JP2002368787A - Explicit path designation relay device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an explicit path designation relay device for supporting an explicit routing of an IPv6 packet that quickly switches a congested link to a bypass so as to relay traffic in a distributed way. SOLUTION: The relay device monitors a traffic state of its own line interfaces 1-4 and informs other surrounding relaying devices of statistic information. A processor 1-3 discriminates whether or not bypass relay is made on the basis of the traffic statistic information notified from the other relaying devices and path information stored in a memory 1-1 of the own relaying device and sets an explicit path to designate a bypass relay path. In this case, explicit path designation information retrieved by a retrieval machine 1-2 is additionally inserted to a path control header at an entrance edge of a domain, and an inserted position pointer of the inserted path and a relay designation stage number are described in a field in an extension header. Then the attached explicit path designation information is deleted at an exit edge of the domain.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an explicit routing repeater, and more particularly, to an IPv6 (Internet Protocol versa).
ion 6) In a system or the like that performs routing and relaying of a packet, an IP packet of a user to be relayed is relayed via a route (explicit route) different from a route determined by a route control mechanism by autonomous distributed control. And I
The purpose of this invention is to set a Pv6 routing control header, distribute user IP packets in a network and relay them, and expand the scope and functions of a traffic engineering system that eliminates the uneven distribution of staying traffic in the network.

[0002] In recent years, the increase in data traffic in the use of the Internet has continued to intensify, and there has been a demand for improved performance of network relay devices constituting the Internet. Above all, there is a demand for an increase in the relay capacity and the realization of quality of service (QoS) guarantee type communication for preferentially relaying a specific IP packet.

[0003] By using "traffic engineering technology" that supports technologies meeting these requirements, the throughput of user data is improved over the entire network, including best-effort packet relay, and the optimal use of relay device resources over the entire network. Technology is expected to be realized.

[0004]

2. Description of the Related Art FIG. 11 shows a functional configuration of a conventional relay device for performing explicit routing. The relay apparatus is a CR-LDP (Constraint base-service) provided in a multiprotocol label switching (MPLS) system.
Label distribution Protocol (RSP) or RSVP (Resourc
e Reservation Protocol) or other explicit routing protocol to perform load balancing relay.

In FIG. 11, 11-1 is a line interface unit, 11-2 is a shortest route calculation control unit, 11-3 is a route information database (DB), and 11-4 is a label distribution control unit (explicit route change count). A processing unit that adds a display object to a message), 11-5 denotes an explicit route determination unit, and 11-6 denotes a statistical information reception processing unit (including a function of determining congestion of the network relay device).

[0006] Also, the IPv6 protocol originally has an automated routing mechanism (for example, OSPF (Op
en Shortest Path First) or the like, a path control header function (this header field is called a Routing header) for alternately relaying IPv6 packets as user data can be added to a path other than the path calculated by the above. .

[0007] The route control header is originally added as an extension header so that communication data transmitted by the transmission source passes through a route other than the default route. However, it has not been conventionally performed that the relay device autonomously arbitrarily changes the route for transferring the user data, that is, additionally processes the routing control header (Routing header) based on the determination of the relay device.

FIG. 12 shows an operation example of load sharing relay in a multiprotocol label switching (MPLS) network. In FIG. 12, 12-1 is a label switching router (LSR) of an ingress edge device having a line interface unit, 12-2 is a label switching router (LSR) of an egress edge device, and 12-3. Is a default route determined by the shortest route calculation control unit, and 12-4 and 12-5 are engineering routes (detour routes for load distribution) calculated and set from the route information database (DB).

[0009] Conventional explicit routing mainly uses multi-protocol label switching (MPLS), and data packets (IPv4 or IPv4) from users.
6), a connection identifier of the data link layer or a label corresponding to the connection identifier is given to the link between the label switching routers (LSRs), from the ingress edge device to the egress edge device. The relay process based on the label is performed by setting the continuous transmission path of the label.

[0010]

In the conventional multi-protocol label switching (MPLS) path reconfiguration, explicit path setting control by a label switching path (LSP) tunnel is performed by each label switching router (L).
Between the SRs, path setting protocols such as CR-LDP and RSVP are exchanged with each other, and each of the relay devices autonomously executes explicit path setting control.

Here, the route change at a specific relay device is
The relay device and the explicit routing object (ER
In the case where a method of performing table management in a relay device through which a protocol message including O) passes is adopted, the relay device that has received the congestion information makes an autonomous determination, and then sets another alternate route. Therefore, it is necessary to send and receive a message of this path setting protocol.

In transmitting and receiving a message of the above-mentioned path setting protocol, since a completion response to the transmitted protocol is waited for, a delay of a certain time or more always occurs until the setting of a new bypass path is completed. Therefore, in the explicit route setting control by transmitting and receiving the protocol message, it is difficult to set a new detour route in a timely manner and change the data relay route.

The present invention does not change the routing table format or the like in the relay device at all, and does not perform path setting processing by transmitting and receiving a protocol message for path setting control between the relay devices, and therefore overhead and processing. Explicit routing relay device that realizes rapid route switching without delay (time lag), reconfigures a path bypassing a specific congested link, and can distribute and relay traffic throughout the network. The purpose is to provide.

In the explicit relay control using the routing header, if the relay device of the address described in the routing header stops operating due to a failure or the like, an error message indicating that the destination address cannot be reached is sent to the transmission source. IPv4 returned or no error message returned
Six packets are discarded, and the user data is transmitted to the network again by the retransmission control of the end host device.

[0015] In this case, the end host device must deal with the problem of the relay process in the network, and the problem of the relay process must be dealt with by the relay device in the network. The present invention actively performs a process of changing the description of the route control header, and when there is a relay device that is not operating, an explicit route specification that can continue the relay process without returning an error message etc. An object is to provide a relay device.

Also, the relay device on the route autonomously inserts route control information into the route control
If the processing for changing the relay route of the six packets is performed, the total length of the user IPv6 packet will be increased. In particular, the source end host device at the beginning
Message transfer capacity during path (MTU: Maximum Transf)
er Unit), when an IPv6 packet having a packet length according to the message transfer capacity (MTU) searched by the protocol (PMTU) for searching is transmitted as user data, the message of the output line interface is sent to the relay device along the route. When the transfer capacity (MTU) is smaller than the message transfer capacity (MTU) of the input line interface, it is necessary to prevent the relay processing accompanying the insertion processing of the path control header.

The present invention provides a method for transmitting a message when the message transfer capacity (MTU) of an output line interface is smaller than the message transfer capacity (MTU) of an input line interface.
Provided is an explicit routing repeater that does not perform a relay process that involves a process of inserting a route control header that increases the total length of a packet.

[0018]

According to the present invention, a relay device autonomously designates a route different from an original route by using a route control function for designating a passing route based on header information of an IPv6 packet. The relay device that cancels the route change process executes a process of deleting only the relay device address of the inserted traffic distribution passage route, and executes a timely traffic distribution process in the traffic engineering domain.

That is, the explicit route specifying relay device of the present invention is (1) a relay device constituting a network for relaying data via a plurality of relay devices, and is used for specifying a bypass relay route. In a relay device having an explicit route setting control means, a means for monitoring the traffic status of a line interface accommodated in the own relay device and collecting it as statistical information, and a means for notifying the other relay devices in the vicinity of the statistical information Means for determining whether or not to perform a bypass relay, based on the statistical information notified from another relay device and the network route information held in the own relay device, and Explicit route setting control means, wherein the explicit route setting control means, when the relay device is located at the entrance edge of the network domain, A relay device address list as explicit routing information to be used only within the domain is additionally inserted into the control header, and the insertion position pointer and the number of relay designation stages of the added route are described in fields in the Internet protocol extension header. Means, and if the relay device is located at the exit edge of the network domain,
Means for deleting the relay device address list information as explicit routing information added and inserted in the ingress edge relay device.

(2) A means for recognizing a suspended relay device in the network domain is provided, and in the relay device address list described as the explicit routing information, the suspended relay device is included. If there is an address, there is provided a means for selecting an address of the relay device in operation from among the relay device addresses of the path control header, and replacing the address with the destination address of the IP header.

(3) means for detecting in advance whether or not the message transfer capacity of the input packet data on the relay path is larger than the message transfer capacity of the input interface; When the transfer capacity is smaller than the message transfer capacity of the input interface, a means is provided for controlling not to insert the relay device address list into the routing header.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows functional blocks of an explicit routing repeater according to the present invention. The explicit routing relay device according to the present invention includes means for adding an address list of the relay device on the specified route to a routing header field for explicit routing, and performs processing for inserting / deleting the routing header. It has a function of transferring IPv6 packets.

The judgment and processing of insertion / deletion of the path control header are executed by the microprocessor (MPU) 1-3. The insertion / deletion of the address list of the relay device is executed according to a local rule applied only to each relay device in the domain. A relay device address list that specifies an explicit relay route is stored in a routing table (SSRA).
M) 1-1.

A user I input from a 4-port line
The Pv6 packet is stored in a data buffer 1-6 after being terminated by a MAC (Media Access Control) processing unit 1-5 via a physical layer interface unit 1-4. The microprocessor (MPU) 1-3 executes header inspection processing on the user's IPv6 packet stored in the data buffer 1-6 according to the processing program stored in the program memory 1-8, and reads necessary header information. Then, the process of inserting / deleting the route control header is determined.

The microprocessor (MPU) 1
-3 extracts a destination address from the header information, and retrieves the destination address to determine an outgoing route.
Transfer to When a response including a memory address pointer storing an additional address list to the routing control header is returned as a result of the search, the microprocessor (MPU) 1-3 converts the contents of the address list into the IPv6 address of the user. Performs processing to insert into packets.

The program memory 1-8 is used as a memory for data and commands, a memory for requests and statuses, a memory for collecting statistical information, and the like. The policy engine 1-7 also provides intelligent integrated packet classification function,
The slicer 1-9 has a function of converting an IP packet into an ATM cell.

FIG. 2 shows a processing flow for inserting the above-mentioned route control header information into a user's IPv6 packet. FIG. 3 shows a processing flow for deleting the inserted route control header information in the relay device on the outgoing side. FIG. 4 shows an IPv6 header format, and FIG. 5 shows a pointer indicating the insertion position of the routing control header information.

The process of inserting the routing control header information into the IPv6 packet will be described with reference to FIG. 2. The router of the ingress edge relay device of the IPv6 domain receives the traffic statistical information while receiving the user's input from the line. The routing table is searched by the destination address (DA) of the IPv6 packet (2-1).

Then, LRTE (Loos
e Route Traffic Engineering: Traffic engineering for thinning out and setting the relay device address of the designated route) It is determined whether or not data indicating execution is set (2-2). As this data, an output interface number, a relay device address to be passed, and the like are set.

If the data is set, it is determined whether or not the received packet has a routing control header (2).
-3) If there is no route control header, load a route control header to be newly inserted into the memory (2-4). Then, the address of the relay device to be passed is loaded from the search result and inserted into the routing header (2-5).

Next, the pointer indicating the insertion position of the inserted relay device address is edited (2-6), and the path control header and the insertion position pointer are inserted into the received packet (2-6).
7) Copy the next header field value of the IPv6 header to the next header field of the routing header,
The value indicating the route control header is written in the next header field of the Pv6 header (2-8). Then, the received IPv6 packet is transferred (forwarded) based on the search result of the routing table (2-9).

Next, the process of deleting the routing control header information will be described with reference to FIG. 3. The router of the egress edge relay device in the IPv6 / MPLS domain transmits the data from the domain in which traffic engineering (TE) is being executed using the routing header. Receiving the IPv6 packet of (3-1),
It is determined whether or not the relay device address insertion position pointer to the route control header is 0 (3-2). If the pointer is not 0, whether the address list is deleted or the route control header is deleted is determined from the relay device address insertion position pointer. It is determined (3-3).

If the address list is to be deleted, the address list inserted into the path control header is deleted according to the insertion position pointer information into the path control header (3-4).
In the case of deleting the route control header, the route control header itself inserted in the route control header is deleted according to the insertion position pointer information in the route control header (3-5), and the next header in the header preceding the route control header is deleted. The next header field value assigned to the deleted route control header is written in the header field (3-6). Then, the payload length is subtracted by the amount corresponding to the deletion of the routing header information, and the
The data is written in the v6 header (3-7), and data transfer (forwarding) to the next path is started (3-8).

As shown in FIG. 4, the header format of IPv6 is a basic header including version, traffic class (priority), flow label, payload length, next header, hop limit, source address and destination address, and hop-by. It consists of a hop option header, a destination option header, a routing header, a fragment header, an authentication header, and an extension header including an upper layer header.

The relay station address of the passing route is written in the routing header of the extension header. FIG. 5 shows the format of the routing header. The figure shows type 0
Of the routing header, the routing header includes a next header, an extension header length, a routing type, a remaining segment, a reserved (Reserved) field, and an area of each relay device address of a passage route.

The next header field is an 8-bit selector, which indicates the type of the header immediately following the routing header, and uses the same value as the protocol field in the IPv4 packet. The extension header length field indicates the length of an 8-octet unit of the routing header by an 8-bit unsigned integer (however, the leading 8 octets).
Octets are not included). Note that the value of the extension header length of the type 0 routing header is equal to twice the number of relay device addresses in the routing header.

The routing type field (in this case, "0") is written in 8 bits in the routing type field. In the remaining segment field, an 8-bit unsigned integer is used to indicate the number of remaining segments on the path, that is, the number of intermediate nodes (relay devices) specified to pass through to reach the final destination. Be included.

The reserved (Reserved) field is a 32-bit information field, which is initialized to "0" at the time of transmission and is ignored on the receiving side. Each of the relay device addresses [1] to [n] of the passage route is a 128-bit address of each relay device numbered from 1 to n.

In the first embodiment of the present invention, the following control field format is inserted as an example in a reserved field as a pointer indicating the insertion position of the relay device address in the path control header. (1) A shim header format field in the multiprotocol label switching (MPLS) system is defined and inserted into the reserved field. (2) The number of stages of the shim header is one. (3) If the data in the Reserved (Reserved) field is not "0", it is determined that the shim header is inserted as significant data. (4) In this case, the label field (20 bits) of the shim header is divided into 2 octets and 4 bit subfields, and the range specification of the insertion path list at the time of traffic engineering by the path control header is described.

FIG. 6 shows a shim header format. The shim header format is RFC3
032 Jan. 2001.
The shim header format is a 20-bit label field and a 3-bit experimental use (Exp)
The field consists of 32 bits, a 1-bit bottom stack (S) field, and an 8-bit time-to-live (TTL) field.

The first 8 in the Label field
In the bit subfield (sub-label-1), a pointer indicating a position at which the contents of the routing header for traffic engineering is inserted is described. In the second 8-bit subfield (sub-label-2), a pointer indicating the number of contents of the routing header for traffic engineering is described. The 4-bit sub-field (sub-label-3) in the label (Label) field is reserved.

As described above, the path change control mechanism is incorporated in the path control header of the IPv6 packet without performing the path setting processing by the path setting control protocol, thereby bypassing the user's IPv6 packet to another path. In this case, in order to notify the explicit routing information to another group of relay devices, the first embodiment of the present invention uses the TLV (Type / Length / Value) included in the extension header field.
Reserve of the routing header, which is a group field
d) In the field, describe the insertion position pointer of the newly added route (relay device address list) and the number of designated relay stages.

In this case, if there is a relay device that does not support the route change control mechanism and executes a process of clearing the Reserved field of the route control header to “0”, a new device is added. The processing of deleting the route cannot be performed by the relay device at the exit edge, and a malfunction occurs in the operation.

That is, an indication that the relay device address list has been additionally inserted is displayed in the reserve (Rese
rved) to control the route change. However, originally, the reserved (Reserved) field is a field reserved for future use, and in other implementations of the IPv6 packet relay mechanism, This reserved (Reserved) field is cleared to zero and relayed,
If the content is checked and it is not zero, it is expected to be used for various judgments and processings such as alarm processing as abnormality detection.

Therefore, a relay device on the route is always defined as a header to be inspected, and a new field definition is added to a hop-by-hop option field which is a general-purpose header field that can be used for other purposes. List inserts can be displayed. What is specified as a hop-by-hop option field is a part of the field value, and the others are unspecified.

As the value of the specified portion, the third bit from the top of the option type is a bit for designating whether or not this option data can be modified halfway to the final destination of the packet. By setting the option data to a value (0) that cannot be changed on the way, it is possible to solve the problem caused by the zero clear described above. However, the load of hop-by-hop option header insertion / deletion processing increases.

However, in order to avoid a problem due to the use of the reserved field, instead of the reserved field of the routing header, a second embodiment of the present invention employs a hop-by-hop protocol defined in the IPv6 protocol standard. The pad N option function of the option header is extended to display the insertion position display pointer and the number of relay stages of the newly added path.

The field definition of the pad N option (alignment requirement: none) in the IPv6 protocol standard specification is described below. The pad N option is used to insert padding of 2 octets or more into the option area of the header. FIG. 7A shows a format of a pad N option (alignment requirement: none) field. In the case of N octets of padding, (N-
The value of 2) is described, and the "0" value of (N-2) octets is described in the option data field.

According to the specification of the IPv6 protocol standard, a series of options in a header must be strictly processed in the order in which they appear in the header. For example, searching for a particular option in the header and all preceding Before an option is processed, it must not be processed. Note that the Pad 1 option is used to insert one octet of padding in the option area of the header, and if padding more than one octet is required, it is shown here rather than using multiple Pad 1 options. Pad N option is used.

The option type identifier is more than just identifying a particular type of option, and its upper two bits determine how the IPv6 node being processed cannot recognize the option type. Encoded to specify whether to handle. The meaning of the sign of the upper two bits of the option type is defined as shown in FIG. 7B, and the pad 1 option and the pad N option follow the definition of the code “00” in the table of FIG. 7B.
Repeater is IP with hop-by-hop option header
When the v6 packet is received, the option type is checked. If the option type is the pad 1 option or the pad N option, processing corresponding to the option is performed, and then the next extended header processing is performed.

Here, regarding the insertion position pointer of the additional route (relay device address list) and the description of the number of designated relay stages in the present invention, a shim (MPLS) system shim (MPLS) system as an example in the option data portion of the pad N option is described. shim) Insert the header. The number of stages of the shim header can be determined by referring to the option data length field and calculating a multiple of 4 octets.

The option type value is “00000XX”
If "X" is used and the option data is not "0",
It is determined that a shim header has been inserted as significant data. In this case, the label field (20 bits) of the shim header is divided into 2 octets and 4 bits of subfields, and the range specification of the insertion path list at the time of traffic engineering by the path control header is described.

I with hop-by-hop option header
When a Pv6 packet is received, an insertion position pointer of the path control header information is described in a form added thereto. When a user IPv6 packet without a hop-by-hop option header is received, a new hop-by-hop option header is set.

In this case, it is necessary to rewrite the next header field before the path control header to be inserted with an indication that the header is a hop-by-hop option header. In the pad N option, an insertion position information pointer of the path control header as shown in FIG. 6 is described. As a format of the position information pointer, for example, a shim header format shown in FIG. 6 can be used.

Next, FIG. 8 shows a processing flow example when the operation of the next hop relay apparatus (router) is stopped in the domain of the IPv6 relay system. Each relay device normally operates a dynamic routing protocol typified by OSPF (Open Shortest Path First), so that it is possible to detect whether or not an adjacent relay device is operating.

When an IPv6 packet is received from the traffic engineering domain using the routing header (8-1), it is determined whether the relay device of the address described in the IPv6 packet header is operating normally (8-1). -2) When the operation is stopped due to a failure or the like, even if the route information to the relay device is described in the route control header (routing table),
Next relay address described in the routing header and I
A process of exchanging the destination address of the Pv6 basic header is executed (8-3), an IPv6 packet is transferred to the next route (8-4), and the relay of the user's IPv6 packet is continued.

As described above, in the middle of the relay process that the end host device cannot know, the relay device autonomously sets the value of the route control header and actively uses the user IPv6 to alleviate the congestion in the network. A process of inserting or deleting a route control header to or from the original route control header added to the packet is performed. In this case, if a relay device in the network domain that has caused a failure and stops the relay operation due to a failure, but the relay device at the immediately preceding hop recognizes the stopped relay device (can be determined by sending and receiving a Keep alive packet), The relay device performs an address list replacement process of the route control header, and relays the user IPv6 packet to the next destination address (operating route).

Such processing is not normally performed on the route control header set by the user. An error message indicating that the destination cannot be reached is returned to the transmission source device, and is retransmitted by the processing of the transmission source device. However, according to the present invention, since the routing control header is created and inserted by the relay device in the middle of the route, the process of replacing the contents of the routing control header with the operating destination address without leaving retransmission control to the transmission source host device. Is executed autonomously, without applying a processing load to the source host device, the user IPv6
Packet relaying can be continued.

Next, a process for controlling the insertion of the routing control information based on the message transfer capacity of the input interface and the output interface will be described with reference to FIG. The explicit routing repeater according to the present invention must be aware of the difference in message transfer capacity (MTU) size of the relay route.

For this reason, “Interface” usually described in the routing table held by the relay device is used.
By inspecting the contents of the field, the control of the additional insertion processing of the path control information is performed based on the message transfer capacity (MTU) size of the relay path. Specifically, “Int
For example, Ethernet (registered trademark), FDDI (Fiber Distributed Data Int)
interface type (link type) such as ATM (Asynchronous Transfer Mode) and ATM (Asynchronous Transfer Mode), the size relationship of the message transfer capacity (MTU) size can be determined by referring to this.

In the IPv6 / multiprotocol label switching (MPLS) system, the label switching router (LSR) at the domain ingress edge has a message transfer capacity (MTU) size of the ingress interface and a message transfer capacity (MT) of the egress interface.
The size of U) is compared with reference to the interface type (link type) information in the routing table to determine the magnitude relationship.

As shown in FIG. 9, first, the user IPv6
The message transfer capacity (MTU) of the receiving interface of the packet is stored (9-1). Next, the routing table is searched to determine the type of the outgoing interface for performing the load distribution processing in the IPv6 routing control header (9
-2). Then, it is determined whether the size of the message transfer capacity (MTU) of the outgoing interface is larger than the message transfer capacity (MTU) of the receiving interface (9-3).

When the size of the message transfer capacity (MTU) of the outgoing interface is smaller than or equal to that of the receiving interface, a flag for prohibiting dispersion in the routing header is set in the interface (9-4). When the prohibition flag is set, the relay device stops the process of setting the additional address list to the routing header in the search result of the routing table. As a result, no routing control by the routing header is performed on the interface.

FIG. 10 shows an example of information held in the routing table. Here, the field described as “Interface” indicates the interface type (link type), and it is determined from the description of “ei31” in FIG. 10 that the interface type (link type) is Ethernet (registered trademark). .

Therefore, the interface type that has received the IPv6 packet of the user and the interface type (Interfac) obtained by searching the routing table are described.
If the message transfer capacity (MTU) of the egress interface is not larger than the content of the e-field, the process of adding an address to the routing header is not performed.

(Supplementary Note 1) In a relay device constituting a network for relaying data via a plurality of relay devices, the relay device having an explicit route setting control means for designating a route of a detour relay , Monitor the traffic status of the line interface accommodated by the own relay device,
Means for collecting statistical information, means for notifying the other relay devices of the surroundings of the statistical information, and means for collecting the statistical information notified from other relay devices and the network route information held in the own relay device. Means for determining whether or not to perform a bypass relay, and explicit route setting control means for specifying a route for the bypass relay, wherein the explicit route setting control means , A relay device address list as explicit routing information used only within the domain is additionally inserted into the routing control header for explicit routing, and the insertion position pointer of the added route and the number of relay designation stages If the relay device is located at the egress edge of the network domain, the Edge relay means for deleting a relay device address list information as additional inserted explicit routing information has been in the apparatus, characterized by including an explicit routing repeater. (1) (Supplementary note 2) The apparatus further comprises means for recognizing the suspended relay device in the network domain, and the suspended relay device is included in the relay device address list described as the explicit routing information. 2. The explicit routing repeater according to claim 1, further comprising: a unit that replaces the relay device address of the routing control header with a route to an active relay device, if any. (2) (Supplementary note 3) A means for detecting in advance whether the message transfer capacity of the input packet data on the relay path is larger than the message transfer capacity of the input interface, and the message transfer of the output interface is provided. 3. An explicit statement according to claim 1 or 2, further comprising means for controlling not to add a relay device address list to the routing header when the capacity is smaller than the message transfer capacity of the input interface. Routing repeater. (3) (Appendix 4) The explicit route setting control means includes means for describing an insertion position pointer of the added route and the number of designated relay steps in a spare field in a route control header in an extension header of the Internet protocol. The explicit routing repeater according to claim 1, wherein: (Supplementary Note 5) The explicit route setting control means includes means for describing an insertion position pointer of the added route and the number of designated relay steps in a hop-by-hop option header field in an extension header of the Internet protocol. 9. The explicit routing repeater of claim 1 wherein:

[0067]

As described above, according to the present invention,
When performing load distribution control to avoid traffic concentration, insert / delete the relay device address list of the route to be changed into / from the route control header, thereby completely changing the format of the routing table in the relay device. LDP / C for path setting control between relay devices
Without causing overhead or processing delay by using a protocol such as R-LDP or RSVP,
Further, it is possible to perform a route change control for distributing and relaying a traffic load in a network without using a signaling function or the like.

Even if a relay device address is temporarily added to the routing header in the domain where the traffic engineering according to the present invention is being executed, by deleting the additional relay device address at the exit edge of the domain,
With other relay devices, the traffic load can be distributed and relayed without any problem.

Further, the device that relays the IPv6 packet can check the header as usual, and relay the IPv6 user packet according to the route described in the routing control header. As a result, it is possible to quickly reduce the input load to the link where congestion has occurred in the network, and it is possible to increase the apparent communication throughput of the end user using the Internet while using this route. In addition, quick route switching without a time lag can be realized without adding a contradiction to a standard protocol specification by adding it to an IPv6 protocol process described in an Internet draft or RFC.

When the suspended relay device is present on the route specified by the route control header, the route to the suspended relay device is replaced with the route to the activated relay device, and the destination is changed. Communication can be continued without returning an unreachable error message to the host device or interrupting the relay process due to discarding the IPv6 packet without returning any error message.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of an explicit routing repeater according to the present invention.

FIG. 2 is a diagram showing a flow of processing for inserting routing control header information.

FIG. 3 is a diagram showing a flow of processing for deleting inserted path control header information.

FIG. 4 is a diagram showing an IPv6 header format.

FIG. 5 is a diagram showing a format of a routing header.

FIG. 6 is a diagram illustrating a shim header format.

FIG. 7 is a diagram showing a format of a pad N option field and code contents of an option type.

FIG. 8 is a diagram showing a processing flow when the operation of the relay device of the next hop is stopped.

FIG. 9 is a flowchart of a process for performing insertion control of routing control information based on a message transfer capacity of an input interface and an output interface.

FIG. 10 is a diagram illustrating an example of information held in a routing table.

FIG. 11 is a diagram illustrating a functional configuration of a conventional relay device that performs explicit routing.

FIG. 12: Multiprotocol label switching (MP
FIG. 9 is a diagram illustrating an operation example of load sharing relay in an LS) network.

[Explanation of symbols]

 1-1 Routing table (SSRAM) 1-2 Search machine 1-3 Microprocessor (MPU) 1-4 Physical layer interface unit 1-5 MAC (Media Access Control) processing unit 1-6 Data buffer 1-7 Policy engine 1 -8 Program memory 1-9 Slicer

Claims (3)

[Claims] 1. A relay device that constitutes a network that relays data via a plurality of relay devices, the relay device including an explicit route setting control unit for designating a route of a detour relay. Means for monitoring the traffic status of the line interface accommodated by the relay device, collecting the statistical information, means for notifying the statistical information to other relay devices in the vicinity, and the statistical information notified from the other relay device; A means for determining whether or not to perform a detour relay based on network route information held in the own relay device; and an explicit route setting control means for designating a detour relay route. When the relay device is located at the entrance edge of the network domain, the setting control means includes, in the routing control header for explicit routing, an explicit Means for additionally inserting a relay device address list as routing information, and describing an insertion position pointer of the added route and the number of relay designation stages in a field in an extension header of the Internet protocol; Means for deleting the relay device address list information as the explicit routing information added and inserted by the ingress edge relay device. 2. A relay apparatus, comprising: means for recognizing a suspended relay apparatus in the network domain, wherein the relay apparatus address list described as the explicit routing information includes:
Means for selecting an address of the relay device in operation from among the relay device addresses in the route control header and replacing the address of the relay device with the destination address in the IP header when the address of the relay device in operation suspension exists. 2. The explicit routing repeater of claim 1, wherein: Means for detecting in advance whether or not the message transfer capacity of the input packet data on the relay path is greater than the message transfer capacity of the input interface; 3. The explicit route according to claim 1, further comprising means for controlling not to insert a relay device address list into the route control header when the message transfer capacity of the input interface is smaller than the message transfer capacity. Designated repeater.