4 of ARS of own network 472
The address resolution request cell having the address information of the terminal 47D is transferred to 82. The ARS 482 that has received the request cell analyzes that the address of the destination terminal written in the received cell is the network 474, and sends the datagram to the CLSF (existing in the network 474) to which the datagram is to be forwarded. Is transferred (AR response) to the terminal 47A for transferring the VCI / VPI information (the VCI / VPI information will be described later).

The resolution of an address in the ARS 482 is performed as follows. The ARS 482 includes the address information (address space information) of the terminal accommodated in the own network 472 and the sub-networks 471 and 47.
3,474 address space information. ARS
Reference numeral 482 compares the address written in the received address resolution request cell with the address space information of each subnetwork, and analyzes the corresponding transfer destination network. Here, it should be noted that the ARS does not necessarily analyze up to the host address of the destination address, but only analyzes up to the network address when performing the comparison.

[0473] There are the following two methods for identifying a datagram intended for the public network 475.

(A) Indicates whether the address information written in the address resolution cell is a datagram intended for the public network 475 or a datagram not intended for the public network. In other words, the terminal knows whether it is for the public network or not at the time of the address resolution request, and the terminal resolves the address resolution in a way that can explicitly identify whether the ARS is for the public or not. Transfer the cell to the ARS. When the address information is not the address for the public network 475, the ARS4
When the address does not exist in the address entry 82, it is determined that the address does not exist.

(B) If the address of the address resolution request cell received in the address entry of the ARS 482 does not exist, the address is set to the public network 4
75.

As described above, the ARS 482 is composed of the terminals and the sub-networks 47 belonging to the network 472.
It has address information of 1,473,474 and address space, and performs address resolution.

FIGS. 52 and 53 show views of the sub-network space in the present embodiment from the viewpoint of ARS. FIG. 52 is an address space view of the neighboring network as viewed from the ARS 482, and FIG.
It is an address space view of an adjacent network viewed from S481. From each ARS, the address space of all sub-networks is resolved.

The address information (ATM layer) received by the ARS 482 from another ARS is the identifier (VCI / VPI information) of the ATM connection from the IWU 476 to the CLSF existing in the target subnet, and includes the identification information of the upper layer. Information). For example,
At the time of datagram transfer from the terminal 47A to the terminal 47D, the identifier of the ATM connection 495, 497 for accessing the CLSF 494 is notified (AR response), and at the time of datagram transfer to the terminal 47F, C is transmitted.
The identifier of the ATM connection 495, 496 for accessing the LSF 491 is notified. ARS48
2 notifies the terminal 47A of VCI / VPI information from the VCI / VPI information received from another ARS such that the ATM connection is normally relayed by the IWU 476. The VCI / VPI number notified to terminal 47A is rewritten by IWU 476 to another VCI / VPI.

[0479] A routing protocol for datagram communication (connectionless communication) between the sub-networks as well as an exchange protocol for address space information of each sub-network operates between the ARSs. Specifically, A between the IWU and the CLSF as shown in FIG.
Performs TM connection setting management. In addition, individual AT
The M connection is separated by the IWU (closed inside the sub-network), and another ATM connection server process and routing server process
It performs M connection path control and ATM connection management (for example, VCI / VPI management).
The S exchanges control messages with these servers and IWUs, and manages ATM connections required for connectionless communication.

FIG. 54 shows the VCI that IWU 476 should have.
An example of the / VPI rewrite table has been shown. The VC added to transfer cells to the same CLSF in the figure
The I / VPI need not necessarily be different numbers. That is, by combining with not only the VCI / VPI information but also the identifier information in the upper layer data unit, it is possible to identify which datagram the cell received by the destination CLSF belongs to. Similarly, the VCI / VPI added to the cell transferred from the same terminal need not always have a different number.

Also, there is a case where it is necessary to use an ATM connection having a route different from the ATM connection used in the normal state due to congestion or failure in the sub-network. In this case, a routing control process for managing and controlling the VCI / VPI and the routing table information set in the table of each switch.
(It is possible that the control process that performs address management can be another process.) The VCI / VPI number shown at the UNI point (user interface point) can be maintained even when the route in the sub-network changes. Then, management control is performed.

Next, when the access points of various servers (for example, CLSF) move, the connection identification numbers (VCI /
VPI) may be rebooted so as to be the same. In addition, a request message (new access information if necessary) for discarding information for old access from the cache table is transferred to the relevant access point upon reboot. In the latter case, in IWU, FIG.
It is necessary to change the entry information of the table as shown in FIG. 4 (using the information in the transmitted message,
Or execute a protocol to acquire new information).

Next, the routing to the destination terminal in this embodiment will be described. The final datagram delivery to each terminal is performed only for the network to which each CLSF belongs (see FIG. 48). That is, for example, CL
The SF 491 distributes datagrams to terminals belonging to the network 471 (the networks 472, 473, 474, and 475 do not provide services). Similarly, CL
The SF 494 distributes datagrams only to terminals in the network 494. If the network address of the datagram received by each CLSF does not exist in the address entry of the CLSF, or if the address of the received datagram is not an element of the network address space of the network, the datagram is incorrect. Is determined to have been delivered. The actions for erroneous datagram delivery are not discussed here. That is, each CLSF only needs to have the address information of the terminal of the network to which it belongs. When the address of the received datagram exists in the own network, an appropriate ATM connection is selected and the datagram is relayed.

FIG. 55 shows an example of protocol processing when a datagram is transferred from the terminal 47A to the terminal 47D. The ATM connection is temporarily terminated by CLSF494. That is, the OSI layer 3 in CLSF494
Protocol is terminated. Layer 3 with CLSF494
Is performed, and the data unit is transferred to the terminal 47D using the ATM connection. As described above, when a datagram is transferred to a terminal other than the own subnetwork, datagram delivery is realized end-to-end at only one end of the ATM connection.

(Embodiment 5-1-1) Next, a more specific embodiment will be described.

In FIG. 56 and FIG. 57, the terminal 47A
An embodiment in which a datagram is transferred from the terminal 47 to the terminal 47D will be briefly described. FIG. 56 shows FIGS. 50 and 49
FIG. 57 is an embodiment based on the management view shown in FIG.
Is an embodiment based on the management views shown in FIG. 52 and FIG.

The address resolution in the present embodiment is performed as follows.

If the terminal 47A does not have the ATM layer address information for transferring the datagram to the terminal 47D when transferring the datagram to the terminal 47D, the terminal 47A has the address information of the terminal 47D to the ARS 482. The address resolution request cell is transferred using the ATM connections 571 and 581. However, the ATM connections 571 and 581 can be realized by a point-to-point connection or a broadcast connection. The ARS 482 that has received the address resolution request cell transmits the address resolution request cell to the ARS 4 via the ATM connection 572 in the example of FIG.
Transfer to 81. The following method can be considered as a procedure from when the ARS 482 receives the address resolution request cell to when the ARS 481 transfers the request cell to the ARS 481.

(1) When the ARS is responsible for all the responses to the address resolution request cells generated in the network, the ARS first screens the address resolution request cells. That is, it verifies whether the address in the received cell is an element of the address space of the own network using a network mask or the like. If the address is in the own network, the address table is searched and the ATM layer address information of the corresponding terminal is returned. On the other hand, if the address is other than that of the own network, the request cell is transferred to a preset upper ARS, in this example, ARS481.

(2) The ARS responds only to a request to send a datagram to a terminal that is not in the network, ie, belongs to another subnetwork, and sends a resolution request to a terminal in its own network. When the terminal to be transferred replies, the response differs depending on whether a broadcast channel is used for address resolution or a point-to-point connection. When a broadcast channel is used, after screening an address, a request cell is fetched, and the request cell is transferred to an upper ARS. On the other hand, when a point-to-point connection is used, in principle, the upper AR
S. In this example, the cell can be transferred to the ARS 481.

In the example of FIG. 56, the ARS 481 that has received the address resolution request cell checks whether the address accommodated in the own network and the address of the request cell exist in the address space. When there is no address corresponding to the address entry,
Judge as an address for the public network. In the example, since the destination is the terminal 47D, the ARS 471
That the address of the requested cell exists in the address space of the network (ie, the network 474).
Address space information that can be used). The ARS 481 that has analyzed the address of the terminal 47D is IWU4
VC of ATM connection from 76 to CLSF494
I / VPI information (VCI / I
(VPI information) to the ARS 482.

The ARS 482 uses the CLSF
VCI, which is the identification information of the ATM connection to be relayed to the ATM connection 494
/ VPI. By using the VCI / VPI, the terminal 47A can directly deliver a datagram to the CLSF494. In the example of FIG. 57 in which the ARS 482 can autonomously perform address resolution, in response to an address resolution request given through the ATM connection 581, the resolution information (AR response) is directly sent to the terminal 47A. Through the ATM connection 582.

The datagram transfer in this embodiment is performed as follows.

The terminal 47A is connected to the ATM connection 57
The cell (having datagram information) is transferred to the CLS 494 by using 5,583. The IWUs 476 and 478 relay the ATM connection by rewriting the received VCI / VPI information of the cell. Since this processing is performed as processing of the ATM layer without being lifted to the upper layer, the ATM passing through the IWU 476,478 is used.
Connections 575 and 583 can be viewed as one ATM connection without an ATM termination point.

At the CLSF 494 which is the terminal point of the ATM connections 575 and 583, the processing of the network layer is performed. The CLSF 494 analyzes the network address and sends the datagram to the ATM connection 57.
6, 584, and transfer it to the terminal 47D.

As described above, the ATM connection is connected to the terminal 4
7A to CLSF494 and CLSF494 to terminal 4
7D, which is terminated only once.

(Embodiment 5-2) In the case where an ATM network of VPI routing is used—No.

First, a description will be given of an address assignment method of the ATM layer in the present embodiment. When performing VPI routing, the VPI field indicates the destination UNI.
(Described as VPI-F). The coding method of the VCI field needs to be defined as follows. Here, the coding of the VCI field under discussion relates to the transfer of cells related to connectionless communication, and the coding here is not necessarily applied to other applications (such as connection-oriented connections). No need to use. That is, generally, the coding method of the VCI field can be performed by negotiation between the receiving terminal and the transmitting terminal. The 16 bits of the VCI field are defined to be composed of two 8-bit subfields, and these are defined as VCI-F1, VCI
Described as -F2.

In this embodiment, the ATM connection is set as follows. The setting of the ATM connection related to the address resolution shown in FIGS. Each ARS 481-484 has at least each A
It manages address information of terminals or networks accommodated by the networks 471 to 474 to which the RS belongs.

FIG. 58 shows a case where the ARS 481 shown in FIG. 48 takes the initiative (becomes a route) to manage the address information of each sub-network. VCI / VPI
The information is rewritten by the IWU and relaying of the ATM connection is performed. For example, ARS482 to AR
The ATM connection to S481 is the connection 59
1,592. V of connection 591
PI-F is VP which is the access address of IWU 476
An I _476-2, VCI-F1 is a VPI ₄₈₂ which is the access address of ARS482. VCI-F2 is
In principle, it can be any value, but IWU
476 is coded to be able to identify from the information inside the VCI-F2 that the received cell is to be relayed to connection 592.
On the other hand, the VPI-F of the connection 592 is ARS481.
VPI ₄₈₁ , which is the access address of
F1 is VPI which is an access address of IWU 476
_476-1 . Note that VCI-F
2 can be any value in principle. VPI _476-1 is an IWU for the network 471.
The access address (VPI) 476 and the VPI _476-2 are the access address (VPI) of the _IWU 476 for the network 472.

[0501] In the IWU 476, the VCI of the received cell is
From the set of information of F2 and VCI-F1, connection 5
The VCI / VPI information to be written into the cell 92 is analyzed. Therefore, the IWU 476 has a VCI field (16 bits) as a table entry, and as a result,
It has a function of analyzing CI / VPI information. The VPI-F of the connection 592 cell is VCI-F1 and VCI-F1.
It is analyzed by the combination of F2. VCI-F1 has V
PI _476-1 is written. Also, VCI-F2 is A
It is coded as an identifier of the ATM connection between the RS 482 and the ARS 481. The value of VCI-F2 is the value of the ATM connection (connections 591, 59
It is determined when setting 2). The assignment of VCI-F2 is based on the subnetwork (networks 471 and 47).
2) A process that manages the VCI-F2 within the terminal may be used, and the terminal (the ARS 481 and the IWU 47)
The process managing the value of VCI-F2 in 6) may be used.

FIGS. 59 and 60 will be described. This is a case where each ARS shown in FIG. 51 independently manages address information of each subnetwork. For example, ARS
ATM connection 617 from 482 to ARS 484
Is formed from connections 6021, 6041, and 6061. The VPI-F of the connection 6021 is IW
VPI 476-2 which is an access address of U476, and VCI-F1 is VPI ₄₈₂ which is an access address of ARS482. VCI-F2 is VCI-F, which is an identification number assigned to connection 6021.
1 and VCI-F2, the connection is identified. The VPI-F of the connection 6061 is IWU47
VPI _478-1 which is an access address of the
I-F1 is VP which is an access address of IWU 476
I _476-1 . VCI-F2 has a connection 606
This is a value set when 1 is set. Connection 604
The VPI-F of No. 1 is a VPI ₄₈₄ which is an access address of the ARS ₄₈₄ , and the VCI-F1 is a VPI _478-1 which is an access address of the _{IWU 478} . VCI-F
2 is a value set when the connection 6041 is set.

[0503] For example, the IWU 478 analyzes VCI / VPI information to be written to the cell of the connection 6041 from a set of the received VCI-F2 and VCI-F1 information of the cell. The IWU 478 has a VCI field (16 bits) as a table entry, and has a function of analyzing VCI / VPI information as a result.

[0504] VPI-F of the cell of the connection 6041
Is analyzed by a combination of VCI-F1 and VCI-F2. VPI _478-1 is written to VCI-F1. Also, VCI-F2 is ARS482 to ARS48.
1 (coded from VCI field information). VCI
The value of -F2 is the value of the ATM connection (connection 60
21, 6041, 6061). V
The assignment of the CI-F2 is based on the VCI-F within the subnetwork (networks 471, 472, 474).
2 or a terminal (ARS 481, IWU 476, ARS 484),
A process that manages the value of I-F2 may be used.

According to the method of setting the VCI / VPI field as described above, the information over a plurality of cells is stored in the ARS.
Can be reassembled without any problem on the receiving side even when is transferred to another ARS. It is necessary to use an upper layer identification field for data identification such as from which terminal the data is transferred or data related to an address resolution protocol.

FIG. 61 shows the setting of the ATM connection required to transfer a datagram from the terminal 47A to the terminal 47F, the public network 475, and the terminal 47D. The two methods are described below.

(1) First, the transfer of a datagram from the terminal 47A to the terminal 47F will be described. In this case, two ATs
It is formed from M connections. One is the connection 621 to 623 and 624 from the terminal 47A to the CLSF 491, and the other is the connection from the CLSF 491 to the terminal 47F.
Connection 625 to the server. Terminal 47A is an IWU
The following cell is transferred to 476. That is, VPI
-F is VPI which is an access address of IWU 476
_476-2 , VCI-F1 is VPI _47A which is an access address of terminal _47A , and VCI-F2 is connection 621
Each has the value set at the time of setting. IWU476
The VCI of the cell received at is analyzed and the corresponding VPI
-F and VCI-F2 are analyzed. VCI-F1
Is written with VPI _476-1 which is an access address for the network 471 of the IWU 476. The VCI-F2 is a value determined when the connection 624 is set, and the value of the VCI-F1 of the cell received by the IWU 476, that is, the VPI _47A that is the access address of the terminal 47A can be copied and written. is there.

The datagram arriving at the CLSF 491 is once subjected to cell reassembly, and the ATM connection is terminated. The CLSF 491 analyzes the address information of the upper layer and recognizes that the destination of the datagram is the terminal 47F. Therefore, the CLSF 491 generates the following cell and transfers the datagram to the terminal 47F. V
PI-F is VPI which is an access address of the terminal 47F.
_47F and VCI-F1 are VPI ₄₉₁ and VCI-F2 which are access addresses of CLSF491, respectively.
5 is an identification number assigned to the device. The terminal 47F
Can uniquely identify the datagram to which the cell belongs from VCI-F1 and VCI-F2. That is,
Datagram reassembly is possible.

Next, transfer of a datagram from the terminal 47A to the public network 475 will be described. In this case, one ATM
Connection (connection 622 + 626 + 627)
Formed from The terminal 47A transfers the following cell to the IWU 476. That is, VPI-F is VPI
_476-2 , the VCI-F1 has a value set at the time of setting the VPI _47A , and the VCI-F2 has a value set at the time of setting the connection 622. The VCI of the cell received by the IWU 476 is analyzed, and the corresponding VPI-F and VCI-F2 are analyzed. VCI-F1 is a network 47 of IWU 476.
VPI _476-1, which is an access address for No. 1, is written. VCI-F2 is connected to connection 626.
Is set when is set, and the value of the VCI-F1 of the cell received by the IWU 476, that is, the terminal 47A
It is also possible to copy and write the VPI _47A which is the access address of the VPI. The IWU 479 determines the V of the received cell.
From the CI information, the VCI / VCI assigned to the ATM connection 627 defined in the public network 475
Is written, and the cell is transferred to the public network 475.

[0510] Finally, the transfer of a datagram from the terminals 47A to 47D will be described. It is formed from two ATM connections. One is connection 623, 628, 629 from terminal 47A to CLSF494, and the other is C
A connection 62A from the LSF 494 to the terminal 47D. The terminal 47A transfers the following cell to the IWU 476. That is, VPI-F is VPI _476-2 which is an access address of _{IWU 476} , and VCI-F1 is terminal 4
VPI _47A , VCI-F which is an access address of 7A
2 has a value set when the connection 623 is set. The VCI of the cell received by the IWU 476 is analyzed, and the corresponding VPI-F and VCI-F2 are analyzed. VCI-F1 is an access address for the network 471 of the IWU 476, which is a VPI _476-1.
Is written. Note that VCI-F2 is a value determined when the connection 628 is set, and the IWU4
It is also possible to copy and write the value of the VCI-F1 of the cell received by 76, that is, the VPI _47A which is the access address of the terminal 47A.

Next, the V of the cell received by IWU 478
The CI is analyzed, and the corresponding VPI-F, VCI-F1, and VCI-F2 are analyzed. VCI / VPI is C
The LSF 494 needs to be allocated so that it can be identified even if cells from all terminals in the network (terminals capable of forwarding datagrams to the CLSF 494) arrive at the same time.

[0512] The datagram arriving at the CLSF 494 is once subjected to cell reassembly, and the ATM connection is terminated. The CLSF 494 analyzes the address information of the upper layer and recognizes that the destination of the datagram is the terminal 47D. Therefore, the CLSF 494 generates the following cell and transfers the datagram to the terminal 47D. That is, VPI-F is VPI _47D which is an access address of the terminal 47D, VCI-F1 is VPI ₄₉₄ which is an access address of CLSF494, and VCI-F2 is an identification number assigned to the connection 62A. In addition,
The terminal 47D can uniquely identify the datagram to which the cell belongs from the VCI-F1 and the VCI-F2. That is, the datagram can be reassembled.

(2) Each CLSF has acquired a VPI address, and all connections identified by the VPI are ATM connections used for cell transfer for datagram communication (connectionless communication). I do. That is, the CLSF server itself needs an ATM connection (not a connectionless connection) in order to communicate with another terminal / server. That is, the CLSF needs to acquire at least two or more access addresses (VPI) at the time of boot.

Similarly, all IWUs also acquire at least two or more access addresses (VPIs) at boot time.
One VPI is used for a cell of connectionless communication. That is, when transferring a cell of a datagram relating to connectionless communication to an external network, each terminal attaches a VPI (for IWU) defined for connectionless communication to the cell and inserts the cell into the network.

[0515] At this time, the following method is used as a coding method of the VCI field of the cell for connectionless communication. The VCI field is divided into two 8-bit subfields (VCI-F1 and VCI-F1).
-F2: The field definition position is not mentioned).

[0516] Next, a description will be given of a coding method of a cell transferred in the external subnetwork. This is the coding of the VCI field when a cell is transferred from the subnetwork to which the terminal belongs to the external subnetwork via the IWU. The identification number of the subnetwork to which the source terminal belongs is written in VCI-F1,
In the VCI-F2, an identification number of the source terminal in the subnetwork is written. For example, each IWU in the network 471 is used as the identification number of the subnetwork.
Can be used as the identification address of the subnetwork (network 47).
1 itself is appropriately coded). Further, the terminal identification number may be an access address (VPI) of the terminal in each subnetwork. For example, the terminal 47D transferred in the external sub-network
_Are VCI-F1 = _VPI478-1 and VCI-F2 =
VPI _47D .

Next, a description will be given of a coding method of a cell transferred in the own sub-network. This is the coding of the VCI field when a cell is transferred to the IWU of the subnetwork to which the terminal belongs. VC
The identification number of the subnetwork to which the destination terminal belongs is written in I-F1, and the identification number of the transmission source terminal in the subnetwork is written in VCI-F2. For example,
The VCI-F2 field of the cell of the connectionless communication transmitted from the terminal 47D transferred in the own network can be VPI _47D .

[0518] With the above configuration, in the address resolution procedure performed prior to the transfer of the datagram, the terminal only needs to acquire the identification number of the sub-network to which the destination terminal belongs, and then the datagram (multiple (Or one cell). Less than,
The transfer of this datagram will be described.

First, the transfer of a datagram from the terminal 47A to the terminal 47F will be described. The terminal 47A is an IWU 47
Then, the following cell is transferred to 6. VPI-F is IWU4
VPI _476-2 , VCI-
F1 is an identification number of the network 471 (for example, CLSF
The use of VPI ₄₉₁₎ is a 491 access address of the possible), VCI-F2 is set to VPI _{47 A} which is the access address of the terminal 47A.

[0520] The VCI of the cell received by the IWU 476 is analyzed, and the corresponding VPI-F and VCI-F2 are analyzed. For example, in the case of the embodiment in which the access address VPI ₄₉₁ of the CLSF ₄₉₁ is written in the VCI-F1, the VPI-F of the cell to be output is obtained by copying the VPI ₄₉₁ that is the VCI-F1 information of the received cell. ,realizable. A VPI _476-1, which is an access address for the network 471 of the _{IWU 476,} is written in the VCI-F1. VCI-F2 can be set to be transparent. That is, for example, CL is added to the VCI-F1 of the cell transferred from the terminal 47A.
In the embodiment where the access address VPI _{49 1} of SF491 is written, (1) VCI-F receiving cell
1 is copied to the VPI-F of the transmission cell, (2) The VCI-F2 of the reception cell is copied to the VCI-F2 of the transmission cell,
(3) The relaying of the cell is realized by the procedure of writing the access address VPI _476-1 of the IWU into the VCI-F1 of the transmission cell.

The datagram arriving at the CLSF 491 undergoes cell reassembly and terminates the ATM connection. The CLSF 491 analyzes the address information of the upper layer and recognizes that the destination of the datagram is the terminal 47F. Therefore, the CLSF 491 generates a cell whose VPI-F is the VPI _47F which is the access address of the terminal 47F, and transfers the datagram to the terminal 47F. Note that CSLF491 is compatible with VCI-F1 and VCI-F
2, the datagram to which the cell belongs can be uniquely identified (the datagram can be reassembled). Therefore, relaying of datagrams (cells) can be performed in a pipeline manner.

Next, transfer of a datagram from the terminal 47A to the public network 475 will be described. Terminal 47A is IWU4
The following cell is transferred to 76. VPI-F is IWU
476 access address VPI _476-2 , VCI-
F1 is set to the identification number of the public network 475 (for example, VPI ₄₇₉ which is an access address of IWU ₄₇₉ can be used), and VCI-F2 is set to VPI _47A which is the access address of terminal 47A.

The IWU 76 analyzes the VCI of the cell received, and analyzes the corresponding VPI-F and VCI-F2. For example, in the case of the embodiment in which the access address VPI ₄₇₉ of the IWU ₄₇₉ is written in the VCI-F1, the VPI-F of the output cell is obtained by copying the VPI ₄₇₉ that is the VCI-F1 information of the received cell. ,realizable. VCI-F1 has an access address VW for the network 471 of the IWU 476.
PI _476-1 is written. VCI-F2 can be set to be transparent. That is, for example, the IWU is added to the VCI-F1 of the cell transferred from the terminal 47A.
In the case of the embodiment in which the access address VPI _{479 of 479} is written, (1) the VCI-F1 of the reception cell is copied to the VPI-F of the transmission cell, and (2) the VCI of the reception cell is copied.
The relaying of the cell is realized by a procedure of copying the I-F2 to the VCI-F2 of the transmission cell and (3) writing the VPI _476-1 which is the access address of the IWU to the VCI-F1 of the transmission cell. . From the received VCI information of the cell, the IWU 479 determines the VC assigned to the ATM connection 627 defined by the public network 475.
Write the I / VPI and transfer the cell to the public network 475.

[0524] Finally, the transfer of a datagram from the terminal 47A to the terminal 47D will be described. Terminal 47A is IWU 47
Then, the following cell is transferred to 6. VPI-F is IWU4
VPI _476-2 , VCI-
F1 is an identification number of the network 474 (for example, IWU4
The use of VPI _478-1 78 which is the access address of the possible), the VCI-F2 set to VPI _47A which is the access address of the terminal 47A.

[0525] The VCI of the cell received by IWU 476 is analyzed, and the corresponding VPI-F and VCI-F2 are analyzed. For example, in the case of the embodiment in which the access address VPI _{478-1 of the IWU 478} is written in the VCI-F1, the VPI-F of the output cell copies the VPI _478-1 which is the VCI-F1 information of the received cell. It can be realized by doing. VCI-F1 is a VP which is an access address for the network 471 of the IWU 476.
I _476-1 is written. VCI-F2 can be set to be transparent. That is, for example, the IWU4 is added to the VCI-F1 of the cell transferred from the terminal 47A.
In the embodiment in which the access address VPI _{478-1 of} 78 is written, (1) the VCI-F1 of the reception cell is copied to the VPI-F of the transmission cell, and (2) the VCI-F1 of the reception cell is copied.
Cell relaying is realized by a procedure of copying the I-F2 to the VCI-F2 of the transmission cell and (3) writing the VPI _476-1 , which is the access address of the IWU, to the VCI-F1 of the transmission cell.

Next, the V of the cell received by IWU 478
The CI is analyzed, and the corresponding VPI-F, VCI-F1, and VCI-F2 are analyzed. CL from IWU478
The VCI field of the cell transferred to the SF 494 can set the VCI field information of the received cell transparently. That is, VCI-F
1 = VPI _476-1 and VCI-F2 = VPI _47A . Note that VPI-F is set to VPI ₄₉₄ which is an access address of CLSF494. CLSF
The datagram arriving at 494 is once subjected to cell reassembly, and the ATM connection is terminated. CLSF
494 analyzes the address information of the upper layer and recognizes that the destination of the datagram is the terminal 47D. Also,
The CLSF 494 generates a cell whose VPI-F is the VPI _47D which is the terminal 47D access address, and transfers the datagram to the terminal 47D. The terminal 47D can uniquely identify the datagram to which the cell belongs from VCI-F1 and VCI-F2. That is,
Datagram reassembly is possible. Specifically, for example, the VCI-F1 identifies the subnetwork to which the source terminal belongs or the IWU accommodating the subnetwork as a VC.
It can be known from the I-F2 as an identification number in the subnetwork.

(Embodiment 5-3) Case where general ATM network is used—No.

FIG. 62 shows a schematic diagram of the network architecture of the present embodiment. This network is 2
It consists of a hierarchical network. Each network 631-635 is internetworked via IWUs 636-639. Network 631
And the public network 635 are connected via the IWU 639. The IWUs 636 to 639 can realize ATM cell relaying without terminating the ATM connection. That is, the VCI / VP of the received cell
It has a function of converting I to VCI / VPI assigned to the corresponding ATM connection in the adjacent network. Note that CLSF-O, CLSF-I and AR
S can also be implemented in the same device. Further, they can be implemented on the IWU.

FIGS. 64 and 65 show the setting of the ATM connection related to the address resolution.
Each ARS manages at least address information of terminals or networks accommodated in the network to which each ARS belongs.

In FIG. 63, from the terminal 63A to (1) the terminal 63
C, (2) Public network 635, (3) ATM connections 641-6 required to transfer datagram to terminal 63B
46 shows the settings. CLSF-O63M to CLSF-
An ATM connection (one-way ATM connection) to I63L, CLSF-I63J and the public network 635 is set. Note that the same connection is set for other sub-networks. The ATM connection set between CLSF-I and CLSF-O is shown in FIG. Note that, in the case of connectionless communication to terminals belonging to the networks 631 to 634 defined from the public network 635, an ATM connection related to connectionless communication from the public network 635 exists in the network 631 from the public network 635. Terminates ATM connection for connectionless communication,
Terminated by a server for relaying datagrams. For the server that terminates this connectionless ATM connection, the same procedure as described below for transferring a datagram from a terminal in a network to a terminal in another network is used to send a message to a target terminal.
Transferred from this server.

[0531] The terminal protocol in this embodiment will be described. When the terminal determines that the datagram is destined for the external subnetwork, the terminal forwards the datagram to CLSF-O. Note that the terminal and CLS
It is assumed that an ATM connection has already been set up with the FO. Each terminal has information on the address space of the sub-network to which the terminal belongs (address mask, etc.), and can determine whether the destination terminal is in its own sub-network or an external sub-network. .

In the case of the network configuration shown in FIG. 62, for example, the following three methods can be used as a datagram delivery system.

(1) CLSF-O and CLSF-I are at the same access point, and a star-shaped ATM connection is set between CLSF and the terminal. All terminals transfer cells to the CLSF when transferring datagrams. All datagram delivery is CLSF
Do. That is, communication between the terminals in the sub-network also passes through the CLSF once.

(2) Communication between terminals in the sub-network is realized without passing through the CLSF, and communication with terminals in the external network is realized via CLSF-O.

(3) CLSF-O is used for communication with terminals on the external network, and CLSF-I is used for communication with terminals inside the own network.

There are two types of protocols of the address resolution server in this embodiment, “backbone ARS-led” and “front-end ARS manual”.
Each is as follows.

(I) Initiative of backbone ARS ARS existing in each network and network 631
A star-shaped ATM connection (bidirectional communication channel) is formed between the ARS 63G and the ARS 63G (FIG. 6).
4). For example, in order to transfer a datagram from terminal 63A to terminal 63B, CLSF-O63M uses CLSF-O
ATM set between 63M and CLSF-I63J
To use a connection, the V
It is necessary to obtain PI / VCI information. CLSF-O
63M sends the terminal 6 to the ARS 63D, which is the ARS of its own network 632, in order to acquire the VCI / VPI.
An address resolution request cell having 3B address information is transferred (address information written in the received datagram). A that received the request cell
The RS 63D analyzes the address of the destination terminal written in the received cell, and when the address cannot be resolved with the information held by the ARS 63D, the ATM connection already set to the ARS 63G existing in the network 631 is established. The relaying of the address resolution request cell is performed using the H.651.

[0538] The ARS 63G uses the VCI / VP for transferring the datagram to the CLSF (existing in the network 633) to which the datagram should be transferred based on the address information of the destination terminal 63B held by the received cell.
It analyzes the I information (VCI / VPI information will be described later) and transfers the VCI / VPI information to the ARS 63D. The ARS 63D receiving the VCI / VPI information,
Based on the VCI / VPI information, the CLSF-O63M returns VCI / VPI information to be used (AR response).

The resolution of the address in the ARS 63G is performed as follows. The ARS 63G has the address information (address space information) of the terminal accommodated in the own network 631 and the address space information of the sub-networks 632 to 634. The ARS 63G compares the address written in the received address resolution request cell with the address space information of each subnetwork, and analyzes the corresponding transfer destination network.

There are the following two methods for identifying a datagram for the public network 635.

(A) Indicates whether the address information written in the address resolution cell is a datagram intended for the public network 635 or a datagram not intended for the public network. In other words, the terminal knows whether it is for the public network or not at the time of the address resolution request, and the terminal resolves the address resolution in a way that can explicitly identify whether the ARS is for the public or not. Transfer the cell to the ARS. When the address information is not the address for the public network 635, the ARS6
If the address does not exist in the 3G address entry, the information that the address does not exist is transmitted to ARS6.
Sent to 3D.

(B) When the address of the received address resolution request cell does not exist in the address entry of the ARS 63G, the address is stored in the public network 6.
35.

As described above, the ARS 63G includes the terminals belonging to the network 631 and the sub-network 632.
~ 634 addresses and address space information,
Perform address resolution. ARS63
G does not necessarily need to have address information up to the terminal level of a sub-network other than the sub-network to which it belongs, and may have information up to the network address level.

The address information (ATM layer) received by the ARS 63D from the ARS 63G is the identifier of the ATM connection from the IWU 636 to the CLSF-I present on the target subnet (VCI / VPI information, which may include upper layer identification information). Information). For example, when transferring a datagram from the terminal 63A to the terminal 63B, the A for accessing the CLSF-I63J is used.
The identifier of the TM connection 645 is notified (AR response), and when the datagram is transferred to the terminal 63C, C is transmitted.
The identifier of the ATM connection 642 for accessing the LSF-I63L is notified.

[0545] The ARS 63D determines from the VCI / VPI information received from the ARS 63G a VC that allows the ATM connection to be normally relayed by the IWU 636.
The I / VPI information is notified to CLSF-O63M. CL
The VCI / VPI number notified to SF-O63M is I
WU 636 rewrites to another VCI / VPI.

(Ii) Front-end ARS Initiative ARS existing in each network and network 631
The ATM connection in the form of a star as shown in FIG. 64 or the ATM connection in the form of a mesh as shown in FIG. Each ARS obtains the address space information and the ATM connection information (VCI / VPI) of the external subnetwork as viewed from its own subnetwork, using the ATM connection defined in FIG. 64 or 65, respectively. FIG. 64 shows that ARS63G is the master ARS
65 can be said to be a distributed type in which each ARS operates independently.

When the network does not have three or more layers, it is more appropriate to use the backbone network as a master as shown in FIG. On the other hand, when there is no limitation on the network hierarchy, which of the forms shown in FIGS. 64 and 65 is selected depends on the form or management form of the network and the number of sub-networks existing in the network. For example, CLSF-O63M transmits VCI / VPI to terminal 63B to transfer the datagram.
CLSF-O63M sends ARS63D to ARS63D, which is the ARS of its own network 632, to obtain ARS63.
The address resolution request cell having the address information of B is transferred. ARS6 which received this request cell
3D analyzes that the address of the destination terminal written in the received cell is the network 633,
The VCI / VPI information for transferring the datagram to the CLSF (existing in the network 633) to which the datagram is to be transferred (the VCI / VPI information will be described later) is transferred to the CLSF-O63M (AR response).

The resolution of the address in the ARS 63D is performed as follows. The ARS 63D stores the address information (address space information) of the terminal accommodated in the own network 632 and the sub-networks 631, 63.
3,634 address space information. ARS
63D compares the address written in the received address resolution request cell with the address space information of each sub-network, and analyzes the corresponding transfer destination network.

Note that there are the following two methods for identifying a datagram intended for the public network 635.

(A) The address information written in the address resolution cell explicitly indicates whether it is a datagram for the public network 635 or a datagram not for the public network. That is, the terminal knows whether or not the ARS is for the public network or not at the time of the address resolution request, and the terminal resolves the address resolution in a manner that can explicitly identify whether the ARS is for the public or not. Transfer the solution cell to the ARS. If the address information is not an address for the public network 635, the ARS
When the address does not exist in the address entry of 63D, it is determined that the address does not exist.

(B) If the address of the received address resolution request cell does not exist in the address entry of the ARS 63D, the address is set to the public network 6.
35.

As described above, the ARS 63D includes the terminals belonging to the network 632 and the sub-network 63.
It has information of 1,633,634 addresses and address space, and performs address resolution. In addition,
The ARS 63D does not necessarily need to have address information up to the terminal level of a subnetwork other than the subnetwork to which the ARS 63D belongs, but may have information up to the network address level.

The address information (ATM layer) received by the ARS 63D from another ARS is the identifier (VCI / VPI information of the ATM connection from the IWU 636 to the CLSF-I existing in the target subnet, including the identification information of the upper layer. Information). For example,
At the time of datagram transfer from the terminal 63A to the terminal 63B, an ATM for accessing the CLSF-I63J is used.
The ID of the connection 645 is notified (AR response), and when transferring the datagram to the terminal 63C, the CLSF-
ATM connection 64 for accessing I63L
2 is notified.

[0554] The ARS 63D uses the VCI / VPI information received from another ARS such that the ATM connection can be normally relayed by the IWU 636.
/ VPI information to the CLSF-O63M. CLS
The VCI / VPI number notified to the F-O63M is the IW
In U636, it is rewritten to another VCI / VPI. Between the ARSs, not only an exchange protocol for address space information of each sub-network but also a routing protocol for datagram communication (connectionless communication) between the sub-networks operates. In particular,
The setting management of the ATM connection as shown in FIG. 63 is performed. Note that individual ATM connections are separated by the IWU (closed inside the sub-network), and another ATM connection server process and a routing server process control ATM connection routing and ATM connection management (for example, VCI / VPI).
ARS), and ARS uses these servers and I
Exchanges control messages with the WU and manages ATM connections required for connectionless communication.

[0555] Routing to the destination terminal in this embodiment will be described. The final datagram delivery to each terminal is performed only for the network to which each CLSF-I belongs. That is, for example, the CLSF-I 63L is a terminal belonging to the network 631 (network 6
32 to 635 are not serviced). Similarly, the CLSF-I63J distributes datagrams only to terminals in the network 633. If the network address of the datagram received by each CLSF does not exist in the address entry of the CLSF, or if the address of the received datagram is not an element of the network address space of the network, the datagram is incorrect. Is determined to have been delivered. The actions for erroneous datagram delivery are not discussed here. That is, each CL
The SF only needs to have the address information of the terminal of the network to which the SF belongs. When the address of the received datagram exists in the own network, an appropriate ATM
Select a connection and relay the datagram.

FIG. 67 shows an example of the protocol processing when a datagram is transferred from the terminal 63A to the terminal 63B. The ATM connection is terminated by CLSF-O63M and CLSF-I63J. That is, CLSF
OSI layer 3 with -O63M and CLSF-I63J
Protocol is terminated. Thus, when transferring a datagram to a terminal other than the own subnetwork,
Datagram delivery is realized end-to-end at the end of two ATM connections.

(Embodiment 5-3-1) Next, a more specific embodiment will be described.

Referring to FIG. 63, an example in which a datagram is transferred from terminal 63A to terminal 63B will be briefly described.

The address resolution in this embodiment is performed as follows.

When transmitting the datagram to the terminal 63B, the terminal 63A analyzes that the terminal 63B belongs to the external sub-network.
Transfer datagram to M. The CLSF-O63M analyzes the address information of the received datagram, and
B (CLSF-I63J) does not have the ATM layer address information for transferring the datagram,
The address resolution request cell having the address information of the terminal 63B is transferred to the ARS 63D.

If the ARS 63D has information that can be resolved, the ARS 63D requests an AR request (CLSF-O63).
VC to transfer cells from M to CLSF-I63J
(I / VPI information) directly. However, ARS6
When resolution of an address cannot be performed in 3D, an appropriate ARS is accessed to perform resolution. When the resolution of the address is completed, the result (AR response) is transferred to CLSF-O63M.

The datagram transfer in the present embodiment is performed as follows.

The terminal 63A is connected to the ATM connection 641
Cell having datagram information using CLSF-O6
Transfer to 3M. CLSF-O63M analyzes the address information of the datagram, converts the datagram into cells,
The cell is connected to the CLSF-I using the ATM connection 645.
Transfer to 63J. IWU636 and IWU637
Performs the relaying of the ATM connection by rewriting the VCI / VPI information of the received cell. A
CLSF-I6 which is the terminal point of the TM connection 645
In 3J, processing of the network layer is performed. CL
SF-I63J analyzes the network address,
The datagram is relayed to the ATM connection 646 and transferred to the terminal 63B.

[0564] As described above, the ATM connection is connected to the terminal 6.
3A to CLSF-O63M; CLSF-O63M to CLSF-I63J; and CLSF-I63J and terminal 63B, and the ATM connection is terminated twice.

(Embodiment 5-4) Case where VPI routing ATM network is used-Part 2

[0566] In this embodiment, the method of realizing the ATM connection set between the ARSs is the same as that of the above-described method, and the description is omitted here.

Referring to FIG. 63 or FIG. 68, terminal 63
Four methods for setting an ATM connection required to transfer a datagram from A to the terminal 63B, the public network 635, and the terminal 63C will be described.

ATM Connection Setting Method (1) First, the transfer of a datagram from the terminal 63A to the terminal 63C will be described. Three ATM connections, namely
(1) Connection 691 from terminal 63A to CLSF-O63M, (2) CLSF-O63M to CLSF-I6
Connection to 3L 692, 693, (3) CLSF-
A connection 694 from the I63L to the terminal 63C is formed. The terminal 63A transfers a cell having VPI-F = VPI _63M . VCI-F1 is VPI _63A which is an access address of terminal 63A. CLSF-O
63M forwards the following cells to IWU 636:
VPI-F is V which is an access address of IWU 636.
PI _636-2, VCI-F1 has VPI _63M which is the access address of the CLSF-O63M, VCI-F2 is set during connection 692 values, respectively.
The VCI of the cell received by the IWU 636 is analyzed, and the corresponding VPI-F and VCI-F2 are analyzed. V
In the CI-F1, a VPI _636-1 which is an access address for the network 631 of the _{IWU 636} is written. Note that VCI-F2 is a value determined when the connection 693 is set, and the value of VCI-F1 of the cell received by the IWU 636, that is, CLSF-O63.
It is also possible to copy and write VPI _63M , which is the access address of M. The datagram arriving at CLSF-I63L is subjected to cell reassembly once, and the ATM connection is terminated. The CLSF-I 63L analyzes the address information of the upper layer and recognizes that the destination of the datagram is the terminal 63C. Therefore, CLSF-I
63L generates the following cell and transfers the datagram to terminal 63C.

[0566] VPI-F is VPI _63C which is an access address of the terminal _63C , and VCI-F1 is CLSF-I63.
VPI _63L , VCI-F2
Is an identification number assigned to the connection 694. The terminal 63C can uniquely identify the datagram to which the cell belongs from the VCI-F1 and the VCI-F2. That is, the datagram can be reassembled.

Next, transfer of a datagram from terminal 63A to public network 635 will be described. It is formed from two ATM connections 691 and 692, 695, 696.
The terminal 63A transfers a cell having VPI-F = VPI _63M . CLSF-O63M transfers the following cells to IWU 636. VPI-F is VPI _636-2 , VC
I-F1 is VPI _63M , VCI-F2 is connection 6
It has the value set at the time of setting 92. The VCI of the cell received by the IWU 636 is analyzed, and the corresponding VPI-F
And VCI-F2 are analyzed. VCI-F1 is I
A VPI _636-1, which is an access address of the _{WU 636} for the network 631, is written. Note that VC
The I-F2 is a value determined when the connection 695 is set, and the VCI-of the cell received by the IWU 636.
The value of F1, i.e. it is also possible to write and copy the VPI _63M which is the access address of the CLSF-O63M. The IWU 639 uses an ATM connection 696 defined by the public network 635 based on the received VCI information of the cell.
Is written, and the cell is transferred to the public network 635.

[0571] Finally, the transfer of a datagram from the terminal 63A to the terminal 63B will be described. Three ATM connections,
That is, (1) a connection 691 from the terminal 63A to the CLSF-O63M, and (2) a connection from the CLSF-O63M to the CLS.
Connection 692, 697, 69 to FI63J
8. (3) Connection 699 from CLSF-I 63J to 63B. Terminal 63A is VPI-F = VPI _63M
Transfer the cell with. CLSF-O63M is IW
The following cell is transferred to U636. VPI-F is I
VPI _636-2 , the access address of _{WU 636} , V
CI-F1 has VPI _63M which is the access address of the CLSF-O63M, VCI-F2 is set during connection 692 values, respectively. The VCI of the cell received by the IWU 636 is analyzed, and the corresponding VPI-
F and VCI-F2 are analyzed. VCI-F1 is
The VPI _636-1 which is an access address for the network 631 of the _{IWU 636} is written. Note that V
CI-F2 is a value determined when the connection 697 is set, and the VCI of the cell received by the IWU 636.
The value of -F1, i.e. it is also possible to write and copy the VPI _{63 M} which is the access address of the CLSF-O63M.

Next, the V of the cell received by IWU 637 is
The CI is analyzed, and the corresponding VPI-F, VCI-F1, and VCI-F2 are analyzed. VCI / VPI is C
LSF-I63J transmits all terminals / servers (such as CLSF) in the network, that is, datagrams to CLSF.
-It must be allocated so that it can be identified even if cells from terminals that can transfer to I63J arrive at the same time.

The datagram arriving at CLSF-I63J is subjected to cell reassembly once, and the ATM connection is terminated. The CLSF-I 63J analyzes the address information of the upper layer, and determines that the destination of the datagram is the terminal 63B.
Recognize that Therefore, CLSF-I63J generates the following cell and transfers the datagram to terminal 63B. VPI-F is VPI _63B which is an access address of the terminal 63B, and VCI-F1 is CLSF-I63J.
The access addresses VPI _63J and VCI-F2 are identification numbers assigned to the connection 699.
The terminal 63B can uniquely identify the datagram to which the cell belongs from the VCI-F1 and the VCI-F2. That is, the datagram can be reassembled.

ATM Connection Method (2) In CLSF-O, datagram reassembly is performed, but address resolution is performed by each terminal. VCI-F1 of the cell in which the address information of the resolved target network is transferred from the terminal to the CLSF-O
Write to. Cell transfer (datagram transfer) from the CLSF-O cannot be executed in a pipeline manner. However, in CLSF-O, it is not necessary to resolve the address of the network to which the destination terminal belongs from the address information in the received datagram.

Each IWU, CLSF-O and CLSF-
I have all obtained VPI addresses, and
The connections identified by are all configured to be ATM connections used for cell transfer for datagram communication (connectionless communication). That is,
The CLSF server and the IWU have their own terminals /
To communicate with the server, an ATM connection (not a connectionless connection) is required. That is, the CLSF and the IWU need to acquire at least two or more access addresses (VPI) at the time of boot.

At this time, the following method is used as a coding method of the VCI field of the cell for connectionless communication. The VCI field includes two 8-bit subfields (VCI-F1 and VCI-F
2: The definition position of the field is not mentioned).

[0577] A coding method of a cell transferred in the external subnetwork in the present embodiment will be described. This is because the subnetwork to which the terminal belongs
Is the coding of the VCI field when a cell is transferred to an external sub-network via the VCI. VCI-
The identification number of the subnetwork to which the source terminal belongs is written in F1 and the identification number of the source terminal in the subnetwork is written in VCI-F2. For example,
Network 631 as the identification number of the subnetwork
It is also possible to use the access address (VPI) of each IWU in (1) as the identification address of the subnetwork (the network 631 itself is appropriately coded). Further, the terminal identification number may be an access address (VPI) of the terminal in each subnetwork. For example, the VCI field of the connectionless communication cell output from the terminal 63B is VCI-
F1 = VPI _637-1 and VCI-F2 = VPI _63B .

(CLSF-O to IWU cell) VC
The identification number of the sub-network to which the source terminal belongs is written in I-F1, and the identification number in the CLSF-O sub-network is written in VCI-F2. For example, the VCI-F2 field of the connectionless communication cell issued from the CLSF-O63M can be VPI _63M .

(Cell from terminal to CLSF-O) VCI
In -F1, the identification number of the subnetwork to which the destination terminal belongs is written, and in VCI-F2, the identification number of the transmission source terminal in the subnetwork is written. For example, V of the cell of the connectionless communication issued from the terminal 63B
The CI-F2 field may be VPI _63B .

[0580] With the above configuration, in the address resolution procedure performed prior to the transfer of the datagram, the terminal only needs to obtain the identification number of the subnetwork to which the destination terminal belongs, and then the datagram (multiple (Or one cell) can be transferred as follows.

First, the transfer of a datagram from the terminal 63A to the terminal 63C will be described. Terminal 63A is a CLSF-
The following cell is transferred to O63M. VPI-F is C
VP which is an access address of LSF-O63M
I _63M and VCI-F1 are set to the identification number of the network 631 (for example, VPI _63L which is an access address of CLSF-I _63L can be used), and VCI-F2 is set to VPI _63A which is the access address of terminal 63A.

The CLSF-O63M transfers the following cell to the IWU 636. VPI-F is IWU636
_Is set to VPI _636-2 , which is the access address of V
The CI-F1 can copy the VCI-F1 of the received cell as it is, and can use the identification number of the network 631 (for example, VP, which is the access address of the CLSF-I 63L).
I _63L can also be used). VCI-F
2 is V which is an access address of CLSF-O63M.
Set to PI _63M .

The IWU 636 analyzes the VCI of the received cell, and analyzes the corresponding VPI-F and VCI-F2. For example, CLSF-I63L is added to VCI-F1.
In the embodiment in which the access address VPI _63L is written, the VPI-F of the cell to be output can be realized by copying the VPI _63L that is the VCI-F1 information of the received cell. IWU636 is included in VCI-F1.
VPI _636-1, which is an access address for the network 631, is written. VCI-F2 can be set to be transparent. That is, for example, the VCI-F of the cell transferred from the CLSF-O63M
1 is the access address VPI _{63L of} CLSF-I63L
Is written in the embodiment, (1) the VCI-F1 of the reception cell is copied to the VPI-F of the transmission cell,
(2) VCI-F2 of the transmission cell is replaced with VCI-F2 of the reception cell
(3) IWU is added to VCI-F1 of the transmission cell.
The cell relaying is realized by the procedure of writing the VPI _636-1 which is the access address of the cell.

The datagram arriving at CLSF-I63L is temporarily reassembled, and the ATM connection is terminated. The CLSF-I 63L analyzes the address information of the upper layer, and sends the datagram to the terminal 63C.
Recognize that Therefore, the CLSF-I63L uses the VP where the VPI-F is the access address of the terminal 63C.
Generate a cell that is I _63C and forward the datagram to terminal 63C. Note that the terminal 63C can uniquely identify the datagram to which the cell belongs from VCI-F1 and VCI-F2. That is, the datagram can be reassembled.

Next, transfer of a datagram from terminal 63A to public network 635 will be described. Terminal 63A is a CLSF
-Transfer the following cell to O63M. VPI-F is VPI which is an access address of CLSF-O63M.
_63M , VCI-F1 is an identification number of the network 635 (for example, VPI which is an access address of the IWU 639).
_{63 9} can also be used), VCI-F2 terminal 63A
_Is set to VPI _63A , which is the access address of.

[0586] The CLSF-O63M transfers the following cell to the IWU 636. VPI-F is IWU636
_Is set to VPI _636-2 , which is the access address of V
The CI-F1 can copy the VCI-F1 of the received cell as it is, and can use the identification number of the public network 635 (for example, I
VPI ₆₃₉ , which is the access address of WU ₆₃₉ , may be used). VCI-F2 is CLS
Set to VPI _63M which is the access address of the F-O63M.

[0587] The IWU 636 analyzes the VCI of the cell received, and analyzes the corresponding VPI-F and VCI-F2. For example, in the case of the embodiment in which the access address VPI ₆₃₉ of the IWU ₆₃₉ is written in the VCI-F1, the VPI-F of the output cell is realized by copying the IWU 639 which is the VCI-F1 information of the received cell. it can. VCI-F1 has an access address VW which is an access address for the network 631 of the IWU 636.
PI _636-1 is written. VCI-F2 can be set to be transparent. That is, for example, the VCI-of the cell transferred from the CLSF-O63M is
In the case of the embodiment in which the access address VPI ₆₃₉ of the IWU ₆₃₉ is written in F1, (1) the V
CI-F1 is copied to the VPI-F of the sending cell, (2) VCI-F2 of the receiving cell is copied to VCI-F2 of the sending cell, and (3) VCI-F1 of the sending cell is assigned the IWU access address. The procedure of writing a certain VPI _636-1 realizes cell relaying. IWU639
Writes the VCI / VPI assigned to the ATM connection 696 defined in the public network 635 from the received VCI information of the cell, and transfers the cell to the public network 635.

[0588] Finally, the transfer of a datagram from the terminal 63A to the terminal 63B will be described. Terminal 63A is a CLSF
-Transfer the following cell to O63M. VPI-F is VPI which is an access address of CLSF-O63M.
_63M and VCI-F1 are identification numbers of the network 633 (for example, VPI _63J which is an access address of CLSF-I _63J and V which is an access address of IWU 637).
PI ₆₃₇ can be used), and VCI-F2
It is set to VPI _63A , which is the 3A access address.

[0589] The CLSF-O63M transfers the following cell to the IWU 636. VPI-F is IWU636
_Is set to VPI _636-2 , which is the access address of V
CI-F1 can be directly copy the VCI-F1 of the received cell, VPI _{637- 1} which is the access address of the identification number (e.g. IWU637 network 633
Can also be used). VCI-F2 is
VPI which is an access address of CLSF-O63M
Set to _63M . VC of cell received at IWU 636
I was analyzed and the corresponding VPI-F and VCI-F2
Is analyzed. For example, in the case of the embodiment in which the access address VPI _637-1 of the IWU 637 is written in the VCI-F1, the VPI-F of the output cell copies the VPI _637-1 which is the VCI-F1 information of the received cell. It can be realized by doing. In the VCI-F1, a VPI _636-1 which is an access address for the network 631 of the _{IWU 636} is written. VCI-F2 can be set to be transparent. That is, for example, the VCI-F of the cell transferred from the CLSF-O63M
In the case of the embodiment in which the access address VPI _637-1 of the IWU 637 is written in
CI-F1 is copied to the VPI-F of the sending cell, (2) VCI-F2 of the receiving cell is copied to VCI-F2 of the sending cell, and (3) VCI-F1 of the sending cell is assigned the IWU access address. The procedure of writing a certain VPI _636-1 realizes cell relaying.

Next, the V of the cell received by IWU 637 is
The CI is analyzed, and the corresponding VPI-F, VCI-F1, and VCI-F2 are analyzed. CL from IWU637
The VCI field of the cell transferred to the SF-I63J can set the VCI field information of the received cell transparently. That is, VC
I-F1 = VPI _636-1, can be a VCI-F2 = VPI _63M. VPI-F is CLSF-I6
VPI _63J, which is an access address of 3J, is set. The datagram arriving at CLSF-I63J is subjected to cell reassembly once, and the ATM connection is terminated. The CLSF-I 63J analyzes the address information of the upper layer and recognizes that the destination of the datagram is the terminal 63B. The CLSF-I 63J generates a cell whose VPI-F is the VPI _63B that is the access address of the terminal 63B, and transfers the datagram to the terminal 63B. The terminal 63B can uniquely identify the datagram to which the cell belongs from the VCI-F1 and the VCI-F2. That is, the datagram can be reassembled.

[0591] ATM connection setting method (3) In this method (2), a cell to which a datagram belongs is transferred in a pipeline manner. In CLSF-O, datagram reassembly is performed, but address resolution is performed by each terminal. In CLSF-O, the resolution of the address of the network to which the destination terminal belongs is determined from the address information in the received datagram. No need to do. The address information of the resolved target network is written to the VCI-F1 of the cell transferred from the terminal to the CLSF-O.

At this time, the following method is used as a coding method of the VCI field of the cell for connectionless communication.

(Inside external subnetwork—same as method 2) The identification number of the subnetwork to which the source terminal belongs is written in VCI-F1, and the identification number of the source terminal in the subnetwork is written in VCI-F2. Is written.

(CLSF-O to IWU cell) VC
The identification number of the subnetwork to which the destination terminal belongs is written in I-F1, and the identification number of the transmission source terminal in the subnetwork is written in VCI-F2. For example,
In the cell relating to the datagram output from the terminal 63A, the VCI-F2 field can be VPI _63A .

(Cell from Terminal to CLSF-O—Method 2)
The same applies to VCI-F1; the identification number of the subnetwork to which the destination terminal belongs is written; and the VCI-F2, the identification number within the subnetwork to which the source terminal belongs.

First, transfer of a datagram from terminal 63A to terminal 63C will be described. Terminal 63A is a CLSF-
The following cell is transferred to O63M. VPI-F is C
VP which is an access address of LSF-O63M
I _63M and VCI-F1 are identification numbers of the network 631 (for example, VPI _63L which is an access address of CLSF-I _63L can be used), VCI-F2
Is set to VPI _63A , which is the access address of terminal 63A.

The CLSF-O63M transfers the following cells to the IWU 636. VPI-F is IWU636
_Is set to VPI _636-2 which is the access address of
The VCI-F1 and VCI-F2 can copy the information of the received cell as it is. Therefore, VCI-
F1 is an identification number of the network 631 (for example, CLSF
Also possible) using the VPI _63L which is the access address of -I63L, VCI-F2 is set to the VPI _63A which is the access address of the terminal 63A. At this time, the cells can be sequentially transferred to the IWU 636 in a pipeline manner without once reassembling the datagram in the CLSF-O63M.

The IWU 636 analyzes the VCI of the cell received, and analyzes the corresponding VPI-F and VCI-F2. For example, CLSF-I63L is added to VCI-F1.
In the embodiment in which the access address VPI _63L is written, the VPI-F of the cell to be output can be realized by copying the VPI _63L that is the VCI-F1 information of the received cell. IWU636 is included in VCI-F1.
VPI _636-1, which is an access address for the network 631, is written. VCI-F2 can be set to be transparent. That is, for example, the VCI-F of the cell transferred from the CLSF-O63M
1 is the access address VPI _{63L of} CLSF-I63L
Is written in the embodiment, (1) the VCI-F1 of the reception cell is copied to the VPI-F of the transmission cell,
(2) VCI-F2 of the transmission cell is replaced with VCI-F2 of the reception cell
And (3) VPI _636-1 which is the access address of the IWU is written in the VCI-F1 of the transmission cell, thereby realizing cell relaying.

The datagram arriving at CLSF-I63L is once subjected to cell reassembly, and the ATM connection is terminated. The CLSF-I 63L analyzes the address information of the upper layer, and sends the datagram to the terminal 63C.
Recognize that Therefore, the CLSF-I63L uses the VP where the VPI-F is the access address of the terminal 63C.
Generate a cell that is I _63C and forward the datagram to terminal 63C. Note that the terminal 63C can uniquely identify the datagram to which the cell belongs from VCI-F1 and VCI-F2. That is, the datagram can be reassembled.

Next, data from the terminal 63A to the public network 635 will be described.
The transfer of datagrams will be described. Terminal 63A is a CLSF
-Transfer the following cell to O63M. VPI-F is
VPI which is an access address of CLSF-O63M
_63M, VCI-F1 are identification numbers of the public network 635 (for example,
VPI which is the access address of IWU639₆₃₉For
VCI-F2 can access the terminal 63A.
VPI which is the address _63ASet to.

The CLSF-O63M transfers the following cells to the IWU 636. VPI-F is IWU636
_Is set to VPI _636-2 which is the access address of
The VCI-F1 and VCI-F2 can copy the information of the received cell as it is. Therefore, VCI-
F1 is set to the identification number of the public network 635 (for example, VPI ₆₃₉ which is an access address of IWU ₆₃₉ can be used), and VCI-F2 is set to VPI _63A which is the access address of terminal 63A. At this time, CLSF
In O63M, cells are sequentially IW-like pipelined without reassembling datagrams.
U636.

The IWU 636 analyzes the VCI of the cell received, and analyzes the corresponding VPI-F and VCI-F2. For example, in the case of the embodiment in which the access address VPI ₆₃₉ of the IWU ₆₃₉ is written in the VCI-F1, the VPI-F of the cell to be output copies the IWU 639, which is the VCI-F1 information of the received cell, by copying realizable. A VPI _636-1, which is an access address of the _{IWU 636} for the network 631, is written in the VCI-F1. VCI-F2 can be set to be transparent. That is, for example, the VCI-F of the cell transferred from the CLSF-O63M
1, the access address VPI ₆₃₉ of the IWU ₆₃₉ is written in (1) VC of the received cell.
I-F1 is copied to VPI-F of the transmission cell, (2) VCI-F2 of the reception cell is copied to VCI-F2 of the transmission cell, and (3) VCI-F1 of the transmission cell has an IWU access address. The cell is relayed by the procedure of writing VPI _636-1 .

The IWU 639 writes the VCI / VPI assigned to the ATM connection 696 defined by the public network 635 from the received VCI information of the cell, and transfers the cell to the public network 635.

Finally, the transfer of a datagram from the terminal 63A to the terminal 63B will be described. Terminal 63A is a CLSF
-Transfer the following cell to O63M. VPI-F is VPI which is an access address of CLSF-O63M.
_63M and VCI-F1 are set to the identification number of the network 633 (for example, VPI _63J which is the access address of CLSF- _I63J can be used), and VCI-F2 is set to VPI _63A which is the access address of terminal 63A.

The CLSF-O63M transfers the following cells to the IWU 636. VPI-F is IWU636
_Is set to VPI _636-2 which is the access address of
The VCI-F1 and VCI-F2 can copy the information of the received cell as it is. Therefore, VCI-
F1 is an identification number of the network 633 (for example, IWU6
Also possible) using the VPI _637-1 is 37 access address, VCI-F2 is set to VPI _63A which is the access address of the terminal 63A. At this time, CL
In the SF-O63M, cells can be sequentially transferred to the IWU 636 in a pipeline manner without performing datagram reassembly once. The VCI of the cell received by the IWU 636 is analyzed, and the corresponding VPI-
F and VCI-F2 are analyzed. For example, VCI-
In the case of the embodiment in which the access address VPI _637-1 of the IWU 637 is written in F1, the VPI-F of the cell to be output is obtained by copying the VPI _637-1 which is the VCI-F1 information of the received cell. realizable. VCI-
In F1, a VPI _636-1 which is an access address for the network 631 of the IWU ₆₃₆ is written.
VCI-F2 can be set to be transparent. That is, for example, in the embodiment in which the access address VPI _{637-1 of the} IWU 637 is written in the VCI-F1 of the cell transferred from the CLSF-O63M, (1) the VCI-F1 of the reception cell is replaced with the VPI-F1 of the transmission cell.
-F, (2) Copy the VCI-F2 of the receiving cell to VCI-F2 of the sending cell, and (3) Copy the VCI-F2 of the sending cell.
F1 has a VPI which is an access address of the IWU.
_By the procedure of writing _636-1 , cell relaying is realized.

Next, the V of the cell received by IWU 637 is
The CI is analyzed, and the corresponding VPI-F, VCI-F1, and VCI-F2 are analyzed. CL from IWU637
The VCI field of the cell transferred to the SF-I63J can set the VCI field information of the received cell transparently. That is, VC
I-F1 = VPI _636-1, can be a VCI-F2 = VPI _63M. In addition, VPI-F is CLSF-I
VPI _63J, which is the access address of _63J, is set.

The datagram arriving at CLSF-I63J is once subjected to cell reassembly, and the ATM connection is terminated. The CLSF-I 63J analyzes the address information of the upper layer, and determines that the destination of the datagram is the terminal 63B.
Recognize that CLSF-I63J is VPI-
A cell in which F is the VPI _63B that is the access address of the terminal 63B is generated, and the datagram is transferred to the terminal 63B. Note that the terminal 63B has VCI-F1 and VCI-F1.
The datagram to which the cell belongs can be uniquely identified from F2. That is, the datagram can be reassembled.

ATM Connection Setting Method (4) This is an example of a method (2) in which datagram transfer to a different sub-network is performed in a pipeline manner. Datagram transfer from a terminal belonging to the same subnetwork to a terminal in the same subnetwork cannot be performed in a pipeline manner, but pipeline transfer is possible if the destination subnetwork is different. In this case, CLSF-O needs to have some buffer space.

In CLSF-O, datagram reassembly is performed, but address resolution is performed by each terminal. In CLSF-O, the address of the network in which the destination terminal exists is determined from the address information in the received datagram. No resolution is required. The address information of the resolved target network is written in the VCI-F1 of the cell transferred from the terminal to the CLSF-O. A datagram transfer procedure in CLSF-O will be described below.

Step 1: A cell is received from the terminal (the first cell for the datagram).

Step 2: Analyze the target sub-network address information written in VCI-F1.

Step 3: Check whether there is any datagram currently being transferred to the analyzed subnetwork.

Step 4: The CLSF-O can recognize the end of the datagram transfer by using an identification code of ATM layer or higher (for example, payload type coding for AAL5). When another datagram is not transferred to the analyzed subnetwork, the received cell is relayed to the target subnetwork. Received cells are relayed without reassembly once in CLSF-O; pipeline transfer.

Step 5: On the other hand, when another datagram is being transferred from the other terminal to the analyzed subnetwork, the received cell needs to be buffered until the datagram transfer is completed. There is.
When the end of the transfer of another datagram is confirmed, the buffered cells are sequentially transferred. At this time, the buffered cell (datagram) does not need to be reassembled, and the transfer of the first cell must be started before the last cell of the datagram to which the buffered cell belongs. Can be. Also, if an appropriate protocol is defined between the CLSF-O and the terminal, it is possible to perform flow control so that buffered cells are not discarded due to buffer overflow.

In the present embodiment, the following method is used as the coding method of the VCI field of the cell for connectionless communication.

(Inside the external subnetwork—same as method 2) The identification number of the subnetwork to which the source terminal belongs is written in VCI-F1, and the identification number of the source terminal in the subnetwork is written in VCI-F2. Written.

(Cell from CLSF-O to IWU) VC
The identification number of the subnetwork to which the destination terminal belongs is written in I-F1, and the identification number of the transmission source network is written in VCI-F2. For example, the cell related to the datagram output from the terminal 63A is VCI-F
The two fields may be VPI _636-1 or the like.

(Cell from Terminal to CLSF-O—Method 2)
The same is applied to VCI-F1, the identification number of the subnetwork to which the destination terminal belongs is written, and the identification number of the transmission source terminal in the subnetwork is written to VCI-F2.

First, the transfer of a datagram from the terminal 63A to the terminal 63C will be described. Terminal 63A is a CLSF-
The following cell is transferred to O63M. VPI-F is C
VP which is an access address of LSF-O63M
I _63M and VCI-F1 are set to the identification number of the network 631 (for example, VPI _63L which is an access address of CLSF-I _63L can be used), and VCI-F2 is set to VPI _63A which is the access address of terminal 63A.

The CLSF-O63M transfers the following cells to the IWU 636 when no other datagram transfer to the subnetwork 631 is performed. VP
IF is VP which is an access address of IWU 636.
I _636-2 is set. The VCI-F1 can copy the information of the received cell as it is. VCI-F1
Is the identification number of the network 631 (for example, CLSF-I
VPI _63L , which is an access address of _63L , may be used). VCI-F2 is, for example, VPI which is an identification number expressing the subnetwork 632.
_636-1 (the access address of _IWU 636). VCI-F2 may be any identifier for identifying its own subnetwork. At this time, CLSF-
In O63M, cells are sequentially IWU-like in a pipeline without performing datagram reassembly.
636.

The IWU 636 analyzes the VCI of the cell received, and analyzes the corresponding VPI-F and VCI-F2. For example, CLSF-I63L is added to VCI-F1.
In the embodiment in which the access address VPI _63L is written, the VPI-F of the cell to be output can be realized by copying the VPI _63L which is the VCI-F1 information of the received cell. A VPI _636-1, which is an access address of the _{IWU 636} for the network 631, is written in the VCI-F1. VCI-F2 can be set to be transparent. That is, for example, the VCI-F1 of the cell transferred from the CLSF-O63M
In this embodiment, the access address VPI _63L of the CLSF-I _63L is written in (1) the VCI-F1 of the reception cell is copied to the VPI-F of the transmission cell, and (2)
Cell relaying is realized by copying VCI-F2 of the receiving cell to VCI-F2 of the transmitting cell, and (3) writing VPI _636-1 , which is the access address of the IWU, to VCI-F1 of the transmitting cell. Is done.

The datagram arriving at CLSF-I63L is temporarily reassembled, and the ATM connection is terminated. The CLSF-I 63L analyzes the address information of the upper layer, and sends the datagram to the terminal 63C.
Recognize that Therefore, the CLSF-I63L uses the VP where the VPI-F is the access address of the terminal 63C.
Generate a cell that is I _63C and forward the datagram to terminal 63C. The terminal 63C can uniquely identify the datagram to which the cell belongs from the VCI-F1 and VCI-F2 (the datagram can be reassembled).

Next, data from the terminal 63A to the public network 635 is transmitted.
The transfer of datagrams will be described. Terminal 63A is a CLSF
-Transfer the following cell to O63M. VPI-F is
VPI which is an access address of CLSF-O63M
_63M, VCI-F1 are identification numbers of the public network 635 (for example,
VPI which is the access address of IWU639₆₃₉For
VCI-F2 can access the terminal 63A.
VPI which is the address _63ASet to.

[0624] The CLSF-O63M performs the IW operation when no other datagram transfer to the public network 635 is performed.
The following cell is transferred to U636. VPI-F is
It is set to VPI _636-2 which is an access address of _{IWU 636} . The VCI-F1 can copy the information of the received cell as it is. VCI-F1 is public network 6
It is set to an identification number of 35 (for example, VPI ₆₃₉ which is an access address of IWU ₆₃₉ can be used). The VCI-F2 can be a VPI _636-1 (an access address of the _IWU 636), which is an identification number representing the subnetwork 632. VCI-F2 may be any identifier for identifying its own subnetwork. At this time, the cells can be sequentially transferred to the IWU 636 in a pipeline manner without once reassembling the datagram in the CLSF-O63M.

[0625] The IWU 636 analyzes the VCI of the cell received, and analyzes the corresponding VPI-F and VCI-F2. For example, in the case of the embodiment in which the access address VPI ₆₃₉ of the IWU ₆₃₉ is written in the VCI-F1, the VPI-F of the cell to be output is obtained by copying the VPI ₆₃₉ which is the VCI-F1 information of the received cell. realizable. The VCI-F1 includes a VP, which is an access address for the network 631 of the IWU 636.
I _636-1 is written. VCI-F2 can be set to be transparent. That is, for example, C
IW is added to VCI-F1 of the cell transferred from LSF63M.
In the embodiment in which the access address VPI _{639 of} U639 is written, (1) VCI-F1
Is copied to the VPI-F of the sending cell, and (2)
Copy CI-F2 to VCI-F2 of the transmission cell, (3)
Cell relaying is realized by the procedure of writing VPI _636-1 , which is the access address of the IWU, in the VCI-F1 of the transmission cell.

[0626] The IWU 639 writes the VCI / VPI assigned to the ATM connection 696 defined by the public network 635 from the received VCI information of the cell, and transfers the cell to the public network 635.

[0627] Finally, the transfer of a datagram from the terminal 63A to the terminal 63B will be described. Terminal 63A is a CLSF
-Transfer the following cell to O63M. VPI-F is VPI which is an access address of CLSF-O63M.
_63M , VCI-F1 is an identification number of the network 633 (for example, VPI which is an access address of the IWU 637).
_{63 7-1} can be used), and VCI-F2
A is set to VPI _63A which is the access address of A.

The CLSF-O63M transfers the following cells to the IWU 636 when no other datagram transfer to the subnetwork 633 is performed. VP
IF is VP which is an access address of IWU 636.
I _636-2 is set. The VCI-F1 can copy the information of the received cell as it is. VCI-F1
Is the identification number of the network 633 (for example, IWU 637)
VPI _637-1 , which is the access address of the user, can be used). VCI-F2 is a VPI _636-1 (IW), which is an identification number representing the subnetwork 632.
U636 access address). V
CI-F2 may be an identifier for identifying its own subnetwork. At this time, CLSF-O63
M, the cells are sequentially pipelined in IWU 636 without reassembling the datagram.
Can be transferred to

[0629] The VCI of the cell received by IWU 636 is
Analyzed and the corresponding VPI-F and VCI-F2
Is analyzed. For example, IWU 637 access to VCI-F1
Access address VPI_637-1Embodiment in which is written
In the case of, the output cell VPI-F is the received cell
VPI which is VCI-F1 information of_637-1Can be copied
And can be realized. VCI-F1 is a network of IWU636.
VP which is an access address for the network 631
I_636-1Is written. VCI-F2
Can be set to Arent. That is, for example, C
For VCI-F1 of the cell transferred from LSF-O63M
Access address VPI of IWU637 _637-1Is written
(1) VCI of the received cell
-F1 is copied to the VPI-F of the sending cell, and (2)
Copy VCI-F2 of file to VCI-F2 of sending cell
(3) IWU access to VCI-F1 of the sending cell
VPI which is the address_636-1To the procedure of writing
More cell relaying is realized.

Next, the V of the cell received by IWU 637 is
The CI is analyzed, and the corresponding VPI-F, VCI-F1, and VCI-F2 are analyzed. CL from IWU637
The VCI field of the cell transferred to the SF-I63J can set the VCI field information of the received cell transparently. That is, VC
I-F1 = VPI _637-1 , VCI-F2 = VPI _636-1
It can be. VPI-F is CLSF-
VPI _63J which is an access address of _{I63J is} set.

The datagram arriving at CLSF-I63J is once subjected to cell reassembly, and the ATM connection is terminated. The CLSF-I 63J analyzes the address information of the upper layer, and determines that the destination of the datagram is the terminal 63B.
Recognize that CLSF-I63J is VPI
VPI _{63 B} where -F is the access address of the terminal _{63 B}
Is generated, and the datagram is transmitted to the terminal 63B.
Transfer to The terminal 63B is connected to the VCI-F1 and the VCI-F1.
The datagram to which the cell belongs can be uniquely identified from -F2 (the datagram can be reassembled).

(Sixth Embodiment) Next, an embodiment relating to a network configuration of a horizontal topology will be described. FIG. 69 shows a system configuration of the present embodiment. As shown in the figure, a plurality of sub-networks 701 to 706
Are internetworking using IWU. The IWU can realize the relaying of the ATM cell without terminating the ATM connection. That is, it has a function of converting the VCI / VPI of the received cell into the VCI / VPI assigned to the corresponding ATM connection in the adjacent network.

The connection line between the sub-networks 701 to 706 may be a connection line in a campus or a dedicated line / switch line of a public network. Further, although public networks 707 and 708 are shown in the figure, this indicates that it is possible to access terminals and networks other than the defined network via the public network. That is, for example, when communicating with a terminal connected to a general public network, the communication is performed through the public networks 707 and 708.

Each of the sub-networks 701 to 706 has a CLSF and can handle cells related to connectionless communication. CL
The setting and operation of the SF and a specific embodiment when connectionless communication is realized between terminals will be described later.

There is no particular limitation on the network distance (indicator of how many relay points (sub-networks) are required to connect arbitrary points inside the network). In the two-layer network described in the previous embodiment,
The network topology is such that a target subnetwork can be reached via at most one relay subnetwork. However, in the network handled here, there is basically no restriction on the distance of this network.

In the following embodiments, no particular reference is made to a method of configuring an ATM connection for realizing address resolution. The method of realizing the address resolution is basically the same as the method described in the above embodiment. That is, each address resolution server is on an equal footing, and a method of analyzing network address information (an image as shown in FIG. 49);
This is a method in which the address resolution server has a logical hierarchical structure (image as shown in FIG. 50) and performs management analysis of address information. The ATM connection set between the address resolution servers is an arbitrary configuration from a full mesh configuration (configuration as shown in FIG. 59) to a minimum spanning tree configuration (configuration as shown in FIG. 58). It is possible to

(Embodiment 6-1) A case where the present invention is applied to a general ATM network.

[0638] The ATM connection in this embodiment is as follows. In order to transfer a datagram from a terminal connected to each subnetwork to a terminal connected to an arbitrary subnetwork, an ATM connection is set from each IWU to CLSFs present in all subnetworks. That is, IWU to CL
An ATM connection (one-way ATM connection) to the SF is set in a full mesh state. In addition, ATM
IWU (Relay IWU) on the connection path
Then, rewriting of ATM header information (at least VCI
/ VPI conversion) and cell relaying
Performed as a layer task. That is, the ATM connection is not terminated in principle via the IWU.

[0639] The position where the CLSF exists can also exist at the position of the IWU. In addition, public network 707
And 708 when the datagram is forwarded to the defined network.
0M, received cell (datagram belongs to)
Is forwarded to the server that temporarily terminates the datagram connection of the public network. Note that this server can also exist in the IWU. The server that terminates the datagram transferred from the public network relays the datagram in the same procedure as the transfer of the datagram from the terminal in the defined network.

FIG. 80 shows the configuration of an ATM connection when a datagram is transferred from the public network 707 to the terminal 70A. The datagram transferred from the public network 707 is once terminated in the CLSF-P811 and then the CLS
The data is transferred to the terminal 70A via F7011. The ATM connections are three connections 812, 813 and 814.

(Terminal Protocol) The procedure for the terminal in this embodiment to transmit a connectionless communication datagram to the destination terminal is as follows.

(1) The terminal issues an address resolution request (AR request). This is performed when the terminal cannot resolve the address by its own ability (for example, when there is no entry in the address resolution cache table), or is always performed.

(2) The terminal sends, from the address resolution server, VCI / VPI which is an identifier of an ATM connection prepared for accessing the corresponding terminal.
Get information.

(3) The terminal inserts a cell into the network with the obtained VCI / VPI, thereby completing the datagram transfer. The terminal does not need to perform the connection setup procedure defined in the ATM network when transferring the datagram.

(Routing) The final datagram delivery to each terminal in this embodiment is performed only for the network to which each CLSF belongs. That is, for example, the CLSF 7061 performs delivery of a datagram to a terminal belonging to the network 706. Similarly, CLSF7
Numeral 031 distributes a datagram to only terminals in the network 703. If the network address of the datagram received by each CLSF is not present in the address entry of the CLSF (or if the address of the received datagram is not an element of the network address space of the network), the datagram is It is determined that the delivery was incorrect. The actions for erroneous datagram delivery are not discussed here.

[0646] That is, each CLSF only needs to have the address information of the terminal of the network to which it belongs.
When the address of the received datagram exists in the own network, an appropriate ATM connection is selected and the datagram is relayed. FIG. 71 shows an example of protocol processing when a datagram is transferred from the terminal 70A to the terminal 70B. The ATM connection is temporarily terminated at CLSF7061. That is, C
The OSI layer 3 protocol is terminated in LSF7061. The layer 3 protocol processing is performed in the CLSF 7061, and the data unit is transferred to the terminal 70B using the ATM connection. Thus, when transferring a datagram to a terminal other than the own subnetwork,
Datagram delivery is realized end-to-end with only one ATM connection termination.

[0647] Similarly, FIG.
FIG. 81 shows a case where a datagram is transferred to the terminal 70A. FIG. 81 shows a protocol configuration when a datagram is transferred from the public network 707 to the terminal 70A.

(Embodiment 6-1-1) Next, a more specific embodiment will be described.

(Regarding Datagram Transfer from Terminal 70A to Terminal 70B) FIG. 70 shows a configuration diagram of an ATM connection. The terminal 70A resolves the address of the terminal 70B when transferring the datagram (analyzes the access address information of the subnetwork to which the terminal 70B belongs). That is, an AR request message containing the address information of the terminal 70B is transferred to the address resolution server ARS. After performing the address resolution, the ARS receiving the AR request returns VCI / VPI information for transferring the cell to the terminal 70B to the terminal 70A as an AR response.

ATM for transferring cell to terminal 70B
The terminal 70A that has acquired the layer address (VCI / VPI) information inputs a cell to which the VCI / VPI is added to the network. Cell is VCI / V in IWU70D.
After the conversion of the PI is performed, it is transferred to the IWU 70G.
Similarly, the cell is transferred to CLSF7061 via IWU70H and IWU70J. C that received the cell
The LSF 7061 analyzes the information of the address information (layer 3) of the datagram and transfers the cell to the terminal 70B. The cell transfer from the CLSF 7061 to the terminal 70B may be performed after the CLSF 7061 has received all the cells belonging to the datagram, or after analyzing the layer 3 address of the datagram, relaying the cells in a pipeline. You may.

(Regarding Datagram Transfer from Terminal 70A to Public Network 708) FIG. 72 shows a configuration diagram of an ATM connection. When transferring the datagram, the terminal 70A performs resolution of the address of the target terminal (analyzes the access address information of the subnetwork to which the destination terminal belongs). That is, an AR request message containing the address information of the destination terminal is transferred to the address resolution server ARS. A that received the AR request
After performing the address resolution, the RS returns VCI / VPI information for transferring the cell to the terminal to the terminal 70A as an AR response.

Terminal 7 that has acquired ATM layer address (VCI / VPI) information for transferring a cell to a destination terminal
OA inputs the cell to which the VCI / VPI is added to the network. Cell is VCI / VP in IWU70D
After the conversion of I is performed, it is transferred to the IWU 70E. Similarly, the cell is transferred to the public network 708 via the IWU 70M.

(Regarding Datagram Transfer from Public Network 707 to Terminal 70A) FIG. 80 shows a configuration diagram of an ATM connection. The source terminal exists in the public network 707, and a cell related to connectionless communication from the public network 707 is transmitted to the CLSF-P811 via the IWU 70K.
Transferred to The VCI / VPI conversion table of the IWU 70K is set so that when a cell having a VCI / VPI assigned to a cell used for connectionless communication arrives in the public network 707, the cell is transferred to the CLSF-P811. ing.

The CLSF-P 811 temporarily terminates the ATM connection and analyzes the datagram layer 3 address information. The address information to be analyzed is CLSF-P81
If it does not exist in the table in 1, an AR request is issued to the ARS. A for transferring a cell to terminal 70A
The CLSF-P811 that has acquired the TM layer address (VCI / VPI) information inputs the cell with the VCI / VPI added to the network. The cell is IWU70G
After the VCI / VPI conversion is performed, the data is transferred to the IWU 70D. In addition, the cell is CLSF from IWU70D.
Transferred to 7011. CLSF70 that received the cell
11 is address information of the datagram (layer 3)
Is analyzed and the cell is transferred to the terminal 70A. In addition,
The transfer of the cell from the CLSF 7011 to the terminal 70A and the transfer of the cell from the CLSF-P 811 to the IWU 70G may be performed after the CLSF 7011 and the CLSF-P 811 have received all the cells belonging to the datagram. After the analysis, the cells may be relayed in a pipeline manner.

(Embodiment 6-2) An embodiment applied to a general ATM network will be described. In order to transfer a datagram from any terminal connected to each subnetwork to a terminal connected to any subnetwork, an ATM connection is established from each CLSF to CLSFs present in all subnetworks. I have. That is, the ATM connection between the CLSFs is set in a full mesh form. In an IWU (relay IWU) existing on the path of the ATM connection, the AT
Rewriting (at least VCI / VPI conversion) of the M header information is performed, and cell relaying is performed as a job of the ATM layer. That is, in principle, the ATM connection is not terminated when passing through the IWU. In addition,
The position where the CLSF exists may be the position of the IWU.

When a datagram is transferred from public networks 707 and 708 to the defined network, the received cells (to which the datagram belongs) in IWU 70K and IWU 70M temporarily terminate the datagram connection of the public network. Forwarded to server.
Note that this server may exist in the IWU. The server that terminates the datagram transferred from the public network relays the datagram in the same procedure as the transfer of the datagram from the terminal in the defined network.

FIG. 82 shows the configuration of an ATM connection when a datagram is transferred from the public network 707 to the terminal 70A. The datagram transferred from the public network 707 is once terminated in the CLSF-P811 and then the CLS
The data is transferred to the terminal 70A via F7031 and 7011.
The ATM connections are four connections 831, 832, 833 and 834.

(Regarding Terminal Protocol) When the terminal determines that the datagram is addressed to the external subnetwork, it transfers the datagram to the CLSF. The CLSF is basically located in the same sub-network as the terminal, but may be present in another sub-network such as, for example, an adjacent node. It is assumed that an ATM connection has already been set between the terminal and the CLSF. Each terminal has information (such as an address mask) of the address space of the subnetwork to which the terminal belongs,
It can be determined whether the destination terminal is in its own subnetwork or an external subnetwork.

In the case of the network configuration shown in FIG. 69, there are the following methods for datagram delivery.

(1) A star-shaped ATM connection is set between the CLSF and the terminal. When transferring datagrams, the terminal transfers all cells to the CLSF. All datagrams are delivered by CLSF. That is, the communication between the terminals in the sub-network is once C
Via LSF.

(2) Communication between terminals in the sub-network is realized without using the CLSF, while communication with terminals in the external network is realized via the CLSF.

(Regarding Routing) The final datagram delivery to each terminal is performed only for the network to which each CLSF belongs. That is, for example, CLSF7
Numeral 061 distributes datagrams to terminals belonging to the network 706. Similarly, the CLSF 7031 distributes datagrams only to terminals in the network 703. If the network address of the datagram received by each CLSF is not present in the address entry of the CLSF (or if the address of the received datagram is not an element of the network address space of the network), the datagram is It is determined that the delivery was incorrect. The actions for erroneous datagram delivery are not discussed here.

That is, each CLSF only needs to have the address information of the terminal of the network to which it belongs.
When the address of the received datagram exists in the own network, an appropriate ATM connection is selected and the datagram is relayed.

FIG. 74 shows the configuration of a connection when a datagram is transferred from terminal 70A to terminal 70B. The ATM connection is terminated at CLSF7011 and CLSF7061. That is, CLSF7
061 and CLSF7011 terminate the OSI layer 3 protocol. Layer 3 protocol processing is performed in the CLSF 7011, and the data unit is transferred to the IWU 70D using an ATM connection. CLSF7
At 061, layer 3 protocol processing is performed, and the data unit is transferred to the terminal 70B using the ATM connection. As described above, when transferring a datagram to a terminal other than the own subnetwork, datagram delivery is realized end-to-end at the end of two ATM connections.

[0665] Similarly, FIG.
08 shows a case where a datagram is transferred.

(Embodiment 6-2-1) Next, a more specific embodiment will be described.

(Regarding Datagram Transfer from Terminal 70A to Terminal 70B) FIG. 74 shows a configuration diagram of an ATM connection. When transmitting the datagram to the terminal 70B, the terminal 70A analyzes that the terminal 70B is a terminal belonging to the external sub-network.
Transfer the datagram to 1. The CLSF 7011 analyzes the address information of the received datagram, and
(CLSF7061), if it does not have the ATM layer address information for transferring the datagram,
The address resolution request cell having the address information of the terminal 70B is transferred to S. After receiving the AR request, the ARS performs address resolution, and then
V for transferring a cell to SF7061 (terminal 70B)
CLSF7011 using the CI / VPI information as an AR response
Return to.

ATM for transferring cells to terminal 70B
CL that has acquired layer address (VCI / VPI) information
The SF 7011 inputs the cell to which the VCI / VPI is added to the network. Cell is VC in IWU70D
After the I / VPI conversion, the data is transferred to the IWU 70G. Similarly, the cell is transferred to CLSF7061 via IWU70H and IWU70J. Upon receiving the cell, the CLSF 7061 analyzes the information of the address information (layer 3) of the datagram and transfers the cell to the terminal 70B. The cell transfer from the CLSF 7061 to the terminal 70B may be performed after the CLSF 7061 has received all the cells belonging to the datagram, or after analyzing the layer 3 address of the datagram, relaying the cells in a pipeline. You may.

(Regarding Datagram Transfer from Terminal 70A to Public Network 708) FIG. 75 shows a configuration diagram of an ATM connection. When transferring the datagram to the target terminal of the public network 708, the terminal 70A resolves the address of the destination terminal (analyzes the access address information of the subnetwork to which the destination terminal belongs). That is, when the terminal 70A cannot resolve the address information of the destination terminal, it transfers an AR request message containing the address information of the destination terminal to the address resolution server ARS. AR that received the AR request
After performing the address resolution, the S returns the access address information of the CLSF 7011 to the terminal 70A with the VCI / VPI information for transferring the cell to the destination terminal as an AR response. When analyzing that the destination is a terminal belonging to the external subnetwork, the terminal 70A receives the received VCI / VPI information or the VC obtained as a result of the analysis.
The I / VPI information is added to the cell, and the cell is
Transfer to 11.

[0670] The CLSF 7011 analyzes the address information of the received datagram, and if it does not have the ATM layer address information for transferring the datagram to the destination terminal, the address resolution having the address information of the destination terminal to the ARS. Transfer the requested cell. After receiving the AR request, the ARS performs the address resolution, and then transmits the VCI / V for transferring the cell to the IWU 70M.
The PI information is returned to the CLSF 7011 as an AR response.

CLS that has acquired ATM layer address (VCI / VPI) information for transferring a cell to a destination terminal
F7011 inputs the cell to which the VCI / VPI is added to the network. The cell is a VCI in IWU70D.
After the / VPI conversion is performed, the data is transferred to the IWU 70E. Similarly, the cell is connected to the public network 7 via the IWU 70M.
08.

(Regarding Datagram Transfer from Public Network 707 to Terminal 70A) FIG. 82 shows a configuration diagram of an ATM connection. The source terminal exists in the public network 707, and a cell related to connectionless communication from the public network 707 is transmitted to the CLSF-P811 via the IWU 70K.
Transferred to The VCI / VPI conversion table of the IWU 70K is set so that when a cell having a VCI / VPI assigned to a cell used for connectionless communication arrives in the public network 707, the cell is transferred to the CLSF-P811. ing.

The CLSF-P 811 temporarily terminates the ATM connection and analyzes the datagram layer 3 address information. The address information to be analyzed is CLSF-P81
If it does not exist in the table in 1, an AR request is issued to the ARS. A for transferring a cell to terminal 70A
TM layer address (VCI / VPI) information (cell is C
VCI / VPI information for transfer to LSF7031)
CLSF-P811 that has obtained VCI / VPI
The cell to which is added is input to the network. CLSF7
Numeral 031 analyzes the address information of the received datagram, and when it does not have the ATM layer address information for transferring the datagram to the terminal 70A (CLSF7011), the address resolution having the address information of the terminal 70A to the ARS. Transfer the requested cell. After receiving the AR request, the ARS performs address resolution, and then uses the VCI / VPI information for transferring the cell to the CLSF 7011 (terminal 70A) as an AR response as a CL response.
It returns to SF7031.

ATM for transferring cell to terminal 70A
CL that has acquired layer address (VCI / VPI) information
The SF 7031 inputs the cell to which the VCI / VPI has been added to the network.

[0675] After the VCI / VPI conversion is performed by the IWU 70G, the cell is transferred to the IWU 70D. Further, the cell is transferred from the IWU 70D to the CLSF 7011. Upon receiving the cell, the CLSF 7011 analyzes the information of the address information (layer 3) of the datagram,
The cell is transferred to the terminal 70A. Note that CLSF7011
Cell transfer from terminal to terminal 70A and CLSF-P81
1 to CLSF7031 and further CLS
Cell transfer from F7031 to IWU70G is performed by CLS
F7011, CLSF7031 and CLSF-P81
1 may receive all the cells belonging to the datagram, or may analyze the layer 3 address of the datagram and then relay the cells in a pipeline.

(Embodiment 6-3) A case where the present invention is applied to an ATM network of VPI routing.

(ATM connection) Each sub-network has an 8-bit sub-network identification number. Within the defined network, the subnetwork can be uniquely identified by this identification number. The identification number of the subnetwork is described as Net _i . For example, the identification number of the sub-network 702 is Ne
t ₇₀₂ . All IWUs have obtained access addresses (VPI) for connectionless communication. Also, the CLSF on the receiving side (a CLSF that handles cells arriving from the IWU, which can be configured differently from a CLSF that handles cells for connectionless communication coming from terminals in its own network) is also connectionless communication. Access address (VPI) has been acquired.

The coding of the VCI / VPI field is as follows.

(1) CLS of own network from transmitting terminal
Cell to F (1-1) VPI-F; CLSF access address (1-2) VCI-F1; Destination network address Ne
t _i or arbitrary (1-3) VCI-F2; access address of own terminal (2) Cell between CLSFs (2-1) VPI-F; access address of next access element (IWU or destination CLSF) (2 -2) VCI-F1; Address Ne of destination network
t _destnation (2-3) VCI-F1; Network address to which transmitting terminal belongs Net _source (3) Cell from CLSF of destination network to destination terminal (3-1) VPI-F; Access address of destination terminal (3) -2) VCI-F1; Can be set to any value. (3-3) VCI-F2: Can be set to any value. Note that it is possible to set any value to a cell received by the destination terminal. Can be set to an arbitrary value if it is set to a value that can be distinguished from a cell arriving from an arbitrary access point.

(Subnetwork where transmitting terminal exists)
First, the cell VCI / VPI from the transmitting terminal to the CLSF
Explain the coding of the field. VPI-F is
This is the CLSF access address. VCI-F2 is
Defined as the access address of the sending terminal. VCI-F1
Coding is performed when the transmitting terminal performs resolution of the subnetwork to which the destination terminal belongs, and when the CLSF
There are two cases in which the above is performed.

(A) The transmitting terminal performs resolution; the transmitting terminal resolves the address of the destination subnetwork, and identifies the identification number Net of the subnetwork.
It writes the _{destin ation} to the VCI-F1.

(B) Resolution by CLSF; Since resolution of the destination network is performed by CLSF, VCI-F1 can be set to an arbitrary value in this case.

Next, the VC of the cell from CLSF to IWU is
The I / VPI coding will be described. VPI-F
Is the access address of the IWU. VCI-F1
Is set to the identification number of the destination network (Net
_destination ). VCI-F2 is set to the identification number of the network of the transmitting terminal (Net _source ). VCI-
The setting method of F1 includes a case where the transmitting terminal performs resolution of the subnetwork to which the destination terminal belongs, and a case where the CLSF
There are two cases in which the above is performed.

(A) The transmitting terminal performs the resolution; the transmitting terminal copies the identification number Net _destination of the destination network written in the VCI-F1, and the VCI-F
Write to 1.

(B) CLSF performs resolution; writes the resolved value.

When transferring cells from the CLSF to the IWU, interleaving between cells belonging to different datagrams to the same destination network is not allowed (cell interleaving is possible for datagrams to different destination networks). That is, a series of cells belonging to one datagram are continuously transferred. That is, the cell is transmitted from each transmitting terminal at an arbitrary timing by CL.
Although cells can be transferred to the SF, cells are transferred from the CLSF to the IWU for each datagram. In the IWU, the VPI-F is based on the Net _destination written in the VCI-F1 of the received cell.
Determine the value of. That is, Ne in the IWU table
A table of VPI-F corresponding to t _destination is set.

[0687] VCI-F1 and VCI-F2 are transferred transparently. Also, Net
_{From the destination} , VPI-F, that is, a routing protocol for determining a subnetwork to be relayed, is separately executed, and a table for each IWU is set.

(Between IWUs) According to the IWU table,
The cell is relayed. That is, the VC of the receiving cell
The VPI-F for relaying the cell to the next IWU is written based on the Net _{destination of the} I-F1.

(Subnetwork to which destination terminal belongs) The IWU to which the destination terminal belongs is the VCI-F1 of the received cell.
And see that the received cell is addressed to its own subnetwork. In the table for VPI-F setting set in the IWU, a VPI number for transferring a cell to the CLSF is set. IWU has its VPI
Is set to VPI-F, and the cell is transferred to CLSF.
At this time, VCI-F1 and VCI-F2 are transparently transferred.

[0690] The received CLSF reassembles the datagram based on the VCI information. At this time, cells to which different datagrams belong are interleaved and CLSF
However, each datagram can be normally reconstructed using the information of the VCI field even when cells are interleaved. The destination address (layer 3 address) of the received datagram is analyzed, and the datagram is transferred to an appropriate terminal.

VP of cell transferred from CLSF to terminal
IF is the access address of the terminal. Next, VC
As the I field, a VCI defined by a cell received by the terminal for a cell for connectionless communication is used (a plurality of VCIs may be defined). This VCI is usually preset between the CLSF and the terminal. When there are a plurality of connectionless VCIs, cell transfer from the CLSF to the terminal can be transferred in a pipelined manner (cells to which different datagrams belong can be interleaved) as long as the identification numbers do not collide. it can.

For example, VCI-F1 and VCI-F2
Is defined in the communication from the CLSF to the terminal, that is, when the VCI field format of the communication in the sub-network is unified in this way, the VCI-F2 of the cell received by the CLSF is It is set as it is in VCI-F2 of the cell to be transferred from the CLSF to the terminal, and set in VCI-F1 of the cell from the CLSF to the terminal as a value that can recognize that the received cell is a cell involved in connectionless communication. If so, the datagram can be completely processed in a pipeline manner.

(Specific Example) First, transfer of a datagram from the terminal 70A to the terminal 70B will be described with reference to FIGS. It is formed from three ATM connections. (1) Connection 791 from terminal 70A to CLSF 7011, (2) CLSF 7011 to CLSF 706
Connection 792 to 796 to (1), CLSF70
A connection 797 from the terminal 61 to the terminal 70B.

The terminal 70A transfers a cell having VPI-F = VPI ₇₀₁₁ . Note that VCI-F2 is the terminal 70A
VPI _70A , which is the access address of the VPI. Also, V
When the terminal 70A resolves the address information (Net ₇₀₆ ) of the sub-network to which the destination terminal belongs by itself, the value Net ₇₀₆ is set to the VCI-F.
Written to 1. When the CLSF 7011 performs resolution, the VCI-F1 can be set arbitrarily.

The CLSF 7011 transfers the following cell to the IWU 70D. VPI-F is VPI _70D-1 which is an access address of IWU _70D , and VCI-F1 is Net ₇₀₆ , V which is an identification address of network 706.
CI-F2 is Net ₇₀₁ , which is the identification address of the network 701. When VCI-F1 copies the value of the received cell as it is (terminal analyzes Net ₇₀₆ )
In some cases, the CLSF 7011 analyzes and sets the Net ₇₀₆ .

[0696] The VCI- of the cell received by the IWU 70D-
F1 is analyzed and the corresponding VPI-F is analyzed. That is, the VPI that can transfer the cell toward the IWU that is to be transferred to the network 706 is analyzed (set in the table). VCI-F1 and V
The CI-F2 copies the field of the received cell as it is. That is, Net _706, which is the network identification number of the subnetwork 706 to which the destination terminal 70B belongs, is written in the VCI-F1. Thereafter, similarly, I
In the WU, an appropriate VPI is selected based on information of Net ₇₀₆ , which is information written in VCI-F1 of the received cell,
The cell is transferred to IWU 70J.

[0697] The cell arriving at the IWU 70J receives the VCI-
It recognizes that the cell has been transferred to the target network from the F1 information, and transfers the cell to CLSF7061. CL
The datagram arriving at the SF7061 is once subjected to cell reassembly, and the ATM connection is terminated. C
The LSF7061 analyzes the address information of the upper layer,
It recognizes that the destination of the datagram is terminal 70B. Therefore, the CLSF 7061 generates a cell and transfers the datagram to the terminal 70B. When the connectionless communication channel to the terminal 70B sufficiently exists, the cell can be transferred to the terminal 70B in a pipeline manner from the reception of the cell of the CLSF 7061.

Next, the transfer of a datagram from the terminal 70A to the public network 708 will be described with reference to FIG.
The terminal 70A transfers a cell having VPI-F = VPI ₇₀₁₁ . Note that VCI-F2 is a VPI _70A which is an access address of the terminal 70A. Also, when the terminal 70A itself resolves the address information (Net ₇₀₈ ) of the sub-network to which the destination terminal belongs, this value Net ₇₀₈ is written to the VCI-F1. When the CLSF 7011 performs resolution, the VCI-F1 can be set arbitrarily.

The CLSF 7011 transfers the following cell to the IWU 70D. VPI-F is VPI _70D-1 which is an access address of IWU _70D , and VCI-F1 is Net ₇₀₈ , VCI-F which is an identification address of public network 708.
F2 is the identification address of the network 701, Net
₇₀₁ . When the value of the received cell is copied as it is (terminal analyzes Net ₇₀₆ ), VCI-F1
There are cases where the LSF 7011 analyzes and sets the Net ₇₀₈ . VCI-F1 of cell received by IWU 70D
Is analyzed, and the corresponding VPI-F is analyzed. That is, the VPI that can transfer the cell toward the IWU to be transferred to the network 708 is analyzed (set in the table). VCI-F1 and VCI
-F2 copies the field of the received cell as it is. That is, VCI-F1 is the network identification number of the subnetwork 708 to which the destination terminal belongs, N
et ₇₀₈ has been written. Thereafter, similarly, in the IWU, Ne, which is information written in the VCI-F1 of the received cell, is used.
Suitable VPI information based on the t ₇₀₈ is selected, the cell I
Transferred to WU70M. The IWU 70M calculates the AT defined by the public network 708 from the received VCI information of the cell.
The VCI / VPI assigned to the M connection is written, and the cell is transferred to the public network 708.

(Seventh Embodiment) Next, an embodiment relating to connectionless communication in a large-scale networking configuration will be described.

When the number of defined sub-networks is very large, it can be handled by internetworking the above-mentioned networks. That is, this is a method in which an adjacent network (a network defined as an aggregate of a plurality of sub-networks described above) is viewed as one sub-network.

FIG. 84 shows the configuration of the network as viewed from the network 861. The actual network configuration is shown in FIG. Thus, the three networks 862 to 864 are seen as one sub-network 851 when viewed from the network 861. Network 8
As for the address space of the network 851 from the viewpoint of 61, a space combining the networks 862 to 864 can be seen.

FIG. 86 shows the simplified network components involved in the following datagram transfer among the network components. In the following, 2 described earlier
An example of datagram transfer from the terminal 87A to the terminal 87B and datagram transfer from the terminal 87A to the terminal 87C will be described with one system configuration.

(Embodiment 7-1) This embodiment is an example in which a terminal can directly transfer a cell to an external network. That is, this method does not use the CLSF in its own network when transferring datagrams to terminals on the external network.

(Regarding Datagram Transfer from Terminal 87A to Terminal 87B) FIG. 87 shows the configuration of an ATM connection. These are the four ATM connections 881-8
84 are configured. The terminal 87A resolves the network layer address of the terminal 87B, and
V for resolving that B belongs to the network 851 and at the same time transferring cells to the CLSF 871
Obtain CI / VPI information. Details are as described above.

The datagram is temporarily terminated by the CLSF871 (the ATM connection is also terminated),
Is performed. The analysis of the network layer address is performed by the CLSF 871, the presence of the terminal 87B in the network 863 is analyzed, and
VCI / VPI information for transferring a cell to the LSF 872 is obtained. Datagram is ATM connection 8
82 to the CLSF 872.

[0707] The datagram is terminated by the CLSF 872 (the ATM connection is also terminated), and the protocol processing of the layer 3 is performed. The analysis of the network layer address is performed by the CLSF 872, and VCI / VPI information for transferring the cell to the CLSF 873 is obtained.
Datagram is CL using ATM connection 883
Transferred to SF873.

The CLSF873 that received the datagram
Analyzes the access address of the terminal 87B by analyzing the network layer address of the datagram. The cell to which the appropriate VCI / VPI has been added is the terminal 8
7B.

As described above, the termination of the ATM connection and the protocol processing in the network layer are performed three times, and the datagram is transferred from the terminal 87A to the terminal 87B.

(Regarding Datagram Transfer from Terminal 87A to Terminal 87C) FIG. 88 shows the configuration of an ATM connection. This is two ATM connections 891,8
92. The terminal 87A resolves the network layer address of the terminal 87C, and
7C resolves that it belongs to the network 851, and at the same time, obtains VCI / VPI information for transferring the cell to the CLSF 871. Details are as described above.

The datagram is temporarily terminated by the CLSF 871 (the ATM connection is also terminated),
Is performed. The analysis of the network layer address is performed by the CLSF 871, and it is analyzed that the terminal 87C exists in the network 864.
VCI / V for transferring cells to network 864
PI information is obtained. Here, the network 862
The transfer of the cell to the IWU between the network and the network 864 is performed.

As described above, one ATM connection termination and protocol processing in the network layer are performed, and the datagram is transferred from the terminal 87A to the network 864 in which the terminal 87C exists. Note that CLSF
The transfer destination of the datagram from 871 may be CLSF in public network 864. In this case, two or more A
Termination of the TM connection and protocol processing at the network layer level are performed.

(Embodiment 7-2) This embodiment is an example in which a terminal cannot directly transfer a cell to an external network. That is, when a datagram is transferred to a terminal on an external network,
This is a method using LSF once.

(Regarding Datagram Transfer from Terminal 87A to Terminal 87B) FIG. 89 shows the configuration of an ATM connection. This is the five ATM connections 901-9
05. The terminal 87A resolves the network layer address of the terminal 87B, and
7A resolves that it belongs to the network 851 (or analyzes that the cell is sent to the CLSF 874), and at the same time, obtains VCI / VPI information for transferring the cell to the CLSF 874. Details are as described above.

The CLSF 874 having received the cell sets (1)
Analyzing the network layer address of the datagram, or (2) using the result analyzed by the terminal 87A, recognizes that the cell needs to be transferred to the network 862 (CLSF 871). At the same time, VCI / VPI information for transferring a cell to the CLSF 871 is obtained, and the datagram is transferred to the CLSF 871.

The datagram is temporarily terminated by the CLSF871 (the ATM connection is also terminated),
Is performed. The analysis of the network layer address is performed by the CLSF 871, the presence of the terminal 87B in the network 863 is analyzed, and
VCI / VPI information for transferring a cell to the LSF 872 is obtained. Datagram is ATM connection 9
03 to the CLSF872.

The datagram is terminated by the CLSF 872 (the ATM connection is terminated), and the layer 3 protocol processing is performed. The analysis of the network layer address is performed by the CLSF 872, and VCI / VPI information for transferring the cell to the CLSF 873 is obtained.
Datagram is CL using ATM connection 904
Transferred to SF873. C that received the datagram
The LSF 873 analyzes the network layer address of the datagram, and analyzes the access address of the terminal 87B. The cell to which the appropriate VCI / VPI has been added is transferred to terminal 87B.

As described above, the termination of the four ATM connections and the protocol processing (3
The datagram is sent to the terminal 87A.
Is transferred to the terminal 87B.

(Regarding Datagram Transfer from Terminal 87A to Terminal 87C) FIG. 90 shows the configuration of an ATM connection. These are the three ATM connections 911-9
13. The terminal 87A resolves the network layer address of the terminal 87C, and
7C resolves (or analyzes sending a cell to CLSF874) that it belongs to the network 851, and at the same time, obtains VCI / VPI information for transferring the cell to CLSF874. Details are as described above.

The CLSF 874 that has received the cell sets (1)
Analyzing the network layer address of the datagram, or (2) using the result analyzed by the terminal 87A, recognizes that the cell needs to be transferred to the network 862 (CLSF 871). At the same time, VCI / VPI information for transferring a cell to the CLSF 871 is obtained, and the datagram is transferred to the CLSF 871.

The datagram is temporarily terminated by the CLSF 871 (the ATM connection is also terminated),
Is performed. The analysis of the network layer address is performed by the CLSF 871, the presence of the terminal 87C in the network 864 is analyzed, and the VCI / VP for transferring a cell to the network 864 is further analyzed.
I information is obtained. Here, cell transfer to the IWU between the network 862 and the network 864 is performed.

As described above, the termination of two ATM connections and the protocol processing (1
The datagram is sent to the terminal 87A.
Is transferred to the network 864 where the terminal 87C exists. Note that the transfer destination of the datagram from the CLSF 871 may be a CLSF in the public network 864. In this case, two or more terminations of the ATM connection and protocol processing at the network layer level are performed.

[0723]

As described above, according to the present invention, A
Communication between ATM networks can be realized without impairing the characteristics of the TM communication system such as high speed, large capacity, and connection orientation.

According to the present invention, datagram delivery using the ATM network can be performed efficiently.

Further, according to the present invention, connectionless communication (datagram delivery) between terminals connected to the ATM network can be performed at high speed.

[Brief description of the drawings]

FIG. 1 is a diagram showing an ATM network according to a first embodiment.

FIG. 2 is a diagram showing the internal structure of the network connection device 13;

FIG. 3 shows a drop table.

FIG. 4 is a diagram showing an internal configuration of an ATM-LAN;

FIG. 5 shows a terminal in an ATM-LAN and a C in an inter-network connecting device.
The figure which shows the connection state of the ATM connection between LSF processing parts

FIG. 6 is a diagram showing another CLS in the ATM-LAN in FIG.
Diagram showing a case where an F processing unit is prepared

FIG. 7 is a diagram showing a format of a broadcast cell;

FIG. 8 is a diagram showing a format of a broadcast cell when performing ARP.

FIG. 9 is a diagram showing a connection state of an ATM connection between a terminal in the ATM-LAN and a call processing unit in the network connection device.

FIG. 10 is a diagram showing another example of the connection state of the ATM connection between the terminal in the ATM-LAN and the call processing unit in the network connection device;

FIG. 11 is a diagram showing an example in which a call processing unit does not exist in the network connection device;

FIG. 12 shows various forms of broadcasting.

FIG. 13 is a diagram showing an ATM network according to a second embodiment;

FIG. 14 is a diagram showing an internal structure of the network connection device 134;

FIG. 15 is a diagram showing an internal configuration of an ATM-LAN;

FIG. 16 is a diagram showing a connection state of an ATM connection between a terminal in the ATM-LAN and a CLSF processing unit in the network connection device;

FIG. 17 is a diagram showing a connection state of an ATM connection between a terminal in the ATM-LAN and a call processing unit in the network connection device.

FIG. 18 is a diagram showing another example of the connection state of the ATM connection between the terminal in the ATM-LAN and the call processing unit in the network connection device.

FIG. 19 is a diagram showing an example in which a call processing unit does not exist in the network connection device;

FIG. 20 is a diagram showing a large-scale ATM network;

FIG. 21 shows ATM- in a large-scale ATM network.
Diagram showing the internal configuration of the LAN

FIG. 22 is a diagram showing an example of VPI value allocation in an ATM-LAN.

FIG. 23 is a diagram showing an example of an ARP cell format.

FIG. 24 is a diagram showing an example of a network layer protocol identification method.

FIG. 25 is a diagram showing an example of a cell format when an ARP response using a broadcast channel is performed using an end-to-end ATM connection.

FIG. 26 is a diagram showing an example of the flow of ARP.

FIG. 27 is a diagram showing a correspondence table between network layer addresses and ATM addresses (VPI values) in an ARP server;

FIG. 28 shows a CLSF processing unit and an ATM-L in the network connection device;
ATM connection state between terminals in AN, and AT
ATM between the CLSF processing units in the M backbone network
Diagram showing an example of the connection state of the connection

FIG. 29 is a diagram showing an example (phase 1) of the arrangement method of the CLSF processing units at the time of initial introduction;

FIG. 30 is a diagram showing a header value added to an ATM cell of a datagram from the network connection device to the CLSF processing unit;

FIG. 31 is a diagram showing a flow of header conversion.

FIG. 32 is a diagram showing an example (phase 2) of a CLSF processing unit arrangement method when a CLSF processing unit is added.

FIG. 33 shows a table in the ARP server.

FIG. 34 is a diagram showing a connection state of an ATM connection between a call processing unit in the network connection device and a terminal in the ATM-LAN, and a connection state of an ATM connection between the call processing units in the ATM backbone network.

FIG. 35 is a diagram showing another example of a call processing method in a large-scale network.

FIG. 36 shows an example of an ATM board.

FIG. 37 is a diagram showing an example of an ARP cell in VP routing.

FIG. 38 is a diagram showing an example of an internal configuration of the IWU 13;

FIG. 39 is a diagram showing another example of the internal configuration of the IWU 13;

FIG. 40 is a diagram showing a datagram communication system in a subnetwork.

FIG. 41 shows a datagram communication sequence.

FIG. 42 is a flowchart showing a datagram transmission procedure.

FIG. 43 shows a datagram communication system in a sub-network.

FIG. 44 shows a datagram communication sequence.

FIG. 45 is a diagram showing VPI routing.

FIG. 46 is a diagram showing a two-layer network.

FIG. 47 is a diagram showing a connection between ARs.

FIG. 48 is a diagram showing an ATM connection required for datagram communication.

FIG. 49 shows an address space view from ARS481.

FIG. 50 shows an address space view from ARS482.

FIG. 51 is a diagram showing an ATM connection between ARSs;

FIG. 52 shows an address space view from ARS482.

FIG. 53 shows an address space view from ARS481.

FIG. 54 is a view showing a VCI / VPI conversion table in the IWU 476;

FIG. 55 shows a datagram transfer protocol process.

FIG. 56 is a diagram showing datagram transfer from the terminal 47A to the terminal 47D.

FIG. 57 is a diagram showing datagram transfer from the terminal 47A to the terminal 47D.

FIG. 58 is a diagram showing an example of a VCI / VPI allocation method between ARSs;

FIG. 59 is a diagram showing another example of a VCI / VPI allocation method between ARSs.

FIG. 60 is a diagram showing another example of a VCI / VPI allocation method between ARSs.

FIG. 61 is a diagram showing a VCI / VPI assignment method for datagram transfer;

FIG. 62 is a diagram showing a network configuration.

FIG. 63 is a diagram showing an ATM connection required for datagram transfer;

FIG. 64 is a diagram showing an ATM connection between ARSs;

FIG. 65 is a diagram showing an ATM connection between ARSs;

FIG. 66 is a diagram showing an ATM connection required for datagram transfer;

FIG. 67 shows a datagram transfer protocol process.

FIG. 68 shows an example of datagram transfer.

FIG. 69 shows a network configuration.

FIG. 70 is a diagram showing a datagram communication example of terminals 70A → 70B.

FIG. 71 is a diagram showing protocol processing.

FIG. 72 shows an example of datagram communication between terminals 70A → 70B.

FIG. 73 shows an example of datagram communication from terminal 70A to public network 708.

FIG. 74 is a diagram showing a datagram communication example of terminals 70A → 70B.

75 shows an example of datagram communication from terminal 70A to public network 708. FIG.

FIG. 76 is a diagram showing an ATM connection between terminals 70A and 70B.

FIG. 77 is a diagram showing a datagram communication example from the terminal 70A to the public network 708.

FIG. 78 is a diagram showing an ATM connection between terminals 70A and 70B.

FIG. 79 is a diagram showing a datagram communication example from the terminal 70A to the public network 708.

FIG. 80 is a diagram showing a datagram communication example between the public network 708 and the terminal 70A.

FIG. 81 is a view showing protocol processing of the public network 708 → terminal 70A.

FIG. 82 is a diagram showing a datagram communication example between the public network 708 and the terminal 70A.

FIG. 83 is a view showing a protocol processing of the public network 708 → terminal 70A.

84 is a diagram showing an image from a network 861. FIG.

FIG. 85 shows a network configuration.

FIG. 86 shows a network configuration.

FIG. 87 is a diagram showing a datagram communication example of terminals 87A → 87B.

FIG. 88 shows a datagram communication example of terminals 87A → 87C.

FIG. 89 is a diagram showing a datagram communication example of terminals 87A → 87B.

FIG. 90 shows an example of datagram communication between terminals 87A → 87C.

[Explanation of symbols]

11, 12 ATM-LAN 13 IWU (inter-network connection device) 21 add / drop processing unit 22 multiplexer / demultiplexer 23 CLSF processing unit (connectionless service function processing unit) 24 call processing unit 25 IWU Management unit 26 Header conversion processing units 41 to 44 Switch nodes 4A to 4G Terminals 91 and 92 Call processing units 101 and 102 Call processing units 111 and 112 ATM-LAN 11A and 11B Call processing units 131 to 133 ... ATM-LAN 134 ... IWU (inter-network connection device) 141 ... ATM switch 142 ... CLSF processing unit (connectionless service function processing unit) 143 ... Call processing unit 144 ... IWU management unit 14A, 14B ... Input processing unit 14X, 14Y ... Output processing units 151-157 ... Switch nodes 15A-15J ... End 171 to 173 call processing unit 181 to 183 call processing unit 191 to 193 ATM-LAN 201 ATM backbone network 202 to 204 ATM-LAN 20A to 20C IWU (inter-network connection device) 211 to 217 switch node 21A-21J terminal 311 ATM-LAN 312 ATM backbone network 31A-31B terminal 31P IWU (inter-network connection unit) 31X CLSF processing unit (connectionless service function processing unit) 361 ATM interface 362 ARP filter Unit 363 ARP processing unit 364 Insert unit 365 Bus interface 411 412 Terminal 413 ARS (address resolution server) 414 CLSF processing unit (connectionless service function processing unit) 415 IWU (network connection unit) 41A to 41C ATM connection 46A to 46E ATM connection 471 to 474 Subnetwork (ATM network) 475 Public network 476 to 479 IWU (internetwork connection device) 47A to 47F Terminal 481 to 484 ARS (address) Resolution server) 485-487 ATM connection 491-494 CLSF processing unit (connectionless service function processing unit) 495-49B ATM connection 571-576 ATM connection 581-584 ATM connection 591-59C ATM connection 60xx ATM connection 611-61A ATM connection 621-62A ATM connection 63A-63B Terminal 631-634 Subnetwork (ATM network) 635 Public 636 to 639: IWU (inter-network connection device) 63D to 63G: ARS (address resolution server) 63H to 63L: CLSF processing unit (connectionless service function processing unit) outside the subnetwork 63M to 63Q: inside the subnetwork CLSF processing unit (connectionless service function processing unit) 641 to 646 ATM connection 651 to 656 ATM connection 661 to 66C ATM connection 691 to 699 ATM connection 701 to 706 Subnetwork (ATM network) 707, 708 Public Network 70A, 70B terminal 70xx CLSF processing unit (connectionless service function processing unit) 70E-70M IWU (network connection device) 771-776 ATM connection 781-784 ATM connector ATM connection 801 to 805 ATM connection 811 CLSF processing unit (connectionless service function processing unit) 812 to 814 ATM connection 831 to 834 ATM connection 851 Subnetwork (ATM network) 861 to 863 ... Sub-network (ATM network) 864 ... Public network 871-874 ... CLSF processing unit (connectionless service function processing unit) 87A-87C ... Terminals 881-884 ... ATM connection 891-892 ... ATM connection 901-905 ... ATM connection 911 913… ATM connection

Claims (8)

[Claims] A plurality of ATs operated in an asynchronous transfer mode
An M network, a plurality of terminals respectively accommodated in the plurality of ATM networks, and an inter-network connection device provided between the plurality of ATM networks and performing internetworking between the ATM networks. The inter-network connection device does not perform processing of the ATM adaptation layer or higher for at least some of packets transmitted from one ATM network connected to the inter-network connection device to another ATM network. At least ATM cell header reference and ATM in layer processing
An ATM communication system for performing at least one of cell switching. 2. A plurality of ATs operated in an asynchronous transfer mode.
An M network, a plurality of terminals respectively accommodated in the plurality of ATM networks, and an inter-network connection device provided between the plurality of ATM networks and performing internetworking between the ATM networks. An ATM communication system characterized in that the network connection device has connectionless service processing means. 3. A plurality of ATs operated in an asynchronous transfer mode
An M network, a plurality of terminals respectively accommodated in the plurality of ATM networks, and an inter-network connection device provided between the plurality of ATM networks and performing internetworking between the ATM networks. An AT characterized in that the network connection device has call processing means, and the call processing means performs at least call / connection processing across the network connection device having the call processing means.
M communication system. 4. A plurality of ATs operated in an asynchronous transfer mode
An M-LAN, a plurality of terminals respectively accommodated in the plurality of ATM-LANs, a datagram server provided in at least one of the plurality of ATM-LANs, and an accommodation included in the same ATM-LAN Connection setting means for setting an end-to-end ATM connection between the plurality of terminals; and datagram delivery between the plurality of terminals accommodated in the same ATM-LAN, by the connection setting means. An ATM communication system comprising: a datagram delivery unit that performs the datagram delivery between the plurality of terminals in the different ATM-LAN through the ATM connection, and performs the datagram delivery through the datagram server. 5. A plurality of ATs operated in an asynchronous transfer mode
M network; a plurality of terminals respectively accommodated in the plurality of ATM networks; a plurality of datagram servers provided at least one in the plurality of ATM networks as a whole; and a plurality of datagram servers provided between the plurality of ATM networks. , An internetworking device that performs internetworking between the ATM networks, wherein the internetworking device is an ATM connected to the internetworking device.
A for requesting address resolution from the network
If the target node of the address resolution does not exist in the ATM network connected to the network connection device in response to the RP request, a reply to the ATM address of the ATM connection connected to the network connection device is sent to the ARP for the ARP request.
A response means for performing an RP response, and an AT returned by an ARP response from the response means
An ATM communication system comprising: a first ATM connection indicated by an M address; and connection connection means for connecting a second ATM connection set between the network connection device and the datagram server. 6. A plurality of ATs operated in an asynchronous transfer mode
M network; a plurality of terminals respectively accommodated in the plurality of ATM networks; a plurality of datagram servers provided at least one in the plurality of ATM networks as a whole; and a plurality of datagram servers provided between the plurality of ATM networks. An internetworking device for performing internetworking between the ATM networks, the internetworking device comprising: a routing protocol termination unit; and an address resolution from a first ATM network connected to the internetworking device. Notification means for notifying an ARP server for performing address resolution when receiving an ARP request for requesting information about routing, and the address resolution for the information passed by the notification means. A represented by the ATM address returned by the address resolution response from the ATM communication system; and a connection connecting means for connecting the second ATM connection is set up between M connections and net-connecting device, the datagram server. 7. The datagram server, wherein the connection connection means connects the first ATM connection and the second ATM connection by at least one of an ATM cell header conversion means and an ATM cell switching means. 7. The ATM communication system according to claim 5, wherein the addition, subtraction, or change of the number is performed by setting at least one of the ATM cell header conversion unit and the ATM cell switching unit. 8. An ATM network operated in an asynchronous transfer mode, comprising: a plurality of terminals accommodated in the ATM network; wherein the terminals have addresses addressed to the terminals from ATM cells from the ATM network. ARP request cell extracting means for extracting an ARP request cell for requesting a resolution, and an ARP response for responding to the address resolution request based on the ARP request cell extracted by the ARP request cell ARP response cell generating means for generating a cell, and multiplexing means for multiplexing the ARP response cell generated by the ARP response cell generating means with an ATM cell stream transmitted from the terminal device. ATM
Communications system.