JP2011124757A - Fault-link specifying system and method of determining monitoring path of the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a fault-link specifying system capable of specifying a single link fault without fail, and to provide a method of determining a monitoring path thereof. <P>SOLUTION: In the monitoring-path setting unit 21 of a fault management device 2, a topology acquisition unit 100 acquires the physical real topology of a network NW to be monitored. A required monitoring-path number obtaining unit 102 executes first integer programming to obtain the required number of monitoring paths Mmin so that any of the monitoring paths passes through all the links of the network NW to specify the single fault link. A monitoring-path group obtaining unit 103 repeatedly executes second integer programming, incrementing the number of monitoring paths M by one from the Mmin, until an executable solution of the group of monitoring paths which are minimum for specifying the fault link and have the minimum total path length can be obtained, where the required number of monitoring paths Mmin is set as an initial value of the number paths M of monitoring. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a system for identifying a faulty link in a network and a method for determining a monitoring route thereof, and more particularly, by simply transmitting monitoring data from a plurality of monitoring devices to each monitoring route while distributing the processing load of each monitoring device. The present invention relates to a faulty link identification system that can identify a single fault without omission and a monitoring route determination method thereof.

  Patent Documents 1 and 2 disclose techniques for detecting a failure between a monitoring device and a server device by transmitting monitoring data from the monitoring device to the server device on the network and monitoring the response time. .

  In addition, monitoring data is sent from the monitoring device to the server device through various monitoring routes, and the failure of each route is detected by monitoring the response time, and the link through which the route in which the failure is detected passes is failed. A technique for specifying a link is disclosed in Patent Document 3.

  However, in the above-described prior art, the determination of the monitoring path through which the monitoring data passes is left to the operator's experience. Therefore, in order to identify the faulty link when assuming a single link fault, a lot of monitoring is required. There has been a technical problem that a route or a long monitoring route is required, the load of monitoring processing increases, and the identification of a fault link is complicated.

  In order to solve such a technical problem, the inventors of the present invention transmit a pseudo service request using a plurality of monitoring paths that connect the monitoring device and the server device to the monitoring device on the network. Based on the response, a technology for identifying a single link failure in the network was invented and a patent application was filed (Patent Document 4).

JP 2002-259243 A JP 2004-171495 A Japanese Patent No. 3885931 Japanese Patent Application No. 2009-238281

  In the technique of Patent Document 4, a system in which each of the monitoring device and the server device is assumed is assumed, and a plurality of monitoring devices and server devices cannot be connected. Therefore, there has been a technical problem that the processing load on the monitoring device and the server device increases as the network size increases and the number of monitoring paths increases.

  An object of the present invention is to solve the above-described problems of the prior art and provide a fault link identification system and a monitoring path determination method thereof that can distribute the processing load of path monitoring by connecting a plurality of monitoring devices to a network. is there.

  In order to achieve the above object, the present invention connects a plurality of quality control devices each having at least one of a monitoring data transmission function and a reception function to a node on a network, and the transmission function changes to a reception function. In a failure link identification system for transmitting monitoring data through a plurality of monitoring routes passing through a relay node and identifying a failure link based on the reachability, and a method for determining the monitoring route,

  A means (procedure) for acquiring the actual topology of the network, a means (procedure) for transforming the real topology into a virtual topology, and the number of monitoring paths required to identify a single link failure in the network based on the virtual topology. Means (procedure) to be solved by the first integer programming method, and the number of necessary monitoring paths as an initial value of the number of monitoring paths. And means (procedure) for solving the problem while increasing the number of monitoring paths until an executable solution is obtained by the second integer programming. And, in the means for transforming the topology, a pseudo-node connected by a virtual link to a plurality of nodes to which a quality management device that exhibits a transmission function is connected, and a quality management device that exhibits the reception function are connected A pseudo arrival node connected to a plurality of nodes by a virtual link is virtually arranged, and the first solving means finds the number of monitoring routes from the pseudo origin node to the pseudo arrival node, and the second solving means Then, the link length of the virtual link is set to zero.

  According to the present invention, the following effects are achieved.

  (1) According to the fault link identification system of the present invention, the load path of the monitoring process is distributed by connecting a plurality of quality monitoring apparatuses to the network, and the minimum sum of path lengths can be specified without fail while the fault link is always specified. By using the monitoring path group, the location where the single link failure occurs can be specified accurately.

  (2) According to the monitoring route determination method of the present invention, in a failure link identification system in which a plurality of quality monitoring devices are connected to a network and the load of monitoring processing is distributed, the minimum and path length that can always identify a failure link A monitoring path group having the smallest sum can be logically determined based on the topology of the monitored network.

It is the block diagram which showed the structure of the network to which this invention is applied. It is the block diagram which showed the structure of the principal part of a failure management apparatus. It is a figure of the virtual topology obtained by transforming the real topology of the network. It is a figure for demonstrating the specific principle of the failure link in this invention. It is the flowchart which showed the operation | movement of this invention. It is the figure which showed the calculation method of the number of the paths which pass a network device node.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a network to which a faulty link identification system and its monitoring route determination method of the present invention is applied. A monitoring target network NW includes a number of network device nodes n and their mutual connections. And a number of links connected to the network.

  In the present invention, a quality monitoring device 1 having at least one of a monitoring data transmission function (t) and a reception function (r) is connected to a plurality of network device nodes n, respectively, and quality monitoring with a transmission function is provided. A monitoring path is set from the device 1 (t) to the quality monitoring device 1 (r) having a reception function. In the present embodiment, it is assumed that all the quality monitoring devices 1 have both a monitoring data transmission function (r) and a reception function (r).

  The failure management apparatus 2 instructs the quality monitoring apparatus 1 (t) having the monitoring data transmission function (t) to transmit monitoring data through the monitoring path. The quality monitoring device 1 (r) having the monitoring data receiving function monitors the quality of the monitoring data received through the monitoring path, and notifies the failure management device 2 when quality deterioration is detected. The failure management device 2 identifies a failure link based on information related to a link through which each monitoring route passes and monitoring route information in which quality degradation is detected.

  In the present invention, the failure management apparatus 2 determines a monitoring route group that can always identify a failure link. In other words, assuming a single link failure, a monitoring route group having a minimum total route length that can always identify the failed link is determined. However, except for the network device node n to which the quality monitoring device having both the monitoring data transmission function (t) and the monitoring data reception function (r) is connected, the network device node n with the same monitoring path is transmitted multiple times. Assume that there is no transit.

  FIG. 2 is a block diagram showing the configuration of the main part of the failure management device 2, and transmits monitoring data from each quality monitoring device 1 to each monitoring route, and a monitoring route setting unit 21 for setting a plurality of monitoring routes. And a link failure identification unit 22 that identifies the failure link. The link failure identification unit 22 transmits monitoring data from each quality monitoring device 1 (t) to the set plurality of monitoring routes, and the monitoring data does not reach the quality monitoring device 1 (r). It is determined that the monitoring data arrives very slowly, or that a failure has occurred on a route that has a very high error rate in the arrived monitoring data. The failure monitoring route has passed, and a failure has occurred. A link that does not pass through any monitoring path is identified as a failed link.

  In the monitoring route setting unit 21, the topology acquisition unit 100 acquires the physical real topology of the monitoring target network NW. The topology transformation unit 101 converts the real topology of the monitoring target network NW shown in FIG. 1 into the virtual topology shown in FIG. 3, as will be described in detail later. As will be described in detail later, the necessary monitoring path number solving unit 102 is required to allow any one of the monitoring paths to pass through all the links of the monitoring target network NW to identify a single failure link. The number Mmin is solved by executing the first integer programming. The monitoring path group solving unit 103 uses the necessary monitoring path number Mmin as an initial value of the monitoring path number M, until an executable solution of the monitoring path group having the minimum total path length and the minimum total path length is determined. The second integer programming is repeatedly executed while increasing the number M of monitoring routes by 1 from Mmin.

  FIG. 4 is a diagram for explaining a failure link identification principle in the link failure identification unit 22. When focusing on the incoming links l1, l2, and l3 that have one network device node (relay node) n0 as the end point and the outgoing links l4 and l5 that have the starting point, the incoming links l1 of the two monitoring paths p1 and p2 If the outgoing link l4 is common, even if a failure occurs in each of the monitoring paths p1 and p2, it is caused by the failure of the incoming link l1 or due to the failure of the outgoing link l4 It cannot be determined whether it is a thing.

  On the other hand, if the incoming and outgoing links of one monitoring path p3 are l2 and l5, and the incoming and outgoing links of the other monitoring path p4 are l3 and l5, the path p3 If only the failure occurs, the incoming link l2 can be identified as the failed link, and if only the route p4 fails, the incoming link l3 can be identified as the failed link, and both the routes p3 and p4 fail. If this occurs, the outgoing link l5 can be identified as a failed link.

  As described above, according to the present invention, it is possible to reliably identify a single link failure by imposing a restriction that only the same route does not pass through any two links (link pairs) on the network. .

  Next, the operation of the embodiment of the present invention will be described with reference to the flowchart of FIG. In step S1, the physical real topology of the monitored network NW is acquired by the topology acquisition unit 100. In step S <b> 2, the acquired real topology is converted into a virtual topology (N, Link) by the topology deforming unit 101.

  FIG. 3 is a diagram showing an example of a virtual topology obtained by modifying the real topology shown in FIG. 1, and a quality monitoring device that exhibits both the monitoring data transmission function (t) and the reception function (r). 1 (FIG. 1) is connected to the network device node n (t) to which the quality monitoring device 1 having only the transmission function (t) is connected and only to the reception function (r). It is virtually divided into a network device node n (r) to which the quality monitoring device 1 is connected. Only the link starting from the node n (t) is connected to the former network device node n (t). Only the link whose end point is the node n (r) is connected to the latter network device node n (R).

  In the virtual topology, a pseudo origin node s and a pseudo destination node d are newly added, and the pseudo origin node s is connected to a quality monitoring device 1 (t) that exhibits a monitoring data transmission function (t). All network device nodes n (t) are connected by a virtual link with a distance of zero. Similarly, all the network device nodes n (r) connected to the quality monitoring device 1 (r) that performs the monitoring data receiving function (r) are connected to the pseudo arrival node d by a virtual link having a distance of zero. The

  In step S3, the necessary integer number of monitoring routes Mmin is obtained by executing the first integer programming for the virtual topology (N, Link) by the necessary monitoring route number solving unit 102. Here, each constant is defined as follows.

n: A node constituting the network.
N: A set of nodes n.
l: Links that make up the network.
L: A set of links l.
s: Simulated source node.
d: Pseudo destination node.
VL: Set of virtual links
MN: A set of nodes whose incoming links or outgoing links are only virtual links
nin: A set of links that end at node n. However, sin is an empty set.
nout: A set of links starting from node n. However, dout is an empty set.
| Nin |: Number of links (number of incoming links) excluding virtual links that end at node n.
| Nout |: Number of links excluding virtual links starting from node n (number of outgoing links).

The variables are defined as follows:
pl: Number of monitoring routes passing through link l.
Mmin: The number of necessary monitoring paths connecting the pseudo source node s and the pseudo destination node d.

  In the first integer programming method, (i) the number of monitoring paths exiting from the pseudo origin node s is Mmin, (ii) the number of monitoring paths entering the pseudo destination node d is Mmin, (iii) pseudo For nodes other than the departure and arrival nodes s and d, based on the three path conservation rules that the sum of the number of monitoring routes coming in from each incoming link is equal to the sum of the number of monitoring routes going out from each outgoing link The following equations (1), (2), and (3) are the constraint equations.

  Furthermore, in order to detect all single link failures, it is necessary for at least one monitoring path to pass through all the links constituting the monitored network NW excluding the virtual links. Therefore, the following equation (4) is set as a condition of the path group for detecting all single link failures.

  Further, as described above with reference to FIG. 4, in order to identify a single link failure, at least all the adjacent link pairs of each relay node excluding the virtual link have no monitoring path passing through the two links. There is a constraint that they must not be the same. Therefore, as shown in FIG. 6, the number of monitoring paths passing through the node n where the number of incoming and outgoing links is m and n must be greater than or equal to Max (m, n, 3/2 (m + n)). I must.

  That is, assuming that m ≧ n, if 2n ≧ m ≧ n, in order to identify a failure between 2 (mn) incoming links and (mn) outgoing links, different incoming sides Two (mn) monitoring paths that pass through the link and pass through the same outgoing link two by two must pass through the node n. Furthermore, in order to identify the failure of the remaining 2n-m ingress links and 2n-m egress links, it is necessary to pass 2 different ingress links and 2 each through the same egress link. (2n-m) / 3 monitoring routes and 2 (2n-m) / 3 monitoring routes that pass through the same ingress link and two different egress links two by two. Need to pass. Therefore, at least 2/3 (m + n) monitoring routes in total need to pass through the node n.

  If m ≧ 2n ≧ n, in order to identify the failure of 2n incoming links and n outgoing links, 2n passing through different incoming links and passing through the same outgoing link two by two The book monitoring path must pass through node n. Furthermore, in order to identify the failure of the remaining m-2n incoming links, m-2n monitoring routes that pass through the remaining m-2n incoming links and any outgoing links Needs to pass through node n. Therefore, at least m monitoring routes in total need to pass through the node n.

  When 2n ≧ m, 2/3 (m + n) ≧ m, and when m ≧ 2n, m ≧ 2/3 (m + n). In order to identify a failure in the side link and n outgoing links, at least Max (m, 2/3 (m + n)) monitoring routes need to pass through the node n. Similarly, if m ≦ n, at least Max (n, 2/3 (m + n)) monitoring routes are used to identify failures in m ingress links and n egress links. Needs to pass through node n. From the above, it is necessary that at least Max (m, n, 2/3 (m + n)) monitoring paths pass through the node n in the case of m ≧ n and m ≦ n.

  From the above, the constraint equation of the following equation (5) is established for all the nodes n except the pseudo source node s, the pseudo destination node d, and the node whose incoming link or outgoing link is only a virtual link. However, here, the monitoring path passing through the virtual link is counted as the number of monitoring paths, but the virtual link is not counted as the number of incoming / outgoing links.

  In the first integer programming method, the objective function to be minimized is the necessary number of monitoring paths Mmin, and the necessary number of monitoring paths Mmin is calculated by solving the above equations.

  In step S4 and subsequent steps, the second integer programming method is executed under the given number of monitoring routes M, and the total route length of the monitoring route group connecting the pseudo origin node s and the pseudo destination node d is minimum. It is asked to become. In the present embodiment, the necessary number of monitoring routes Mmin obtained by executing the first integer programming method is adopted as the initial value of the number of monitoring routes M, and the monitoring route is obtained until an executable solution of the monitoring route group is obtained. The execution of the second integer programming is repeated while increasing the number M by one.

  That is, in step S4, the necessary number of monitoring routes Mmin is set as the initial value of the number of monitoring routes M. In step S5, the monitoring path group solving unit 103 executes a second integer programming method for finding a monitoring path group with a minimum total path length that can always identify a faulty link. In this second integer programming, constants are defined as follows:

LP: A set of link pairs excluding virtual links.
(l1, l2): Link pair.
dl: Link l distance
M: Number of monitoring paths connecting the pseudo source node s and the pseudo destination node d.

  Furthermore, the following variables are defined here.

Xl (k): A binary variable that is “1” if the k (1 to M) th monitoring path passes through the link l and “0” if it does not pass.
Y (l1, l2) (k): A binary variable that is “1” if the k (1 to M) th monitoring path passes through the link pair (l1, l2), and “0” otherwise.
SumLength: Total route length in the monitoring route group.

  The constraint equation in the second integer programming method is as follows. That is, since all the monitoring routes pass through one of all the links l starting from the pseudo source node s, the following equation (6) is one of the route conservation rules.

  Similarly, since all the monitoring routes pass through any one of all the links l having the pseudo arrival node d as an end point, the following equation (7) is one of the route conservation rules.

  All the monitoring routes pass through the outgoing link 1 starting from each relay node other than the pseudo arrival / departure nodes s and d and the incoming link 1 as the end point only once, or neither pass, so the following formula (8 ) Is one of the path conservation laws.

  Since any one of the monitoring paths must pass through all links except the virtual link, the following equation (9) is a condition of the path group for detecting all single link failures.

  Further, the following equation (10) is a condition for the monitoring path to pass through each link pair excluding the virtual link.

  Furthermore, for all link pairs except virtual links on the network, all of the paths that pass through one link pass through the other link, and all of the paths that do not pass through one link do not pass through the other link. Then, since the failure of the one side link and the failure of the other side link cannot be identified, the path passing through the one side link and the path passing through the other side link must not be exactly the same. Therefore, the following equation (11) is a condition of the path group for specifying all single link failures.

  In this embodiment, since the monitoring path is determined so that the total of the monitoring path length is minimized, the objective function to be minimized is given by the following equation (12).

  In this embodiment, the value of the binary variable Xl (k) (1 to M) obtained by executing the second integer programming is a minimum number M of monitoring route groups in which the total route length is minimized. Will show.

  In step S6, it is determined whether the monitoring route group has been successfully solved. If the solution is unsuccessful, the process proceeds to step S7, increments the number M of monitoring routes, returns to step S5, and the second integer programming is executed again.

1 ... Quality monitoring device 2 ... Failure management device

Claims (6)

A plurality of quality control devices each having at least one of a monitoring data transmission function and a reception function are connected to nodes on the network, and monitoring data is transmitted from the transmission function to the reception function through a plurality of monitoring paths passing through the relay node. In a failed link identification system that identifies a failed link based on its reachability,
Comprising a failure management device for solving a monitoring path group including the plurality of monitoring paths;
The failure management device is
Topology acquisition means for acquiring the actual topology of the network;
Topology modifying means for transforming the real topology into a virtual topology;
First solving means for solving the number of monitoring paths necessary for identifying a single link failure in the network based on the virtual topology by a first integer programming;
Using the number of necessary monitoring paths as an initial value of the number of monitoring paths, a feasible solution can be obtained by the second integer programming method for a monitoring path group that can identify all single link failures and minimize the total path length. A second solution means for solving the problem while increasing the number of monitoring routes until
The topology transformation unit includes a pseudo-originating node connected by a virtual link to a plurality of nodes connected to a quality management device that exhibits the transmission function, and a plurality of nodes connected to a quality management device that exhibits the reception function Virtually placing a pseudo-arrival node connected with a virtual link,
In the first solving means, the number of monitoring routes from the pseudo origin node to the pseudo destination node is obtained,
In the second solution solving means, the link length of the virtual link is set to zero.   The topology deforming unit is a quality management device that exhibits a transmission function when only a link starting from the node connected to a quality management device that exhibits both the monitoring data transmission function and the reception function is connected. Is virtually divided into a first node connected to the second node connected to a quality control device that performs a reception function by connecting only a link having the node as an end point. The fault link identification system according to claim 1, wherein the fault link identification system is connected to a node and a pseudo arrival node.   In the first integer programming method, a restriction condition is set for all relay nodes on the network so that a monitoring route passing through the relay node does not pass through the same adjacent link pair. 3. The fault link identification system according to 1 or 2.   In the second integer programming method, the constraint condition is set so that the monitoring paths passing through the two links are not exactly the same for all link pairs on the network. The fault link identification system according to any of the above.   In the first integer programming, the number of monitoring paths passing through a node where the number of incoming links is m and the number of outgoing links is n is greater than or equal to Max (m, n, 3/2 (m + n)). The failure link identification system according to any one of claims 1 to 4, further comprising: A plurality of quality control devices each having at least one of a monitoring data transmission function and a reception function are connected to nodes on the network, and monitoring data is transmitted from the transmission function to the reception function through a plurality of monitoring paths passing through the relay node. In the monitoring route determination method of the fault link identification system that identifies the fault link based on its reachability,
To get the actual topology of the network,
Transforming the real topology into a virtual topology;
A step of finding the number of monitoring paths necessary for identifying a single link failure of the network based on the virtual topology by a first integer programming method;
Using the number of necessary monitoring paths as an initial value of the number of monitoring paths, a feasible solution can be obtained by the second integer programming method for a monitoring path group that can identify all single link failures and minimize the total path length. And increasing the number of monitoring routes until the solution is obtained,
In the procedure of transforming the topology, a pseudo source node connected by a virtual link to a plurality of nodes connected to the quality management device that performs the transmission function, and a plurality of connected quality management devices that perform the reception function A pseudo-arrival node connected with a node of the virtual link is virtually arranged,
In the first solving means, the number of monitoring routes from the pseudo origin node to the pseudo destination node is obtained,
The monitoring path determination method for a faulty link identification system, wherein the second link finding unit sets a link length of the virtual link to zero.

Images