JPH04105441A - Restoration system for fault of consecutive multiplex link fault - Google Patents

Abstract

PURPOSE:To restore a fault of a consecutive multiplex link fault by detecting it when a fault takes place in a consecutive link and stopping the fault recovery processing in a node included in a faulty block. CONSTITUTION:When a node detecting a fault broadcasts an ID of its own node and an ID of a node through which a path affected with the fault passes to adjacent nodes and the fault information noticed is compared to detect a fault of a consecutive link, and when the occurrence of the fault of the consecutive link is detected, the fault restoration processing of a node resident in the fault block comprising the consecutive link is cancelled. Thus, even when consecutive links have a fault simultaneously, a bypass is set to restore the path.

Description

[Detailed Description of the Invention] [Summary] The purpose of the present invention is to provide a failure recovery method that can cope with consecutive multiple link failures, with respect to a method of searching for a path to bypass nodes or links in a network when failure occurs. Each node has a path tracer generation means and a path tracer storage means, and when a link or node failure occurs, the failure detection node sets its own node as the end point of the detour route and stores the N nodes passed by the path. In the event of a failure, by broadcasting a message that designates a node as a candidate for the other end of the detour, and sending a detour setting message back to the node at the end of the detour if the node that received the message corresponds to the node candidate. In a network node in which each node autonomously searches for a detour, a failure detection node broadcasts its own node ID and the ID of the node through which the path has passed to adjacent nodes, and also transmits the notified failure information. If consecutive link failures are detected by comparing them with each other, failure recovery processing for a node located in the middle of the consecutive link failure sections is stopped.

(Industrial Application Field) The present invention relates to a method in which, when a node or link in a network fails, each node autonomously searches for a path to bypass the failed node or link by using flooding. This invention relates to a recovery method for successive multiple link failures that searches for a detour and performs failure recovery when multiple consecutive link failures occur.

In the fault recovery method by forming a detour path, when a fault occurs in a node or link in the network, a path that bypasses the faulty node or link is searched, and a cross-connect device installed in the node switches the line to recover from the fault. Create a detour that does not pass through nodes or links to recover from a failure.

If such a detour search is performed using a centralized control method, not only will the time required to search for a detour increase as the scale of the network increases, but it will also take longer if the center that searches for the detour itself fails. However, since there is a problem that the detour may stop functioning, a distributed control method is preferable, in which each node can autonomously search for a detour.

It is desired that such a detour search method using distributed control be able to cope with continuous multiple link failures.

[Background Art] When a failure occurs in a node or link in a network, it is necessary to search for a path that bypasses this node or link and use a cross-connect device to switch lines and recover from the failure. .

Conventionally, a network control center monitors the state of the line, and when a node reports a failure, it searches for a path that bypasses the failed link.

Such a centralized control method has the disadvantage that as the scale of the network increases, the time it takes to search for a detour becomes longer. Furthermore, if the network control center itself fails, there is a problem that it will no longer function at all.

As an alternative to this, a distributed control method has been proposed. As an example of a distributed control method, there is a detour search method using multiple flooding, and according to this method, it is possible to perform failure recovery in a shorter time than a centralized control method.

Here, in the detour route search method using distributed control, each node autonomously searches for a detour route by flooding.

Flooding is when a node that detects a failure sends a message specifying only the destination without specifying the route to all links connected to that node. If not, the message is relayed to all links connected to that node and sent. By repeating this procedure, the message finally reaches the destination node, and is conventionally known as a routing method in packet communication, and has extremely high reliability. It is possible to perform routing with

Even in a detour search method using distributed control, it is possible to search for a detour using such flounding.

Hereinafter, a conventional technique for failure recovery using this detour route search method will be explained.

FIGS. 4(a), (b), and (C) show examples of path tracers in the detour search method.
(a) explains the path overhead, (b) explains the read and write of the path tracer, and (C) explains the accumulation of the path tracer.

The signal between nodes consists of serial information as shown in Figure 4(a), and one frame of information consists of a section overhead for writing control information only between adjacent nodes, and a section overhead for writing control information for each path. It consists of a path overhead (shown with diagonal lines) to be written, and a path information byte (in which transmission data for each path is written).

These path tracers are sequentially connected to node A via links, as shown in FIG. 4(b).

Read/write (R/W) in B, C, -, respectively.
) to be done.

The contents of the path tracer accumulated at each node are:
As shown in FIG. 4(C), it consists of an identifier (ID) indicating the upstream node through which the path has passed. However, in this example, it is assumed that the number N of nodes stored in the path tracer is two.

FIG. 5 shows an example of the configuration of nodes in the detour search method. In the figure, reference numeral 11 indicates a link, and path information bytes 12. It is shown that it consists of a path overhead 131 shown by a solid line and a section overhead 14 shown by a dotted line.

15 is a matrix switch, and path information hide 12. A path cross-connection consisting of a path overhead 13 is performed. Reference numeral 16 denotes an alarm generating circuit (ALM), which detects a signal disconnection of the link 11 and generates a failure alarm. A section overhead interface (SEC○VHINF) 17 performs section overhead interface processing with the central processing unit (CPU) 21.

18 is the path overhead interface (PATH
OVHI NF) performs interface processing between the path overhead 13 and the CPU 21.

19 is the matrix switch ink face (SW
INF), matrix switch 15 and CP
Performs interface processing with the TJ21.

A L M2S, S E COVHINF17. P.A.
THOVHINF18,5WrNF19 and CPU U21
are connected by a path 20.

CPU 21 controls the overall operation in the node. Further, a storage device (MEM) 22 connected to the path 20 stores information such as the above-mentioned path tracer and the usage status of each path.

FIG. 6 shows an example of the configuration of a network in the detour route search method, and includes nodes B, C, D, node X, and Y indicated by 31 to 35, and links BC, CD, and links indicated by 36 to 39. XC, link YC, links 40 to 45, and paths indicated by 46 to 49 (paths 1 to 4
) and detour routes (paths 1 to 4) indicated by 50 to 52 are shown.

As shown in FIG. 6, under normal conditions, paths 1 to 4 each pass through the next node.

Path 1: Node B - Node C - No FD Path 2: Node B - Node C - No FD Path 3: Node X - Node C -
Node D path 4: Node Y - Node C - Node D It is now assumed that a failure occurs in node C, node D detects this as a signal disconnection in link CD, and a failure alarm is generated.

In node D, the detour search software is activated by this alarm and executes the following processing.

(1) First, a path in the link CD that passes through the node D where the alarm has occurred is searched to identify the path affected by the fault (faulty path). The failed paths in this case are paths 1 to 4.

(2) Examine the path tracers of paths 1 to 4 stored in node D, divide them into groups having the same path tracer, and calculate the number of paths belonging to each group. The groups and the number of paths belonging to each group in this case are as follows.

Group 1 (Node C - Node B): 2 pieces Group 2
(Node C - Node A detour message is sent to neighboring nodes using the section overheads of all output links that have protection paths connected to node D. Each node constantly searches for the backup path and stores the information in the storage device. Please write the following information in this detour search message (.

■ ID of the source node of the detour search message (starting point of the detour) ■ ID of the first node written in the path tracer
(Candidate for the end point of the detour) ■ ID of the 20th node written in the path tracer
(Candidates for the end point of the detour) ■ Number of paths that require detour ■ IDI of path that requires detour ■ ID2 of path that requires detour ■ Requires detour In the case of the path IDM, for example, group 1, the information of the above-mentioned (1) to (2) is as follows.

■ Node D ■ Node C ■ Node B ■　2 ■Pass 1 ■Pass 2 becomes.

(4) The node that receives the detour search message reads the above-mentioned fields ■ and ■ in the detour search message, and checks whether its own node is a candidate for the end point of the detour.

If the own node is not a candidate for the end point of the detour,
This detour search message is relayed and sent to all output links that have a backup path.

If the own node is a candidate for the end point of the detour,
A detour setting message having the same content as the detour route search message is sent back to the source of the detour route search message via a route opposite to that of the detour route search message.

In the case of Group 1 mentioned above, the candidates for the nodes that will be the end points of the detour are Node C and Node B, but since Node C is in a failure state, it does not send back a detour setting message, and only No FB is used for the detour. route setting messages can be sent back.

The backup path selected each time a node is passed is written in the detour search message, and each node passed through remembers the content of the detour search message.
The detour setting message can reach the source node D through the reverse route of the detour search message from the backup path written in the message and the backup path stored in the node. The end point of the route is determined, and a detour route is determined.

In this way, it is possible to simultaneously determine the end points of the detour and search for a detour route between the end points.

Note that although the above explanation has been made regarding the case of a failure in a node, a failure in a link can also be dealt with in the same way.

(Problem to be solved by the invention) However, in the detour search method using such a distributed control method, when two consecutive links fail at the same time, depending on the form of the detour, it may be possible to recover, or it may not be possible to recover. There are cases where it is not possible.

FIG. 7 shows a series of recoverable link failures.

As shown in FIG. 7, if a failure occurs simultaneously in link B-C between node B and node C and link C-D between node C and node D, then If there is a detour 11 connecting the two nodes, and a detour 12 connecting the node C-0, each node B.

By switching the path in the detour direction at C and D, the path from node A to node E can be restored.

Note that in FIG. 7, the arrow indicates the direction in which the detour search message 7 is sent.

FIG. 8 shows an unrecoverable continuous link failure.

In FIG. 8, a detour 13 connecting nodes A and C,
If there is only the detour 14 connecting nodes B and D, and the detour 14 is set, the detour 14 at node A is
The switched path reaches node C, but is not connected to any node beyond that, so the path is not restored.

In this way, when two consecutive multiple link failures occur, there is a problem in that the path may become unrecoverable.

However, even in this case, if the failure recovery process is canceled at node C in the middle of consecutive failed sections, and the setting of detour route 14 is canceled to leave only detour route 13, nodes B and D will be connected to the detour direction. It is clear that the path from node A to node E can be restored by switching the path.

The present invention aims to solve the problems of the prior art as described above, by detecting a failure in consecutive links and performing failure recovery processing at a node included in the middle of the failure section. The purpose of this is to enable recovery from consecutive multiple link failures by canceling the link.

[Means for Solving the Problems] The present invention, as shown in the principle configuration in FIG. By the means 1, the IDs of the N nodes just before each path passes through the node are sent to the path tracer storage means 2, and each node through which the path passes accumulates the IDs. When a failure occurs in a link or a node, the detour search message sending means 3 causes the node that detects the failure to make its own node the end point of the detour, and to determine the number of paths affected by the failure (N) that have passed through.
These nodes are written as candidates for the other end of the detour, and these are written in a message and broadcast, and the detour setting message return means 4 determines whether or not the node that received this message corresponds to the N node candidates. When a failure occurs, each node autonomously searches for a detour by sending a failure notification message to the nodes in the network by determining the detour and sending a detour setting message back to the node at the end of the detour if applicable Providing means 5,
When a failure is detected, the node that detected the failure broadcasts its own ID and the ID of the node through which the path affected by the failure has passed, to adjacent nodes,
Continuous link failure detection means 6 is provided to detect successive link failures by mutually comparing the notified failure information, and failure recovery processing canceling means 7 is provided to detect successive link failures. If this is detected, the failure recovery process for a node located in the middle of a failed section made up of consecutive links is canceled.

In addition, in such a recovery method for continuous multiple link failures, before a node that detects a failure sends a detour search message, it must check its own node's ID and the ID of the node through which the path affected by the failure has passed. This system broadcasts a failure notification message stating the following.

Furthermore, in such a recovery method for continuous multiple link failures, when a detour search message arrives, the failure detection node ID included in the message and the node ID through which the path affected by the failure passed are stored. and means for broadcasting the detour search message regardless of the presence or absence of spare capacity of the output link of the node.

In addition, in these continuous multiple link failure recovery methods, when the own node is the failure detection node, the failure detection node ID and the path affected by the failure are determined from the received failure notification message or detour route search message. means for extracting a passed node ID; means for detecting that the own node is the last node among the nodes passed by the extracted path affected by another failure; and detecting the event. means to release the spare secured by the detour search message sent by the own node and send a cancellation message to the link that secured the spare, and when the cancellation message is received, the detour search message corresponding to this is secured. This system is provided with means for releasing the reserved spare and sending a cancellation message back to the link that secured the spare and the link that received the detour search message.

Furthermore, in these continuous multiple link failure recovery methods, the failure detection node ID and the node ID through which the path affected by the failure passed are extracted and stored from the received failure notification message or detour route search message. When the node that detected a failure is the last node through which a path affected by another failure passes, it reads out the accumulated information and aborts failure recovery for the node that detected the failure. a means for determining that the node is a node; a means for sending back a cancellation message when a detour search message is received from a failure recovery cancellation note; A means for releasing the reserve reserved by the search message and sending an abort message to the link that sent the detour search message, and a means for releasing the reserve reserved by the corresponding detour search message when receiving the abort message from another node. The system is provided with means for releasing the reserve and sending a cancellation message to the link that sent the detour search message.

(Operation) At each node, the IDs of the previous N nodes that have passed before each path passes through the node are sent to the path overhead, and each node that the path passes accumulates them, and When a node fails, the node that detects the failure sets its own node as the end point of the detour, and sends a message to the N nodes that the path affected by the failure has passed as candidates for the other end of the detour. The node that received this message is written to and broadcast, determines whether the node that received this message corresponds to N node candidates, and if so, sends a detour setting message back to the node at the end of the detour. At the time, each node autonomously searches for a detour, and when a failure is detected, the node that detected the failure enters its own ID,
I of the node traversed by the path affected by the failure
By broadcasting D to adjacent nodes and mutually comparing the notified failure information, failures of consecutive links are detected. Since failure recovery processing is canceled for a node located in the middle of a failed section made up of links, even if consecutive links fail at the same time, it is possible to set a detour and restore the path.

In this case, before a node that detects a failure sends a detour search message, it broadcasts a failure notification message that includes its own node ID and the ID of the node through which the path affected by the failure has passed. You can.

Also, when a detour search message arrives.

The failure detection node ID included in the message and the node ID passed by the path affected by the failure are accumulated, and the detour route search message is sent regardless of the presence or absence of spare capacity of the output link of that node. may also be broadcast.

In these cases, when the own node is the failure detection node, the failure detection node ID and the node ID through which the path affected by the failure passed are extracted from the received failure notification message or detour route search message. ,
When the own node detects that it is the last node among the nodes that have passed through the path that has been affected by another extracted failure, it releases the reserve secured by the detour search message sent by the own node. , sends a cancellation message to the link that has secured the spare, and when the cancellation message is received, releases the spare that was secured by the detour search message in response to this, and receives the link that secured the spare and the detour search message. An abort message may be sent back to the linked link.

Furthermore, in these cases, the fault detection node ID and the node ID through which the path affected by the fault passed are extracted and stored from the received fault notification message or detour route search message, and the stored information is stored. When the node that detected a failure is the last node passed by a path affected by another failure, the former failure detection node is determined to be the failure recovery abort node, and failure recovery is aborted. When a detour search message is received from a node, a cancellation message is sent back, and if a detour search message has already been received from a node that has canceled failure recovery, the spare reserved by the detour search message is released and the A cancellation message is sent to the link that sent the detour search message, and when the cancellation message is received from another node, the corresponding detour search message releases the reserved reserve and sends the detour search message. A link cancellation message may be sent.

Embodiment: When a failure occurs, a node that detects the failure broadcasts a failure notification message to all output links before broadcasting detour search messages from other nodes. The failure notification message has almost the same content as the detour route search message, and includes the ID of the source node and the ID of the passing node affected by the failure on the path.
Contains.

The second example illustrates the broadcast of a failure notification message in one embodiment of the present invention.

Now, as shown in the figure, link B between node B and node C
-C and link C-D between node C and node D at the same time, the failure of link B-C is detected at node C as a signal disconnection. Similarly, link C-D
failure is detected by Note D as a signal loss.

Here, it is assumed that two consecutive links are faulty at the same time, so the nodes passed through in the fault notification message sent by node C are node B and node A. On the other hand, the node through which the failure notification message sent by the node D is the node FC and the node B.

When a failure notification message arrives at a neighboring node, the neighboring links broadcast it to all output links. Like the detour search message, the failure notification message has a hop count indicating how many nodes it was sent through, and if the number exceeds the predetermined hop limit, Discarded without being broadcast.

In this way, the failure notification message sequentially passes through adjacent nodes and reaches the node through which the path affected by the failure passes. When a detour search message from another source arrives at a node that has detected a failure, that node stores the failure notification message and checks whether consecutive links have failed.

In this case, the check is performed by checking whether the ID of the own node is present at the beginning of the ID of the node through which the path affected by the failure is written in the received failure notification message. Hello, own node ID
If there is, it can be seen that the link from the local node to the source of the message has also failed.

FIG. 3 illustrates a comparison of destination nodes of failure notification messages.

In Figure 3, li sent from source node D
In the damage notification message, node C is listed as the first node ID field of the path affected by the failure, so node C determines that adjacent node D has also detected the failure. . Therefore, in this case, link C
- It can be seen that D has caused a problem.

On the other hand, since node C does not receive any signal from node B, it knows that link B-C has failed.

Therefore, at this time, in node C, it can be determined that link BC and link CD have failed at the same time, and that both links are continuous.

In this way, if failures in consecutive links are detected and the own node is found to be in the middle of the failure section,
In order to cancel the failure recovery (detour route search) that the node is executing, that node (node C) sends a cancellation message to the link that broadcast the detour route search message.

The adjacent node that receives the abort message examines the corresponding detour search message, releases the reserve assigned to it, and sends the abort message to the link to which the reserve was allocated.

As a result, the detour route 14 set from node C to node A is canceled, and the path is restored.

In the above embodiment, a failure notification message is broadcast in order to notify adjacent nodes of the occurrence of a failure, but a detour search message may also be used to do the same.

In this case, when the detour route search message arrives, the adjacent node that received the detour route search message will identify the failure detection node ID included in the message and the node ID through which the path affected by the failure has passed. is stored, and the detour search method 2 is broadcast even if there is no spare output link. However, if there is no spare, the field of the secured detour route band field in the detour search message is set to O. As a result, the fault notification function can be performed without performing detour setting processing.

Furthermore, in the above embodiment, the source node of the detour search message that detects failures in consecutive links starts the process of canceling failure recovery, but adjacent nodes also receive failure notification messages or With the ability to accumulate pathfinding messages and detect successive link failures, it is also possible for these nodes to initiate an abort of failure recovery.

In this case, from the received fault notification message or detour route search message, the fault detection, note ID, and the note ID through which the path affected by the fault has passed are extracted and stored. Read out this accumulated information and
When the node that detected a fault is the last node traversed by a path affected by another fault,
The former failure detection node is determined to be a failure recovery abort note. Then, when a detour search message is received from a failure recovery cancellation node, a cancellation message is sent back.

If a detour route search message has already been received from the failure abort node, the spare reserved by the detour route search message is released, and an abort message is sent to the link that sent the detour route search message. When an abort message is received from another node, the spare reserved by the corresponding failure recovery message is released and the detour search method is activated.
Sends an abort message to the link that sent the message.

In this way, the received and accumulated failure notification messages or detour search messages are mutually referenced, and the sender ID of one message is in the first field of the D field for the note affected by the failure of another message. Check whether it is correct.

In addition, the adjacent node that detects the failure of consecutive links in this way broadcasts an abort message in the direction from which the detour search message was sent, and also sends an abort message back to the link where the message arrived, in the same way as described above. This allows neighboring nodes to initiate abort of failure recovery.

〔Effect of the invention〕

As explained above, according to the present invention, even if multiple failures occur in consecutive links, it is possible to set a detour and restore the path, thereby improving the network recovery rate. .

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the basic configuration of the present invention, FIG. 2 is a diagram illustrating broadcasting of a failure notification message in an embodiment of the invention, and FIG. 3 is a diagram illustrating a comparison of destination nodes of failure notification messages. Figure 5 is a diagram showing an example of the configuration of a node in the detour search method, Figure 6 is a diagram showing an example of the network configuration in the detour route search method, and Figure 7 is a diagram showing an example of a recoverable continuation. Diagram showing link failure, No. 8
The figure shows continuous link failures that cannot be recovered. 2 is a path tracer generation means, 2 is a path tracer storage means, 3 is a detour search message sending means, 4 is a detour setting message returning means, 5 is a failure notification message sending means, 6 is a continuous link failure detection means, Reference numeral 7 indicates a means for canceling failure recovery processing.

Claims (5)

[Claims] (1) In each node of a network in which multiple nodes are sequentially connected by links including multiple paths, the IDs of the previous N nodes that have passed before each path passes through the node are passed to the path overhead. A path tracer generation means (1), a path tracer storage means (2) in which each node through which the path accumulates the generated information, and a path tracer storage means (2) in which each node through which the path passes accumulates it, and when a failure occurs in a link or a node, a node that detects the failure detours its own node. a detour search message sending means (3) for writing and broadcasting the N nodes through which the path affected by the failure has passed in a message as candidates for the other end points of the detour; Detour route setting message return means (4) that determines whether or not the node that received this message corresponds to one of the N node candidates, and if so, sends back a detour route setting message to the node at the end point of the detour route. In preparation, when a failure occurs, each node in a network autonomously searches for a detour. failure notification message sending means (5) for broadcasting to adjacent nodes the ID of the node through which the path has passed; a fault detection means (6); and a fault recovery processing canceling means (7) for canceling fault recovery processing for a node located in the middle of a fault section consisting of consecutive links when it is detected that a fault has occurred in a continuous link. ) A recovery method for successive multiple link failures. (2) In the recovery method for consecutive multiple link failures according to claim 1, before a node that detects a failure sends a detour search message, the self-node ID and the path affected by the failure pass through. 1. A method for recovering from continuous multiple link failures, characterized by comprising means for broadcasting a failure notification message containing an ID of a node that has failed. (3) In the recovery method for consecutive multiple link failures according to claim 1, when a detour search message arrives, the failure detection node ID included in the message and the path affected by the failure are A recovery method for consecutive multiple link failures, comprising means for accumulating passed node IDs, and means for broadcasting a detour search message regardless of the presence or absence of spare capacity of the output link of that node. (4) In the recovery method for consecutive multiple link failures according to claim 1, 2, or 3, when the own node is the failure detection node, the failure is detected from the received failure notification message or detour route search message. Detected node ID and node ID passed by the path affected by the failure
means for detecting that the self-node is the last node among the nodes through which the path affected by the extracted other faults has passed; A means for releasing the reserve reserved by the detour route search message sent by a node and sending an abort message to the link that secured the reserve, and a means for releasing the reserve reserved by the detour route search message sent by the node, and when receiving the abort message, releasing the reserve reserved by the detour route search message in response to the abort message. 1. A method for recovering from consecutive multiple link failures, characterized in that the method includes means for sending an abort message back to the link that released the link and secured the reserve, and the link that received the detour search message. (5) In the recovery method for consecutive multiple link failures according to claim 1, 2, or 3, the failure detection node ID and the influence of the failure are determined from the received failure notification message or detour route search message. A method for extracting and storing the IDs of nodes passed by paths affected by other faults, and a means for reading out this accumulated information so that a node that has detected a fault is the last node passed by a path affected by another fault. means for determining the former failure detection node as a failure recovery abort node; a means for sending back an abort message when a detour search message is received from a failure recovery abort node; When a detour search message is received, a means for releasing the reserve secured by the detour search message and sending an abort message to the link that sent the detour message, and a means for sending an abort message from other nodes. Recovery from continuous multiple link failures, characterized by comprising means for, when received, releasing the reserve secured by the corresponding detour search message and sending an abort message to the link that sent the detour search message. method.