JP2591186B2 - Distributed failure recovery system and device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to recovery from a failure in a network communication composed of a plurality of nodes at the time of a link failure, and more particularly to a technique for recovering from a failure due to distributed automatic detour routing.

(Prior Art) Conventionally, this type of failure recovery technique can be roughly classified into two types: a method using a centralized control method and a method using a distributed control method.

In the centralized control method, when the control unit alarm information is collected, a detour search algorithm is activated to determine a detour,
The detour is formed by transferring the detour information, and the optimal rerouting (re
routing) is possible. However, if a multi-transmission path failure is also taken into consideration, the setting of the collection time of the alarm collection and the algorithm for forming the detour route become complicated. Further, as the network becomes larger, the time required for alarm collection and the time required for transferring the detour information increase, and a great deal of time is required for fault recovery.

In the distributed control method, each node in the network exchanges information by transmitting and receiving packets to search for a detour in an autonomous and distributed manner so as to recover from a failure. Until now (a) WDGrover, "THE SELFHEALING ^TM NETWORK" , proce
eding of Globecom'87, Nov. 1987. (b) HCYang and S. Hasagawa, “FITNESS: FAILURE IM
MUNIZAITON TECHNOLOGY FOR NETWORK SERVICE SURVIVAB
ILITY ", Proceeding of Globecom" 88, Dec. 1988. (c) HRAmirazizi, "CONTROLLING SYNCHRONOUS NETW
ORKS WITH DIGITAL CROSSCONNECT SYSTEMS "Proceeding
of Globecom'88, Dec. 1988., the method (a) shows a method of recovering from a single transmission failure. The method (c) is a recovery method for a single transmission line failure, and has a problem that a control packet forms a loop and delays failure recovery. Further, each method is a "failure-end circuit" in which a channel which has become normal due to a transmission path failure is bypassed between both ends of the failure line, and is not a recovery in consideration of the priority of the failure channel. However, when recovering from a failure in an actual network (especially a corporate network), "line end recovery" is more desirable in view of effective use of a transmission line and effective use of a transmission line.

Accordingly, the present invention provides an adaptation to the recovery of a multi-line failure, a packet loop avoidance, a high-speed failure recovery, a failure recovery capable of priority control, a line end recovery, and an optimal rerouting in order to improve the disadvantages of the distributed control system. It is an object of the present invention to provide a fault recovery technology that allows for troubleshooting.

(Configuration of the Invention) According to the present invention, in a network in which a plurality of nodes are connected in a mesh by transmission paths and all paths (lines) are prioritized, a plurality of failures occur when a transmission path failure occurs. Since there may be a path, if the nodes at both ends of the failed path are the first and second nodes, there are a plurality of first and second node sets, and the first and second nodes are provided for each failed path. Transmits an open packet to a connection node of a failed path according to packet transmission control, the open packet has control information 0 including at least the identifier of the failed path, and is a packet transmitted on the same line as the packet transmission control. This is to store the packet in the queue table so that the packet is transmitted in the order of higher priority. When transmitting a packet, the packet transmission control is all performed according to the packet transmission control. When the open node receives the open packet, it releases the link of the failed path and transfers the open packet to the next connected node. The packet is sent to all the adjacent nodes having the spare band, and the search for the adjacent node having the spare band equal to or larger than the band of the failed path is repeated a predetermined number of times with a certain waiting time, and the control packet is transmitted at least in the failed path. It has control information 1 including an identifier, band information of a failed path, a node trace indicating a node through which the control packet has passed, and a hop count indicating the number of nodes through which the control packet has passed. And a node excluding the second node is defined as a third node. When the hop count is received, the control packet 1 is recorded, and if the hop count is equal to or less than a predetermined value, the control information 1 is updated. Transfers the control packet to an adjacent node having a spare band equal to or larger than the band of the failed path, and, when the second node receives the control packet, returns to the node connected to the transmission path from which the control packet was sent. Send back the packet,
The return packet has at least an identifier of the failed path, band information of the failed path, and control information 2 including a node trace indicating a node through which the return packet has passed. From the control packet reception record stored in the packet and the stored control packet reception record, an adjacent node reaching the first node by the shortest path is selected from the connection transmission lines having the spare band, and node trace information is transmitted to the node. Transmit the updated return packet, transmit a positive acknowledgment packet to the transmission source node of the received return packet, and allocate the spare band of the transmission path to the detour path for the bandwidth of the failed path, and if there is no connection transmission path having the spare band, Negative acknowledgment packet for the node that sent the return packet When the first node receives the return packet, the node receiving the negative ack packet releases the band allocated to the alternate path of the failed path and searches for another alternate path again. Is assigned, and an end packet is sent back to the second node through the failure detour 1. When each of the nodes on the failure detour 1 receives the end packet, it controls a switch in the node to cause a failure. A decentralized fault recovery method and apparatus, wherein a detour is formed, and when the end packet is received by the second node, a fault detour is formed by diverting a fault path to the fault detour (1). Is obtained.

(Example) Next, the present invention will be described with reference to the drawings. FIG. 1A is a block diagram showing a process of multiplexing data and transmitting and receiving control data as one embodiment of the present invention. The data input terminal 101 is connected to an interface circuit (INF) 102, and after level conversion, the data is input to a multiplex circuit (MU).
X) 103, control data is sent to the packet control circuit (PA
CKET-CONT) 108, the path alarm data is stored in the CPU processing unit 11
Connected to 8. Here, the path is the path 217 shown in FIG.
A logical connection line extending between two points via a plurality of (single) lines as shown in FIG.

The switch circuit (TSI) 104 exchanges the time slots of the data multiplexed by the MUX 103 based on the switch information held in the built-in RAM, and is connected to the interface circuit (INF) 106 via the separation circuit (DEMUX) 105. You.

PACKET-CONT108 converts the control data extracted by INF102 into C
The control data is transferred to the PU processing unit 118, and the control data generated by the CPU processing unit 118 is distributed to the INF.

 INF 106 is connected to data output terminal 107.

The CPU processing unit 118 stores the control data obtained through the PACKET-CONT 108 in a packet storage queue (PACKET-QUEUE) 116, and according to the contents of the packet, a control packet reception processing unit (detailed in FIG. 1D). CONT-PAK-P
RO) 109, a return packet reception processing unit (RET-PAK-PRO) 110 detailed in FIG. 1 (e), a positive ack reception processing unit (POS-ACK-PRO) detailed in FIG. 1 (f). PRO) 1
11. Negative ACK reception processing unit (NEG-ACK-PRO) 112 detailed in FIG. 1 (g), end packet reception processing unit (END-PAK-PRO) detailed in FIG. 1 (h) 113, first
Open packet reception processing unit (OP
EN-PAK-PRO) 115 for processing. At this time,
Each of the processing units 109 to 115 sets a packet while setting and searching for a flag associated with the packet stored in the PACKET-QUEUE 116 or changing and searching for each band in the assigned table (ASSIGN-TABLE) 117, and sets an alarm. Detection processing unit (ALM-D
ET-PRO) 114, PACKET-CPNT
Transfer to 108. Especially END-PAK-PRO113 and OPEN-PAK-PR
O115 changes the RAM data for storing the switch information in the TSI 104.

Next, the algorithm of each of the processing units 109 to 115 of the CPU processing unit 118 will be described. The present invention comprises four phases from failure detection to failure recovery. One is a release phase in which open packets transmitted from both nodes of the failed path release the link of the failed path toward the opposing node via the connection node. The processing in each node is ALM-DET-PRO1
14 or OPEN-PAK-PRO115. The second is the broadcast phase, where one end node of the failed path (SENDE
R) broadcasts a control packet as an adjacent node failure message having an available bandwidth equal to or greater than the failed path bandwidth, and searches for a possible detour. ALM-DET-PRO114 or CONT-
There is PAC-PRO-109. Third, in the return phase, when the other node (CHOOSER) on the faulty path receives the control packet, it generates a return packet, and the return packet is sent back to SENDER to search for a detour. Also, there is a positive ACK packet or a negative ACK packet as a response of the return packet between each relay node. RET as processing in each node
-PAK-PRO110, POS-ACK-PRO111 or NEG-ACK-PR
There is O112. The fourth is the confirmation phase.
An end packet of the detour selected by the return phase is passed to connect the detour. The processing in each node includes RET-PAK-PRO110 and END-PAK-PRO113. Further, a fault path ID is added to the above six packets.

When the alarm warning is transferred to the CPU processing unit 118 in FIG. 1A, the processing in the ALM-DET-PRO 114 is executed.
FIG. 1B is an explanatory diagram showing the operation of the ALM-DET-PRO 114. In step S00, an open packet to which a failure path ID is added is set for a connection node of the failure path, and the packet is transferred to the PACKET-CONT 108 shown in FIG. 1A. SENDER (for example, it is uniquely determined as a node having a smaller node number among both nodes of the failed path), and if it is SEND
If it is ER, the process proceeds to step S03, and it is determined whether or not there is a free bandwidth equal to or greater than the faulty path bandwidth in the line. The operation of performing the transfer to the PACKET-CONT 108 shown in FIG.
For all the lines connected to the own node. In addition, as the band state of the line, an unused state (SPARE) 121, a reservation state (RESERVED) 122,
There are four states, a use state (USED) 123 and a disconnection state (FAIL) 124, and a band state and a path allocation state are set in the ASSIGN-TABLE 117. In the ASSIGN-TABLE 117 in both ends nodes A and B of a certain path and in the relay node, the state of the path as viewed from the own node is set, and the band state of each path usually matches. However, the ASSIGN-TABLE 117 is changeable, the nodes on both ends of the failed line that have detected the line alarm perform state changes 125, 131, and 133, and the nodes A and B that have detected the path alarm perform state change 129 of the path,
At the time of recovery from a failure, 126 and 132 are performed, and when a return packet is sent to the path, the state of the band assigned to the path is changed.
7 is performed, the state change of the path band is performed at the time of receiving a positive ack, the state change of the path band is performed at the time of a negative ack, the path band is changed at the time of a negative ack, and In order to perform the bandwidth state change 130, the path of the ASSIGN-TABLE 117 of each of the nodes A and B from when the node A (or B) sends out a packet to when the node B or A processes the packet. Band state settings may be different.

When the own node receives an open packet as a distributed control packet, the open packet is transferred to the CPU processing unit 118 in FIG. 1A, and the processing in the OPEN-PAK-PRO-115 is executed. FIG. 1C is an explanatory diagram showing the operation of the OPEN-PAK-PRO 115. In step S10, T in FIG.
The switch information in the SI 104 and the ASSIGN-TABLE 117 are rewritten to release the link of the failed path, and in step S11 an open packet is set for the next connection node and transferred to the PACKET-CONT 108 in FIG. Is performed.

When the own node receives a control packet as a distributed control packet, the control packet is transferred to the CPU processing unit 118 in FIG. 1A, and the processing in the CONT-PAK-PRO 109 is executed. FIG. 1D shows CONT-PAC-PRO109.
It is an explanatory view showing the operation of. In step S20, it is determined whether or not the own node is a CHOOSER (for example, uniquely determined as a node having a large node number among the nodes at both ends of the failed line). The received control packet is S
If the number of hops indicating the number of nodes transmitted from the ENDER to the reception of the own node is smaller than a predetermined number S, a connection line having an available bandwidth equal to or greater than the failure path bandwidth in accordance with the broadcast rule in step S26. Then, an operation of setting a control packet toward the PACKET-CONT 108 shown in FIG. 1A is performed. The broadcast rule is that the transmission destination is limited to a node other than the trace node (passing node) of the control packet that has not been transmitted for the path failure and that the received control packet is transmitted. This is to record the own node in the node trace data. Here, the node trace data is composed of a bit sequence of the number of nodes constituting the network, is data representing a node which has passed a packet by setting 1 to a bit position of a passed node number, and is added to a control packet and a return packet. You. If the number of hops is X, control packets are not relayed. If the own node is XHOOSER, the process proceeds to step S21, and it is searched whether a return packet for the failed path has been transmitted.
Proceeding to S22, if there is an available bandwidth equal to or greater than the failure path bandwidth on the line receiving the control packet,
The state of the band assigned to the failed path of SSIGN-TABLE 117 is represented by R
ESERVED, a return packet is set for this line in step S24, and the PACK shown in FIG.
The operation of transferring to ET-CONT108 is performed. At this time, node trace data is added to the return packet. If there is no free bandwidth equal to or greater than the failed path bandwidth in the receiving line in step S22, and if there is no RESERVED bandwidth equal to or greater than the failed path bandwidth in step S27, the processing is terminated.

In FIG. 1A, the processing in the RET-PAK-PRO 110 in which the return packet is transferred to the CPU processing unit 118 is executed. FIG. 1E is an explanatory diagram showing the operation of the RET-PAK-PRO 110. It is. In step S30, it is determined whether or not the intersection of the return packets to which the same band is allocated has occurred.If the intersection of the return packets has occurred, the return packet transmitted in step S38 is searched, and in step S39, The priority of the transmitted return packet and the priority of the received return packet are compared, and if the priority of the received return packet is low, a negative ACK packet is set for the received line in step S40, and FIG. ) PA
The operation of transferring to CKET-CONT108 is performed. Step S
If there is no return packet intersection at 30,
Proceeding to step S31, it is determined whether or not the own node is a SENDER, and if it is not a SENDER (relay node), a step S32 is performed.
Then, among the control packets stored in the PACKET-QUEUE 116 of FIG. 1 (a), the control packet which is not included in the node trace data of the return packet received by the transmission source node and has the smallest hop number is searched. 1 as determined in step S33
If there is no control packet, the process proceeds to step S40. If there is a control packet, the state of the band allocated to the control packet reception line in step S34 is RE.
SERVED, and in step S35, a return packet in which the own node is added to the node trace data is set for the control packet transmission source node, and the packet is transferred to the PACKET-CONT 108 shown in FIG. In S36, the status of the band to which the detour path on the receiving line of the return packet is allocated is set to USED, and step S37
Then, a positive acknowledgment packet is set for the return packet transmission source node, and the PACK shown in FIG.
The operation of transferring to ET-CONT108 is performed. Step S31
In step S41, if the own node is a SENDER, the switch information RAM in the TSI 104 in FIG. 1A is rewritten in order to connect the faulty path and the detour (the return packet passage). In steps S42 and S43, the same processing as in steps S36 and S37 is performed. The passage until the return packet sent from the CHOOSER reaches the SENDER is a detour for failure recovery. Next, in step S44, an end packet is set for the source node of the return packet, and the packet is transferred to the PACKET-CONT 108 shown in FIG. 1A.

In FIG. 1A, when a positive ACK packet is transferred to the CPU processing unit 118, the processing in the POS-ACK-PRO 111 is executed. FIG. 1F is an explanatory diagram showing the operation of the POS-ACK-PRO 111. In step S50, the state of the band to which the detour path on the line receiving the positive acknowledgment packet is allocated is set to USED.

In FIG. 1A, when a negative ACK packet is transferred to the CPU processing unit 118, the processing in the NEG-ACK-PRO 112 is executed. FIG. 1 (g) is an explanatory diagram showing the operation of NEG-ACK-PRO-112. The state of the band allocated on the line received in step S60 is set to SPARE, and step S61
Then, among the control packets stored in the PACKET-QUEUE 116 of FIG. If there is no control packet in step S62, the process proceeds to step S65,
If there is a control packet, the state of the band allocated on the control packet reception line in step S63 is RESE.
RVED, and in step S64, set a return packet toward the control packet transmission source node.
The packet is transferred to the PACKET-CONT 108 shown in FIG. If the current node is a CHOOSER in step S65, the process ends and the CH
If it is not OOSER, in step S66, the own node sets a negative ACK packet toward the transmission source node of the reception return packet from which the return packet is transmitted, and the PACKET-CONT-108 shown in FIG. Is performed.

When the end packet is transferred to the CPU processing section 118 in FIG. 1A, the processing in the END-PAK-PRO 113 is executed.
FIG. 1H is an explanatory diagram showing the operation of the END-PAK-PRO 113. In step S70, it is determined whether the own node is a CHOOSER.
If it is SHOOSER, in step S71, the switch information RAM in the TSI 104 in FIG. 1A is rewritten in order to switch the route of the faulty path to the detour (passage path of the return packet), and the process ends. If the own node is not a CHOOSER, the flow advances to step S72 to search for the received return packet, and in S73, the switch information RAM in the TSI 104 of FIG.
In step S74, an end packet is set for the next detour connection node, and the packet is transferred to the PACKET-CONT 108 shown in FIG. 1 (a).

Next, the operation of the distributed failure recovery system of the present invention will be described.
FIG. 2 shows a network model composed of 10 lines of node (1) 201, node (2) 202, node (3) 203, node (4) 204, node (5) 205, node (6) 206, and 207 to 216. It is. There are also paths 217 to 223 in the network.
The SPARE band of each line is line 2
07 is 0 Kbps, 208 is 64.0 Kbps, 209 is 19.2 Kbps, 210 is 0 Kbps, 2
11 is 48.0 Kbps, 212 is 14.4 Kbps, 213 is 19.2 Kbps, 214 is 0 Kbps
s, 215 is 4.8 Kbps and 216 is 38.4 Kbps.

The operation of the failure recovery system when a single line failure occurs on the line 207 in the network shown in FIG. 2 will be described below. Here, pass 217 (speed 19.2) on line 207
Kbps) and path 218 (speed 9.6Kbps) and path 219 (speed 48.0K)
bps).

As described above, the present algorithm is composed of a release phase, a broadcast phase, a return phase, and a confirmation phase.
This is shown in FIG. 5, FIG. 5, and FIG. In FIG. 2, when a line fault occurs, nodes (2) 202 and (6) 206 have a path 217, and nodes (3) 203 and (6) 206 have a path 218.
The node (1) 201 and the node (5) 205 detect a failure of the path 219, and each node executes the alarm detection processing shown in FIG. 1 (b). That is, in FIG. 3, the node (2) 202 which is the terminal node of the failed path sends the open packet 301 to the connection node (1) 201 of the failed path 217, and the node (3) 203 connects the failed path 218. The open packet 302 is sent to the node (1) 201, and the node (5) 205
The open packet 303 is transmitted to the connection node (6) 206 at 219. The node that has received the open packet performs the processing shown in FIG. 1 (c) open packet reception processing.

In FIG. 4, node (1) 201 is the SE of the faulty path 217.
Connection line 21 with free bandwidth of 19.2Kbps or more as NDER
The control packet 401 is sent to 5 and the control packet 421 is sent to line 216, and the node (3) 203
Control packets 417, 4 on connection lines 212, 213, 214, 216 having free bandwidth of 9.6 Kbps or more as SENDER 8
19, 403, 422, and the node (1) 201
Connection line 20 with 48Kbps or more free bandwidth as SENDER
The control packet 405 is transmitted to the control packet 8. Node (1)
Upon receiving the control packet 401, the control unit 201 sends the control packet 407 to the node (4) 204 that has executed the control packet reception processing shown in FIG. Similarly, control packet 42 of failure path 217
0,416,410,412,409,418 and control packet 402,406,404,408,415,411,413 of faulty path 218 and faulty path
219 control packets 414T are transmitted and received. The same operation is continued until the hop number of the control packet becomes X and disappears.

Further, in FIG. 5, the node (6) is the faulty path in FIG.
When the control packet 218 is received, the control packet reception processing shown in FIG. That is, the received control packet 408
A return packet 513 is sent back to the transmission source node (4) 204 of. Upon receiving the return packet 513, the node (4) 204 executes the return packet receiving process shown in FIG. 1E, and transmits the control node 417 having reached the minimum number of hops to the source node (3) 203, that is, the failure path. 218 SEN
At the same time as sending a return packet to the DER, a positive ack 514 is sent to the source node (6) 206 of the return packet 513. Similarly, when the node (5) 205 receives the control packet 414 of the failed path 219 in FIG. 4, the node (5) 205 executes the control packet receiving process shown in FIG. That is, a return packet 509 is returned to the transmission source node (4) 204 of the received control packet 414. Upon receiving the return packet 509, the node (4) 204 executes the return packet receiving process shown in FIG. 1E, and sends the control node 406 having reached the minimum number of hops the source node (1) 201, that is, the failure path. At the same time, a positive packet 510 is transmitted to the transmission source node (5) 205 of the return packet 509. Similarly, when the node (6) 206 receives the control packet 412 of the faulty path 217 in FIG. 4, it executes the control packet reception processing shown in FIG. 1D as SHOOSER. That is, the return packet 511 is returned to the transmission source node (5) 205 of the received control packet 412. Upon receiving the return packet 511, the node (5) 205 executes the return packet receiving process shown in FIG. 1 (e), and sends it to the transmission source node (3) 203 of the control packet 420 arriving with the minimum number of hops. The source node (6) 206 of the return packet 511 received at the same time as transmitting the return packet 507
Sends a positive ack 512 to Upon receiving the return packet 507, the node (3) 203 executes the return packet reception processing shown in FIG. 1E, and sends the control packet 421 reaching the minimum hop number at the transmission source node (2) 202, Return packet 5 sent at the same time as sending return packet 501 to SENDER of faulty path 217
A positive ACK 508 is transmitted to the transmission source node (3) 203 of 07.

The SENDER that has received the return packet executes the return packet reception processing shown in FIG. 1E, that is, the node (3) 203 that is the SENDER of the faulty path 218 returns the return packet.
Upon receipt of 506, a positive acknowledgment packet 505 shown in FIG. 5 and an end packet 602 shown in FIG.
Switch to the detour route assigned above. Similarly, upon receiving the return packet 503, the node (1) 201 which is the SENDER of the faulty path 219 returns a positive ack packet 504 shown in FIG. 5 and an end packet 607 shown in FIG. 6 to the transmission source node (4) 204. At the same time, the failed route is switched to the detour route assigned on the line 208. Also, when the SENDER node (2) 202 of the failure path 217 receives the return packet 501, it returns a positive acknowledgment packet 502 shown in FIG. 5 and an end packet 601 shown in FIG. 6 to the transmission source node (3) 203. At the same time, the failure route is switched to the bypass route assigned on the line 216.

Each node receiving the end packet in FIG. 6 executes the end packet reception processing shown in FIG. 1 (h).
That is, the end packet 606 for the faulty path 218, the end packet 605 for the faulty path 219, and the end packets 602 and 603 for the faulty path 217 are transmitted and received, and a detour connection is made.

As a result of the above processing, a route connecting the line 216, the line 213, and the line 210 as a bypass of the failure path 217, a route connecting the lines 212 and 209 as a bypass of the failure path 218, and a line serving as a bypass of the failure path 219 208 and line 211
Is established.

In this method, the operation shown in FIGS. 1A to 1H can be executed and a detour can be generated without any awareness of the case of a multi-line failure.

(Effects of the Invention) As described above, according to the present invention, since a line end is recovered (path recovery), recovery on the shortest detour route and transmission and reception of an open packet cause a failure assigned to a line on which no failure has occurred. There is an effect that it is possible to make effective use of the bandwidth to make the bandwidth of the path a spare bandwidth, and to perform recovery according to the priority of the failed path by priority transmission control of the packet.

[Brief description of the drawings]

FIG. 1 (a) is a block diagram showing the processing transition of the present invention, and FIGS. 1 (b) to 1 (i) are diagrams for explaining the operation of each part in FIG. 1 (a), and FIG. FIG. 3 is a network configuration diagram illustrating an operation example of a single failure according to the present invention; FIG. 3 is a transition diagram of transmission / reception of an open packet in a release phase of the operation example;
The figure shows the transition diagram of transmission and reception of control packets in the broadcast phase, FIG. 5 shows the transition diagram of transmission and reception of return packets, positive ack packets and negative ack packets in the return phase, and FIG. 6 shows the transition of transmission and reception of end packets in the confirmation phase FIG. In the figure, 102: interface circuit, 103: multiplex circuit, 104: switch circuit, 105: separation circuit, 106: interface circuit, 108: packet control circuit, 109 ...
Control packet reception processing unit, 110: return packet reception processing unit, 111: positive ack packet reception processing unit, 112: negative ack packet reception processing unit, 1
13: end packet reception processing unit, 114: alarm detection processing unit, 115: open packet reception processing unit, 116: packet storage queue, 117: assignment table, 118: CPU processing unit.

 ────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-3-58541 (JP, A) JP-A-3-58542 (JP, A) Globecom'88 1549-1554 C.I. H. Yang et. al "F ITNESS: FAILURE IMM UNIZATION TECHNOLO GY FOR NETWORK SER VICE SURVIVABILITY Y" IEICE Technical Report CS 89-56 IEICE Technical Report IN 89-70 Globalcom'87. 1090-1095W. D. GROVER “THE SELFHEALING ▲ Upper TM ▼ NETWORK”

Claims (4)

(57) [Claims] When a transmission path failure occurs in a network in which a plurality of nodes are stitch-connected by a transmission path, a plurality of failed paths (circuits) can exist. Assuming that both end nodes are a first node and a second node, there are a plurality of first and second node sets, and for each failed path, the first node transmits a control packet to the failed path with a spare bandwidth equal to or greater than the bandwidth of the failed path. The control packet includes at least the identifier of the failed path, bandwidth information of the failed path, a node trace indicating the node through which the control packet has passed, and a hop count indicating the number of nodes through which the control packet has passed. A node having the control information 1 and excluding the first and second nodes for each faulty path is defined as a third node. When the control packet is received, control packet reception recording is performed, and if the hop count is equal to or less than a predetermined value, the control information 1 is updated so that the node trace has no own node and the control packet of the same failure path is If the control packet has not been transmitted, the control packet is transferred to an adjacent node having a spare band equal to or larger than the band of the faulty path. When the second node receives the control packet, it is connected to the transmission line from which the control packet was transmitted. Returning a return packet to the node, the return packet including at least the identifier of the failed path, bandwidth information of the failed path, and control information 2 including a node trace indicating a node through which the return packet has passed; Node receives the return packet and receives it From the control information 2 of the turn packet and the stored control packet reception record, an adjacent node reaching the first node by the shortest path is selected from the connection transmission line having the spare band, and the node trace information for the node is selected. And transmits a positive acknowledgment packet to the transmission source node of the received return packet, and allocates a spare band of the transmission path to a detour path corresponding to the band of the failed path, and there is no connection transmission path having a spare band. For example, a negative acknowledgment packet is returned to the node that sent the return packet, and the node that has received the negative acknowledgment packet releases the band allocated to the detour path of the failed path and searches for another detour path again. 1 node assigns the fault bypass route 1 formed by the return packet Then, the end packet is sent back to the second node through the faulty detour 1, and each node on the faulty detour 1 receives the end packet and controls a switch in the node to form a faulty detour. A distributed failure recovery method, wherein when a node receives the end packet, a failure path is formed by diverting a failure path to the failure bypass circuit 1. 2. In a network in which a plurality of nodes are stitch-connected by a transmission path, if a transmission path failure occurs, a plurality of failed paths (lines) may exist. If the both end nodes are the first and second nodes, there are a plurality of first and second node sets, and the first and second nodes of each failed path send open packets to the connected nodes of the failed path, The open packet has control information 0 including at least the identifier of the failed path. When the connection node of the failed path receives the open packet, it releases the link of the failed path and transfers the open packet to the next connection node. , And the first of each failed path
Sends a control packet to all adjacent nodes having a spare band equal to or larger than the band of the failed path, and searching for an adjacent node having a spare band equal to or larger than the band of the failed path is predetermined with a certain waiting time. The control packet repeats control information 1 including at least the identifier of the failed path, bandwidth information of the failed path, a node trace indicating the node through which the control packet has passed, and a hop count indicating the number of nodes through which the control packet has passed. Nodes other than the first and second nodes for each failure path are referred to as a third node. When the third node receives the control packet, the third node performs control packet reception recording and sets the hop count in advance. If the value is equal to or less than a predetermined value, the control information 1 is updated and the If there is no own node in the trace and the control packet of the same faulty path has not been transmitted, the control packet is transferred to an adjacent node having a spare band equal to or larger than the band of the faulty path. A return packet is returned to a node connected to the transmission path from which the packet has been sent, the return packet being a node indicating at least the identifier of the failed path, bandwidth information of the failed path, and a node through which the return packet has passed. When the third node receives the return packet, the third node has control information 2 of the received return packet.
From the control packet reception record stored therein, selects an adjacent node that reaches the first node by the shortest path from among the connection transmission lines having the spare band, and returns a return packet in which node trace information is updated for the node. Transmitting, transmitting a positive acknowledgment packet to the transmission source node of the reception return packet, and allocating a spare band of the transmission path to a detour path corresponding to the bandwidth of the failure path, and sending a return packet if there is no connection transmission path having the spare band. A negative acknowledgment packet is returned to the failed node, and the node receiving the negative acknowledgment packet releases the band allocated to the alternate path of the failed path and searches for another alternate path again, and the first node returns the return path. Allocate the fault detour 1 formed by the packets and detour the end packet When the end packet is received by the nodes on the detour route 1 and the end packet is received, the switches in the nodes are controlled to form a detour route, and the second node receives the end packet. Then, a failure bypass is formed by diverting the failure path to the failure bypass circuit 1. 3. In a network in which a plurality of nodes are connected in a stitch shape by a transmission path and all paths (lines) are prioritized, when a transmission path failure occurs, a plurality of failure paths may exist. Therefore, if the nodes at both ends of the failed path are the first and second nodes, there are a plurality of first and second node sets, and the first and second nodes are provided for each failed path.
Sends an open packet to the connection node of the failed path according to the packet transmission control, the open packet has control information 0 including at least the identifier of the failed path, and is a packet transmitted on the same line as the packet transmission control. Are stored in a queue table so that packets are transmitted in descending order of priority. All packets are transmitted in accordance with the packet transmission control. The link of the path is released, the open packet is transferred to the next connection node, and for each failed path, the first node sends a control packet to all adjacent nodes having a spare band equal to or larger than the band of the failed path, and The search for an adjacent node having a spare band equal to or larger than the path band is predetermined with a certain waiting time. The control packet includes control information 1 including at least the identifier of the failed path, bandwidth information of the failed path, a node trace indicating a node through which the control packet has passed, and a hop count indicating the number of nodes through which the control packet has passed. A node excluding the first and second nodes for each faulty path is set as a third node. When the third node receives the control packet, the third node performs control packet reception recording and sets the hop count in advance. If the value is equal to or less than a predetermined value, the control information 1 is updated, and the control packet is transmitted to an adjacent node having no spare node in the node trace and no control packet of the same failed path and having a spare band equal to or larger than the band of the failed path. Transfer the second
When receiving the control packet, the node returns a return packet to the node connected to the transmission path from which the control packet was sent, and the return packet includes at least the identifier of the failed path, the bandwidth information of the failed path, and The third node has control information 2 including a node trace indicating the node through which the return packet has passed. When the third node receives the return packet, the control information 2 of the received return packet and the control information received from the stored control packet reception record are used as a backup. Selecting a neighboring node that reaches the first node by the shortest path from the connection transmission path having a band, transmits a return packet with updated node trace information to the node, and transmits the return packet to the transmission source node of the reception return packet. Sends a positive ack packet and The reserved bandwidth is allocated to the detour path corresponding to the bandwidth of the failed path, and if there is no connection transmission path having the reserved bandwidth, a negative ack packet is returned to the node that sent the return packet, and the node that received the negative ack packet fails the failed path. When the first node receives the return packet, it allocates a faulty detour 1 formed by the return packet, and releases the end packet when the first node receives the return packet. When the end packet is received by the nodes on the faulty detour 1, the switch in the node is controlled to form a faulty detour, and the end packet is sent to the second node. Is received, a faulty bypass is formed by diverting the faulty path to the faulty bypass 1 Distributed fault recovery method characterized. 4. In a network in which a plurality of nodes are stitch-connected by a transmission path, each processing of the nodes is controlled by a node type of the own node and a received packet type. There are first and second nodes at both ends of the failure path and a third node other than the first and second nodes. The received packet types include open packets, control packets, return packets, positive ack packets, negative ack packets, and end packets. Means for each of the nodes to detect the node type of the own node, means for detecting the received packet type, means for monitoring a path terminated at the own node and generating a fault alarm, Activated at the time of detection to release the bandwidth of the faulty path and perform open packet transmission processing toward the connected node. In the case of the first node, an alarm detection processing means for performing control packet transmission processing to an adjacent node having a spare band equal to or greater than the band of the failed path, and is activated upon receipt of the open packet to release the link of the failed path An open packet reception processing means for transferring the open packet to the next connection node, and a control packet broadcast process for the third node which is started when the control packet is received, and for the second node Control packet reception processing means for performing return packet return processing, and in the case of the first node which is activated upon reception of the return packet, switches the transmission path to the failure path and connects a detour to connect the return packet Sends a positive ACK packet and an end packet to the third node. If the return packet is forwarded to the only adjacent node, a positive acknowledgment packet is returned to the source of the return packet, and if there is no connection transmission path having a spare band and the transfer destination of the return packet cannot be found, the negative packet is returned. Return packet reception processing means for returning an attack packet, positive attack packet reception processing means activated when the positive attack packet is received, and negative attack packet reception processing activated when the negative attack packet is received and performing another detour search processing Means for starting the end packet reception, switching the transmission path in the case of the third node and forming a detour, transferring the end packet to the source of the return packet, and transmitting the end packet in the case of the second node End switch that completes the detour by performing a road switch 1. A distributed fault recovery apparatus comprising: a packet reception processing unit; a memory for storing all the received packets; and a table for enabling and changing a bandwidth use state of a transmission path.