JP2001251320A - Burst data communication system in atm network - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a burst data communication system for an ATM network that needs no server for address resolution, and converts an Ethernet address into an ATM address for the communication. SOLUTION: In a network system where terminal units 30 employing the ATM transmission system and accommodating at least burst data are interconnected in a mesh, when transmission lines are configured to be a mesh, an ATM switch panel mounted in an adjacent terminal unit 30 sets a VP in advance and a burst data interface panel in each terminal unit 30 sets the VP while preventing a communication path from being a loop in order that transmitted data are not received in a plurality of directions.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a burst data communication system in an ATM network. Here, the burst data refers to packet data of the Ethernet (registered trademark) type LAN. More specifically, the present invention relates to a system for accommodating packets generated in bursts from an Ethernet terminal in an ATM (asynchronous transfer mode switch device) for communication.

[0002]

2. Description of the Related Art Conventionally, when storing burst data in an ATM network, it has been realized by using LAN emulation technology. The LAN emulation emulates existing Ethernet and token ring LAN applications on an ATM network, and provides an end system with an interconnection with the existing LAN without being aware of the ATM. Here, the Ethernet refers to a local LAN in an office, and the token ring refers to a system that performs communication via a ring called a token.

[0003] The MAC (M
This is a method for performing connection at the media access control (media access control) level. When sending LAN data to an ATM, it is necessary to convert (address resolution) to an ATM address. In LAN emulation, the processing is performed by a server.

A device for connecting an existing LAN environment on the Ethernet side to an ATM environment is provided to a LAN emulation server by an LEC (LAN emulation server).
Client).

[0005] Servers are classified into the following three types according to the functions to be provided.

[0006] LECS (LAN emulation
LECS is an ELAN containing the LES ATM address
The (LAN emulated by LAN emulation) information is notified to the LEC. LEC is EL
When participating in AN, the first access is LECS
It is.

LES (LAN Emulation Server) LES is a server that the LEC accesses according to the ATM address notified from the LECS. In the LES, the MAC address of the LEC (the frame address of the Ethernet) and the ATM address are registered.
An ARP (address resolution protocol) function for resolving an ATM address from an address is provided. Here, the ARP is a protocol for searching for an address by referring to a table when an inquiry about the address is made.

BUS (Broadcast / Unknown Server) BUS is a multicast (selective broadcast) on the ELAN,
Broadcast (broadcast all) and send / receive data. The BUS takes over transmission and reception of broadcast data to an unspecified partner.

FIG. 8 is a conceptual diagram of a conventional network system. The figure shows an example in which each LAN emulation server is mounted on an ATM switch. In the figure, 1
Is an existing terminal (Ethernet terminal: client), 2 is an LEC (ATM bridge) to which the existing terminal 1 is connected, and 10 is an ATM switch which is connected to the LEC 2 and performs switching. That is, each existing terminal 1
Is connected to another existing terminal 1 via an ATM switch 10. In the figure, 3 is LECS,
4 is LES, 5 is BUS, and both are ATM switches 10
It is provided within. In the LES 4, an ATM address corresponding to the MAC address is registered. The LAN emulation operation will be described using such a system as follows.

(Step 1) Initialization First, the client accesses LECS3, and
Acquire the ATM address of S4. Here, LECS3
Access to the default well-known address.

(Step 2) Connection setting to server Each client uses the ATM address of the previously acquired LES for control with the server in order to receive a service from the LAN emulation server (LES) that the client wants to participate in. Set up a connection. Also, a connection is set in BUS5 in preparation for broadcasting. First, AT from broadcast address
An ARP operation is performed on the M address to acquire the address of BUS5. The connection is set to BUS5 with the address.

(Step 3) Address registration Next, each client registers its own MAC address and AT
Register the M address in the server (LES4). here,
If the client is a directly connected terminal, the MAC address of the client itself is registered. ATM bridge (LE
C2), the proxy (P
(roxy: proxy) operation, and registers at least one MAC address of a terminal under the bridge.

(Step 4) Address Resolution When there is a frame to be transferred, the client checks the target MAC address and resolves the ATM address of the target client to set up a connection up to the ATM client corresponding to the MAC address. (LE
-ARP). The client sends an LE-ARP request to the server using the control connection.

If the ATM address of the target client corresponding to the target MAC address is registered in the server 4, the server returns the address to the client. For a MAC address that is not registered, the server broadcasts an ARP request packet to all clients over a multipoint connection. In this case, the corresponding receiving client returns its ATM address as a response to the ARP request via the server. The LE-ARP information thus obtained is cached (held) on each client side.

FIG. 9 is a diagram showing an example of a frame format of packet data. In the figure, reference numeral 20 denotes a preamble (synchronous sequence) having a length of 7 bytes, reference numeral 21 denotes a 1-byte length of SFD (frame start code), reference numeral 22 denotes a DA (data transmission destination address) of 6 bytes, and reference numeral 23 denotes an SA (data transmission source address). ) Is 6 bytes long, 24 is DATA (data)
Is a length of 48 bytes to 1502 bytes, and 25 is a 4-byte length of FCS (Error Detection Code).

(Step 5) Data Transfer When the client knows the ATM address of the client corresponding to the target MAC address obtained in this way, it sets up a data connection to the client. Once the data connection is established, a correspondence table between the connection and the MAC address is created for each client, and thereafter, the frame of the MAC address is transferred through the ATM connection.

All frames arriving while the data connection has not been set (LE-ARP / connection setting procedure) are all unknown, and are transferred to the BUS5. A frame is sent from the client to BUS5 by a point-to-point connection, and from BUS5, all clients are broadcast by a point-to-multipoint connection.

After the LE-ARP, when a data connection is set up, frame transfer starts directly between the clients on the connection. In this way, the frame is first transferred to BUS5, and then switched to data connection transfer. At this time, in order to prevent the frame order from being out of order, the transmitting client uses a message to notify the receiving client that the transfer via the BUS5 is completed when the data connection is established.

On the other hand, the transfer of a frame having a multicast MAC address is performed point-to-point by BUS.
5, a data frame is first transferred and broadcast to all clients from BUS5 through a point-to-multipoint connection.

From the above, when accommodating an existing LAN in an ATM network by LAN emulation, a server for address resolution is required.

[0021]

In the prior art, as described above, a server for resolving an ATM address from an Ethernet address was indispensable for the system. Conventionally, it is necessary to install a plurality of servers (LECS, LES) on a network. Then, in order to know the ATM address of the LES, each client needs to inquire the LECS at initialization.

Each client needs to set up a control connection with the LES. Each client needs to register its own MAC address and ATM address in the server. Also, each client
If you do not know the desired ATM address, ARP to LES
You need to make a request and resolve the address. In this way, a failure of a specific node affects the entire system.

Further, since the LES centrally manages the ATM address of the target client corresponding to the target MAC address, if a failure occurs in the LES, the address cannot be resolved when a new connection is established.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and does not require a server for performing address resolution, and can convert an Ethernet address into an ATM address and perform burst data communication. It is intended to provide a communication system.

[0025]

(1) FIG. 1 is a block diagram showing the principle of the present invention. In the figure, reference numeral 30 denotes a terminal device (hereinafter referred to as a node) to which terminals are connected and which performs data transfer, and is provided in a mesh as shown in the figure.
Each node is connected by a virtual path (VP: virtual path). These VPs are represented by VP1 to VP12 as shown in FIG.

Of these, VP1, VP2, VP3, VP
4, VP5, VP8, VP9, and VP10 are indicated by bold solid lines, and indicate valid open VPs. Other VPs VP6, VP7, VP11, and VP12 indicate closed invalid VPs. By arranging the VPs in this way, a loop does not occur in the data transfer system.

That is, N: N between any Ethernet terminals
ATM switch board (ATM-SW board) between adjacent nodes 30 on the ATM network to realize communication
VP is set in advance. Further, a VP is set so that a communication path does not form a loop in a burst data interface board (hereinafter abbreviated as a burst data I / F board) in the terminal station device so that transmitted data is not received from a plurality of directions. I have.

With this configuration, even if the server is not installed on the network, the burst data can be communicated from the initial state in which the VP is set on the ATM network in advance.

(2) According to the second aspect of the present invention, when the connection between the communicating terminals is not yet set, the AT is set so that the data does not form a loop.
It is characterized in that the M VP and the packet switch section in the burst data I / F board are switched and broadcast in the network.

With this configuration, LECS, LE
Even without a specific server such as S or BUS, address learning can be performed by address learning on the packet data interface board and broadcast communication in the network, thereby enabling burst data communication.

(3) The invention according to claim 3 has a configuration in which address learning is distributed in each burst data I / F board, and an OAM cell is always provided for a working / backup VP between adjacent terminal stations. Are transmitted and received, the normality of the network is monitored, and if the VP is disconnected due to a network failure, switching to a backup VP is performed.

With this configuration, when a failure occurs in the working connection, a communication path can be secured by switching to a spare VP whose normality has been confirmed in the OAM cell. In addition, since the address learning is decentralized, there is no influence on the entire network such as the failure of address learning due to the occurrence of a failure.

[0033]

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

First, the operation of the present invention will be described with reference to FIG.

A first feature of the present invention is that, in order to realize N: N communication between arbitrary Ethernet terminals, a VP is set in advance in an ATM switch panel (described later) between adjacent nodes on an ATM network. The communication path does not form a loop in a burst data I / F board (described later) so that transmitted data is not received from a plurality of directions.

The connection between the burst data I / F boards mounted on the node 30 on the ATM network is set in advance between adjacent nodes by VP. The setting procedure is performed by using a switching table of an ATM switch panel at the time of setting a node, for example, as configuration information from a CPU panel mounted on each node and monitoring and controlling the node. Since the connection becomes a loop in the entire network by setting all the VPs between the adjacent nodes, the VP is enabled / disabled so that the connection does not form a loop in the burst data I / F board.

When communication is performed with the connection in a loop state, the transmitted packet data is received at each node from a plurality of optical transmission lines, and the data is transmitted again to the plurality of optical transmission lines by broadcast. Broadcast data circulates on the network and does not end. In order to prevent such a state from occurring, the VP is set so that the connection does not form a loop as shown by a thick solid line in FIG. That is, each node 30 does not constitute a loop.
In the burst data I / F board, only the thick line VP is valid.

According to this embodiment, communication of burst data becomes possible from an initial state in which a VP is set in advance on an ATM network without installing a server on the network.

FIG. 2 is a block diagram showing a detailed configuration example of a node. 1 are denoted by the same reference numerals. In the figure, the detailed configuration of the node 2 is shown. In the figure, reference numeral 40 denotes a burst data I / F board for preventing transmitted data from being received from a plurality of directions, that is, for preventing a communication path from forming a loop.
/ F panel 40 connected to an ATM switch panel (ATM-S
W board) and 32 are relay interface (I / F) boards connected to the ATM switch board 31. Relay I / F board 3
Reference numeral 2 denotes an interface for connecting to each node 30, which is a VP connection unit.

In the burst data I / F board 40, 4
1 is a packet switch (SW) unit for switching a packet, 42 is a learning table unit for storing switching information by learning, 43 is a data processing unit connected to the packet switch unit (packet SW unit) 41, and 44 is a data processing unit The connection between the unit 43 and the ATM switch panel 31 is performed to convert an Ethernet packet into an ATM cell or to perform an ATM conversion.
This is an SAR unit that converts cells into Ethernet packets. The operation of packet data switching and address learning in a node using the system configured as described above will be described below.

Node 1 for which no connection has been set
Packet data transmitted from terminal A (see FIG. 9)
Sends A to VP1 and VP3 that are set to be valid by node 1.
Sent after TM cell conversion. The node 2 receives the data received by the VP 1 via the relay I / F board 32. ATM
The switch board 31 changes the VP value of the received data (VP1 → VP1 + n), and outputs the burst data I /
Switching to F board 40 is performed.

In the burst data I / F board 40, the SAR
The unit 44 performs the cell → packet conversion, and the data processing unit 4
The data is given to the packet switch unit 41 via the communication unit 3.
Here, the switching information (D
A: There is no destination address, so it is broadcast to ports other than the input port. At this time, the source address (SA) of the terminal A is stored in the learning table section 42 corresponding to the input port of the packet switch section 41.

After the broadcasted packet data is converted into cells by the SAR unit 44, it is transmitted to the adjacent nodes 3 and 5 via the VP2 and VP4 set valid by the node 2.

As described above, the packet data is transmitted through the VP which is not formed in a loop in advance, so that the packet is broadcast in the network. At the same time, the source address of the data is learned (stored) in the packet switch unit 41 of each node 30. ), The communication connection between the terminals is stored.

Therefore, in the second and subsequent operations, the switching information (DA) stored in the learning table section 42 of the packet switch section 41 of each burst data I / F board 40.
The data is switched based on the data, and reaches the communication destination terminal.

As described above, according to this embodiment,
Even without a specific server such as LECS, LES, BUS, etc.
The address can be resolved by learning the address of the packet data on the burst data I / F board and the broadcast communication in the network, and the burst data can be communicated.

FIG. 3 is a diagram showing another operation of the present invention. 1 and 2 are denoted by the same reference numerals. It is assumed that the configuration of each node 30 is as shown in FIG. Here, the decentralization of the address learning will be described with reference to FIG. An OAM cell (test cell) is transmitted / received to / from a working / standby VP between burst data I / F boards 40 mounted on adjacent nodes, and the normality of the network is monitored. In FIG. 3, the thick solid line indicates the working VP, the thin solid line indicates the backup VP, and the square indicates the OAM cell.

For example, when transmitting an OAM cell, CRC-10 (check bit) is added to the payload of the cell, and the data is transmitted periodically with a gap between the data cells. The receiving side monitors the arrival of cells, and if they do not arrive at specified intervals, it is regarded as an optical transmission line failure and is used as a VP switching factor. Further, the CRC-10 added to the payload is inspected, and if the number of errors exceeds a certain threshold, it is regarded as an optical transmission line failure and is used as a VP switching factor.

For example, if the OAM cell transmitted from the node 3 cannot be detected by the burst data I / F board 40 of the node 6 or if a CRC-10 error of the payload is detected, it is regarded as an optical transmission line failure and the node 3 and the node 6 make the VP5 invalid, and the node 6 makes the VP7 valid. The node 5 is instructed via the node 2 to set the VP 7 to valid. As a result, node 3
The data transmitted from the node arrives at the node 6 through the route of VP2 → node 2 → VP4 → node 5 → VP7.

As described above, according to this embodiment,
When a failure occurs in the working connection, a communication path can be secured by switching to a spare VP whose normality has been confirmed in the OAM cell. In addition, since the address learning is decentralized, there is no influence on the entire network such as the failure of address learning due to the occurrence of a failure.

FIG. 4 is a block diagram showing a network system according to the present invention. In the embodiment shown in the figure,
There are nine nodes (terminal devices) from node 1 to node 9. Each node is connected by a VP. The number of each VP is the same as that shown in FIG. However, the VP number is the V number in the ATM switch panel 31 (see FIG. 2) in the node.
To change the P value, the number of VP1 changes to VP101. That is, VP1 is a VP value between adjacent nodes,
VP101 is a VP value in the burst data I / F board 40. Hereinafter, the same applies to the remaining VP values.

Each node is connected to Ethernet. In the example shown in the figure, terminal A is connected to the Ethernet of node 1 and terminal B is connected to the Ethernet of node 9.
Is connected.

FIG. 5 is a block diagram showing a configuration example of a node according to the present invention. The same components as those in FIG. 2 are denoted by the same reference numerals. In the figure, reference numeral 30 denotes a node (terminal device). Reference numeral 31 denotes an ATM switch panel for switching an ATM cell to a target output port, and reference numeral 34 denotes a monitoring / control panel for performing alarm, status monitoring, control, and inter-node communication in a node.

Reference numeral 40 denotes a burst data I / F for accommodating packet data of an existing Ethernet LAN in an ATM.
The board 32 is a relay I / F board which exchanges ATM cells with adjacent nodes via light. Reference numeral 33 denotes a synchronous data I / F board for storing existing synchronous data in an ATM.
Reference numeral 50 denotes an Ethernet terminal connected to the burst data I / F board 40, and reference numeral 51 denotes a synchronization terminal connected to the synchronization data I / F board 33.

FIG. 6 is a block diagram showing a specific configuration example of the burst data I / F board 40. The same components as those in FIG. 2 are denoted by the same reference numerals. In the figure, 60 is DP
A conversion unit for performing parallel-to-serial conversion of DPI 4-bit to UTOPIA 8-bit for performing mutual conversion between I and UTOPIA is connected to an ATM switch I / F. Here, UTOPIA refers to an interface specification proposed by the ATM Forum. 44 is a SAR section for performing mutual conversion between packet data and ATM cells;
Reference numeral 43 designates valid / invalid for packet data from each port of the packet switch unit 41 in transmission (Ethernet → ATM), and designates an output port to the packet switch unit 41 in packet units in reception (ATM → Ethernet). This is the data processing unit to be determined.

Reference numeral 41 denotes M in the Ethernet packet data.
A packet switch unit that learns a source address (SA) of an AC address and performs switching to a specific output port from switching information (DA). Reference numeral 49 denotes a PHY unit for performing processing of the physical layer.
And are connected through each port.

Reference numeral 45 denotes a transmission / reception cell buffer memory for buffering cells for converting between packet data and ATM cells. 46 temporarily buffers packet data from each port before passing it to the SAR unit 44. A transmission packet buffer memory 47;
This is a reception packet buffer memory for temporarily buffering packet data from the packet switch unit 41 before passing it to the packet switch unit 41.

Reference numeral 42 denotes a learning table for learning the source address (SA) of the packet data, and 48 stores the data. Reference numeral 48 denotes a learning table before switching to a specific output port.
This is a memory for storing packet data. Reference numeral 61 denotes a C for performing transmission / reception of OAM cells and setting of a data processing unit
This is a PU unit. The specific operation of the present invention will be described using the device configured as described above.

(Initial Setting Operation) In a configuration in which a network as shown in FIG. 4 is extended over a mesh network by optical transmission lines, VPs are set in advance between all nodes. VP is the AT of FIG.
Set on the M switch board 31. In the node 5 of FIG. 4, the VP is set for each adjacent node as follows.

(Example) Between node 5 and node 2 → VP4 and between node 4 → VP6 and between node 6 → VP7 and between node 8 → VP9 Next, burst data is transmitted so as not to receive the transmitted data from a plurality of directions. The VP is set to valid / invalid so that the communication path does not form a loop in the I / F board 40. In the case of the network configuration of FIG. 4, each node 30 enables only the thick line VP. That is, a tree structure is used.

The valid / invalid setting of the VP is performed by the data processing unit 43 in FIG. Each port of the packet switch unit 41 is made to correspond to an ATM VP, and packet data output from the port corresponding to the VP set to be valid is sent to the SAR unit 44.

The packet data output from the port corresponding to the invalid VP is not sent to the SAR unit 44 but is discarded. In the node 5 of FIG. 4, the valid / invalid setting is performed as follows.

(Example) Node 5 Packet switch unit port 5 → VP4 (valid) Packet switch unit port 6 → VP9 (valid) Packet switch unit port 7 → VP6 (invalid) Packet switch unit port 8 → VP7 (invalid) All A VP is set between adjacent nodes, and the normality of the network is monitored by constantly flowing OAM cells to the working (valid setting VP) and the spare (invalid setting VP). If the working connection is disconnected due to network failure,
By switching to the spare VP, a new connection can be established and a communication path can be secured.

(Operation of Address Learning Function) First, storage of a communication path by broadcast will be described with reference to a network configuration diagram of FIG. 4, a node configuration diagram of FIG. 5, and a configuration diagram of a burst data I / F board of FIG. .

In the case where the network is a mesh network as shown in FIG.
Although P is set, only the thick line VP is enabled in the burst data I / F panel 40 so that the communication path does not form a loop (tree structure).

The packet data (SA (source address) = A, DA (destination address) = B) transmitted from the Ethernet terminal A accommodated in the node 1 in FIG. 4 is broadcast-communicated in the following procedure. Here, VP value change in the ATM switch panel 31 is set to VP value ± 100.

The operation of each node will be described below.

Node 1 The packet transmitted from the terminal A in FIG.
The packet is input from the port 1 of the packet switch unit 41 through the unit 49 and is temporarily stored in the transmission packet storage memory 48. The switching information (DA) is stored in the learning table unit 4
2 and is broadcast to a port other than the input port (the SA of terminal A is stored in port 1).

The packets sent to the ports 5 to 7 (Uplinks 1 to 3) are processed by the data processing unit 43 from the ports 5 to V
P101 (VP1), port 6 → VP103 (VP3)
Is enabled, and ports 7 and 8 are disabled.

The packet is temporarily stored in the transmission packet buffer memory 46 and passed to the SAR unit 44 in packet units. S
In the AR unit 44, A
The AL5 is configured and splits it into 53-byte ATM cells. Here, AAL5 means that packet data is AT
This is a protocol for generating packet data into intermediate data before converting to M cells. DPI-UTOPIA
Converted to DPI by the conversion unit 60, the ATM switch board 3
Sent to 1. In the ATM switch panel 31, the input VP
Is switched to the corresponding output port, and transmitted to the optical transmission line via the relay I / F board 32 in FIG.

Node 2 The burst data I / F board 40 of the node 2 in FIG.
Since cells 101, 102, and 104 are enabled, the cell of VP1 is received from the node 1 via the ATM switch board 31, and the DPI-UTOPIA conversion unit 60 shown in FIG.
Converted to IA.

In the SAR section 44, the AAL5 is formed in the transmission / reception cell buffer memory 45, confirms the normality of the cell, and converts the cell into a packet. The data is temporarily stored in the reception packet buffer memory 47 from the data processing unit 43 and the output port is determined. The data of the VP 101 is sent to the port 5.

The packet input from the port 5 of the packet switch unit 41 is broadcast to a port other than the input port because the switching information is not in the learning table unit 42 (the SA of the terminal A is stored in the port 5).

Ports 6, 7, 8 (Uplink2, 3,
The packet sent to 4) is converted by the data processing unit 43 into a VP
Are set as valid / invalid, and are converted to ATM cells by the SAR unit 44, and then transmitted to the VPs 102 (VP2) and VP104 (VP4). At this time, port 6 → VP102, port 7
→ The VP 104 is set valid, and the port 8 is set invalid.

Packets sent to ports 1-4 are P
The data is passed to the connected Ethernet terminal via the HY unit 49.

Node 3 The burst data I / F board 40 of the node 3 in FIG.
Since nodes 102 and 105 are enabled, V
The cell of P2 is received and converted into a packet (VP102
→ Port 5). The packet input from the port 5 of the packet switch unit 41 is broadcast to a port other than the input port since the switching information (DA) is not in the learning table unit 42 (the SA of the terminal A is stored in the port 5).

Ports 6, 7, 8 (Uplink 2, 3,
The packet sent to 4) is converted by the data processing unit 43 into a VP
Is set to valid / invalid, and is converted to an ATM cell by the SAR unit 44, and then transmitted to the VP 105 (VP5). At this time, the port 6 → the VP 105 is set valid, and the ports 7 and 8 are set invalid.

Packets sent to ports 1-4 are P
The data is passed to the connected Ethernet terminal via the HY unit 49.

Node 4 The burst data I / F panel 40 in FIG.
08 is enabled, the cell of VP3 is received from the node 1 and converted into a packet (VP103 → port 5). The packet input from the port 5 of the packet switch unit 41 is broadcast to a port other than the input port since the switching information (DA) is not in the learning table unit 42 (the SA of the terminal A is stored in the port 5).

Ports 6, 7, 8 (Uplink2, 3,
The packet sent to 4) is converted by the data processing unit 43 into a VP
Is set to valid / invalid, and is converted to an ATM cell by the SAR unit 44 and then transmitted to the VP 108 (VP8). At this time, the port 6 → VP 108 is set valid, and the ports 7 and 8 are set invalid.

Packets sent to ports 1-4 are P
The data is passed to the connected Ethernet terminal via the HY unit 49.

Node 5 The burst data I / F board 40 of the node 5 in FIG.
Since nodes 104 and 109 are enabled, V
The cell of P4 is received and converted into a packet (VP104
→ Port 5). The packet input from the port 5 of the packet switch unit 41 is broadcasted to a port other than the input port because the switching information (DA) is not in the learning table unit 42 (the SA of the terminal A is stored in the port 5).

Ports 6, 7, 8 (Uplink 2, 3,
The packet sent to 4) is converted by the data processing unit 43 into a VP
Is set as valid / invalid, and is converted to an ATM cell by the SAR unit 44, and then transmitted to the VP 109 (VP9). At this time, the port 6 → VP 109 is set valid, and the ports 7 and 8 are set invalid.

Packets sent to ports 1-4 are P
The data is passed to the connected Ethernet terminal via the HY unit 49.

Node 6 The burst data I / F board 40 of the node 6 in FIG.
Since nodes 105 and 110 are enabled, V
The cell of P5 is received and converted into a packet (VP105
→ Port 5). The packet input from the port 5 of the packet switch unit 41 is broadcast to a port other than the input port because the switching information (DA) is not in the learning table unit 42 (the SA of the terminal A is stored in the port 5).

Ports 6, 7, 8 (Uplink 2, 3,
The packet sent to 4) is converted by the data processing unit 43 into a VP
Is set to valid / invalid, and is converted to an ATM cell by the SAR unit 44, and then transmitted to the VP 110 (VP10). At this time,
Port 6 → VP 110 is set valid, and ports 7 and 8 are set invalid.

The packets transmitted to ports 1 to 4 are P
The data is passed to the connected Ethernet terminal via the HY unit 49.

Node 7 The burst data I / F board 40 of the node 7 shown in FIG.
Since the node 108 is enabled, the cell of the VP 8 is received from the node 4 and converted into a packet (VP 108 → port 5).

The packet input from the port 5 of the packet switch unit 41 is broadcast to a port other than the input port since the switching information (DA) is not in the learning table unit 42 (the SA of the terminal A is stored in the port 5). ).

Ports 6, 7, 8 (Uplink2, 3,
All the packets transmitted to 4) are invalidated by the data processing unit 43 (data termination).

Packets sent to ports 1-4 are P
The data is passed to the connected Ethernet terminal via the HY unit 49.

Node 8 The burst data I / F panel 40 of the node 8
Is valid, the cell of VP9 is received from the node 5 and converted into a packet (VP109 → port 5).

The packet input from the port 5 of the packet switch unit 41 is broadcast to a port other than the input port since the switching information (DA) is not in the learning table unit 42 (the SA of the terminal A is stored in the port 5). ).

Ports 6, 7, 8 (Uplink 2, 3,
All the packets transmitted to 4) are invalidated by the data processing unit 43 (data termination).

Packets sent to ports 1-4 are P
The data is passed to the connected Ethernet terminal via the HY unit 49.

Node 9 Since the burst data I / F board 40 of the node 9 has enabled the VP 110, it receives the cell of the VP 10 from the node 6 and converts it into a packet (VP 110 → port 5).

The packet input from the port 5 of the packet switch unit 41 is broadcast to a port other than the input port because the switching information (DA) is not in the learning table unit 42 (the SA of the terminal A is stored in the port 5). ).

Ports 6, 7, 8 (Uplink 2, 3,
All the packets transmitted to 4) are invalidated by the data processing unit 43 (data termination).

Packets sent to ports 1-4 are P
The data is passed to the connected Ethernet terminal (terminal B in this case) via the HY unit 49.

Broadcast communication from nodes 2 to 9 can be performed similarly to node 1.

Next, data communication after storing the communication path will be described with reference to the network configuration diagram of the embodiment of FIG. 4, the node configuration diagram of FIG. 5, and the configuration diagram of the burst data I / F board of FIG.

Since SA = A of terminal A is stored in packet switch unit 41 of FIG. 6 of each node in the broadcast communication, the communication destination terminal (terminal B) performs communication based on the switching information.

The procedure by which the Ethernet terminal B (SA = B, DA = A) accommodated in the node 9 of FIG. 4 communicates with the terminal A is described below for each node.

Node 9 The packet input from the port 1 of the packet switch unit 41 in FIG. 6 contains switching information (DA = port 5).
Are stored in the learning table section 42, and are switched to the port 5 (the SA of the terminal B is stored in the port 1).

For the packet transmitted to port 5 (Uplink 1), the data processing unit 43 sets the validity of the VP, the SAR unit 44 converts the packet into cells, and then the VP 110 (VP10).
Sent to VP1 at the ATM switch panel 31 in FIG.
FIG. 5 corresponding to the output port after conversion from 10 to VP10
Is transmitted to the adjacent node 6 via the relay I / F board 32 of the above.

Node 6 The burst data I / F panel 40 of the node 6
0 and 105 are enabled, so that VP1
0 cells are received. The packet input from the port 6 of the packet switch unit 41 in FIG. 6 is switched to the port 5 because the switching information (DA = A → port 5) is stored in the learning table unit 42 (the terminal is connected to the port 6). B's SA).

The packet transmitted to the port 5 (Uplink 1) is set to the VP validity in the data processing unit 43, is converted into a cell in the SAR unit 44, and is transmitted to the VP 105 (VP5). VP10 at the ATM switch panel 31 in FIG.
5 is converted to VP5, and then transmitted to the adjacent node 3 via the relay I / F board 32 of FIG. 5 corresponding to the output port.

Node 3 The burst data I / F panel 40 of Node 3 is
Since nodes 5 and 102 are enabled, VP5
Receive the cell. The packet input from the port 6 of the packet switch unit 41 in FIG. 6 is switched to the port 5 because the switching information (DA = A → port 5) is stored in the learning table unit 42 (the terminal is connected to the port 6). B's SA).

The packet transmitted to the port 5 (Uplink 1) is set to the VP validity by the data processing unit 43, is converted into a cell by the SAR unit 44, and is transmitted to the VP 102 (VP2). VP10 at the ATM switch panel 31 in FIG.
After being converted from VP2 to VP2, it is transmitted to the adjacent node 2 via the relay I / F board 32 of FIG. 5 corresponding to the output port.

Node 2 The burst data I / F panel 40 of Node 2 is
Since cells 1, 102, and 104 are valid, the cell of VP2 is received from node 3. The packet input from the port 6 of the packet switch unit 41 in FIG. 6 is switched to the port 5 because the switching information (DA = A → port 5) is stored in the learning table unit 42 (the terminal is connected to the port 6). B's SA).

The packet transmitted to the port 5 (Uplink 1) is set to the VP validity in the data processing unit 43, is converted into a cell in the SAR unit 44, and is transmitted to the VP 101 (VP1). VP10 at the ATM switch panel 31 in FIG.
After being converted from VP1 to VP1, it is transmitted to the adjacent node 1 via the relay I / F board 32 of FIG. 5 corresponding to the output port.

Node 1 The burst data I / F panel 40 of node 1
1 and 103 are enabled, so VP1
Receive the cell. The packet input from the port 5 of the packet switch unit 41 in FIG. 6 is switched to the port 1 because the switching information (DA = A → port 1) is stored in the learning table unit 42 (the terminal is connected to the port 5). B's SA).

The packet sent to port 1 is a PHY
It is passed to the terminal A via the unit 49.

(Dispersion operation of address learning) As described in the description of the address learning function, each burst data I / F
The address of the data passing through the board 40 is learned, and the connection is stored (distribution of address learning).

A switching method when a connection is disconnected due to a network failure will be described using the network of the embodiment shown in FIG.

When the network is a mesh network as shown in FIG. 7, OAM cells are exchanged between VPs set between adjacent nodes, and the normality of the path is constantly monitored.
If an error occurs in the VP set valid in the burst data I / F panel 40, the VP is switched to the VP set invalid.

For example, the voltage between node 3 and node 6 in FIG.
When an abnormality (indicated by x in the figure) occurs in P5, node 6 → VP5 is invalidated, and node 6 → VP7
Make valid setting.

The node 3 → invalidates VP5, and instructs node 5 via node 2 to enable VP7.

Since the VP has been renewed in this manner, the packet switch unit 41 shown in FIG.
Ports forget address learning after a certain period of time. Therefore, a process of storing the communication path again in the broadcast communication is performed.

(Processing of OAM Cell) In the case of transmission, the contents of the OAM cell are set from the CPU unit 61 to the SAR unit 44 in FIG. Periodically O for the set VP
Transmit the AM cell.

In the case of reception, the data processing unit 43 of FIG.
The AM unit is identified, and the CPU unit 61 analyzes the content.

As described above, when accommodating the existing Ethernet LAN packet data in the ATM network,
In the prior art, a server resolves an ATM address from an Ethernet address by address resolution, so that a server must be installed on a network.

According to the present invention, the VP is set in advance so as not to form a loop on the ATM, and the communication path is stored in each burst data I / F board in broadcast communication, so that address resolution processing can be performed. Since the configuration is performed, a server which is indispensable in the related art becomes unnecessary.

Further, since address learning is decentralized to detect that a connection has been disconnected due to a network failure due to an OAM cell and switch to a spare VP, a communication path is ensured. Node failure does not affect the entire system. ADVANTAGE OF THE INVENTION According to this invention, a network is superior to the prior art which depends on a specific node, and greatly contributes to the technique which accommodates burst data in an ATM network.

In the above description, the case where an optical transmission line is used as the transmission line of the mesh network is taken as an example. However, the present invention is not limited to this, and the same applies to transmission lines of any other forms. Can be applied to Also, the number of mesh nodes is not limited to nine, and an arbitrary number of meshes can be used.

[0126]

As described above, according to the present invention, the following effects can be obtained.

(1) According to the first aspect of the present invention, when the transmission path is in a mesh form, the VP is set in advance in an ATM switch board mounted in an adjacent terminal device, and the transmitted data is transmitted. By setting the VP so that the communication path does not form a loop on the burst data interface board in the terminal device so that the data is not received from a plurality of directions, the server can be installed on the ATM network without installing the server on the network. The burst data communication becomes possible from the initial state in which the VP is set in advance.

(2) According to the second aspect of the present invention, when the connection between the communicating terminals has not been set yet, the ATM VP and the burst data I / O set so that the data does not form a loop. By switching the packet switch unit in the F board and broadcasting in the network, address learning on the packet data interface board and broadcast communication in the network without a specific server such as LECS, LES, BUS, etc. Thus, address resolution can be performed, and communication of burst data becomes possible.

(3) According to the third aspect of the present invention, the address learning is distributed among the burst data I / F boards, and the active / backup VP is always provided between adjacent terminal stations. Monitors the normality of the network by transmitting and receiving OAM cells, and switches to a backup VP if a VP is disconnected due to a network failure, and confirms the normality of the OAM cell if a failure occurs in the working connection By switching to the spare VP that has been set,
A communication path can be secured. In addition, since the address learning is decentralized, there is no influence on the entire network such as the failure of address learning due to the occurrence of a failure.

As described above, according to the present invention, it is possible to provide a burst data communication system capable of performing communication by converting an Ethernet address into an ATM address and performing communication without requiring a server for performing address resolution. .

[Brief description of the drawings]

FIG. 1 is a principle block diagram of the present invention.

FIG. 2 is a block diagram illustrating a detailed configuration example of a node;

FIG. 3 is a diagram showing another operation description of the present invention.

FIG. 4 is a block diagram showing a network system according to the present invention.

FIG. 5 is a block diagram illustrating a configuration example of a node according to the present invention.

FIG. 6 is a block diagram showing a specific configuration example of a burst data I / F board.

FIG. 7 is a network configuration diagram of the embodiment.

FIG. 8 is a conceptual diagram of a conventional network system.

FIG. 9 is a diagram illustrating an example of a frame format of packet data.

[Explanation of symbols]

 30 node VP virtual path

 ──────────────────────────────────────────────────の Continuing on the front page (72) Kenichi Ishikawa 3-22-8 Hakata Ekimae, Hakata-ku, Fukuoka, Fukuoka Prefecture Inside Fujitsu Kyushu Digital Technology Co., Ltd. (72) Kenji Taniguchi Nakahara-ku, Kawasaki-shi, Kanagawa 4-1, Kamikodanaka, Fujitsu Limited (72) Inventor Go Tsuyoshi Matsumoto 4-1-1, Kamikodanaka, Nakahara-ku, Kawasaki, Kanagawa Prefecture F-term within Fujitsu Limited (reference) 5K030 GA12 HA10 HB11 HC14 HD10 MD02 5K033 AA06 CC01 DA00 DB18 EB06 EC04 9A001 CC04 CC08 JJ27 KZ56

Claims (3)

[Claims] In a network system in which ATM is used as a transmission method and at least terminal devices accommodating at least burst data are connected in a mesh form, when a transmission path is in a mesh form, an ATM mounted in an adjacent terminal apparatus is provided. Set the VP in advance on the switch panel, and set the VP on the burst data interface panel in the terminal device so that the communication path does not form a loop so that the transmitted data is not received from multiple directions. A burst data communication system in an ATM network. 2. When a connection between communication terminals has not been set, data is switched between an ATM VP set so as not to form a loop and a packet switch section in a burst data interface board, and a network is switched. 2. The burst data communication system according to claim 1, wherein the data is broadcast. 3. In a configuration in which address learning is distributed in each burst data interface board, OAM cells are always transmitted and received between working terminal devices and protection VPs between adjacent terminal devices, and the normality of the network is monitored. 2. The burst data communication system according to claim 1, wherein switching to a backup VP is performed when the VP is disconnected due to a network failure.