Network failure location identification device, network failure location identification method, and program

Abstract

To identify a fault point as much accurately as possible by only a limited number of path measurements under probability routing.SOLUTION: Provided is a network fault point identifying device for identifying a fault point in a target network, said device comprising: a test optimization unit for calculating an optimum path measurement test to be executed on the target network; a test execution unit for executing the test calculated by the test optimization unit upon the target network; and a test result analysis unit for narrowing down a network state in accordance with the result of test by the test execution unit. The test optimization unit calculates a test using the analysis result obtained by the test result analysis unit, the test execution unit executes the test, and the test result analysis unit repeats the process of narrowing down a network state in accordance with the test result one time or more.SELECTED DRAWING: Figure 5

Description

Technical Field

[0001] The present invention relates to a technique for identifying a failure location in a network.

Background Art

[0002] In recent years, networks have become more complex, and failures have occurred that cannot be detected by conventional methods that rely solely on data such as logs and metrics output by network devices.

[0003] As a means for detecting such failures, a means called "network tomography" has attracted attention. Network tomography measures the end-to-end communication status between a plurality of remote nodes (this is called path measurement), and integrates the records regarding the connectivity to identify the failure location (failed node or failed link).

[0004] In particular, those that estimate the binary state (with / without failure) of each network component (node, link) are also called binary network tomography and are actively studied [Non-Patent Document 1].

[0005] Many of the existing techniques of binary network tomography assume that routing is deterministic. However, in reality, there are protocols such as a load balancing mechanism and ECMP that distributes traffic to paths with equal weights, and situations occur where routing behaves probabilistically. Regarding network tomography under probabilistic routing, although there are few, there are existing methods [Non-Patent Documents 2, 3].

Prior Art Documents

Non-Patent Documents

[0006]

Non-Patent Document 1

[0007] In probabilistic routing, even when a node pair is determined, the path connecting them is not uniquely determined. Therefore, a large number of path measurements are required to correctly identify the failure location. Generally, a large number of path measurements should be avoided as much as possible because they prolong the time to identify the failure location or impose a large load on the network.

[0008] However, existing network tomography techniques [Non-Patent Documents 2 and 3] under probabilistic routing assume that a large amount of path measurement data has already been obtained, and there is a possibility that an unnecessarily large number of path measurements will be carried out.

[0009] The present invention has been made in view of the above points, and an object thereof is to provide a failure location identification technique for accurately identifying a failure location with a limited number of path measurements under probabilistic routing. [Means for Solving the Problems]

[0010] According to the disclosed technology, a network fault location identification device for identifying a fault location in a target network, a test optimization unit that calculates a path measurement test to be executed on the target network, a test execution unit that executes the test calculated by the test optimization unit on the target network, and a test result analysis unit that narrows down the state of the target network according to the result of the test by the test execution unit. The network fault location identification device is provided with: the test optimization unit uses the analysis result obtained by the test result analysis unit Furthermore to calculate a test, the test execution unit executes the Further calculated test, and the test result analysis unit repeats the process of narrowing down the state of the target network according to the result of the Further calculated test one or more times. The test optimization unit calculates a test that is an optimal solution or an approximate optimal solution to the problem of maximizing the mutual information amount between the probability variable representing the execution result of the test and the probability variable representing the state of the target network. A network fault location identification device is provided. The test optimization unit calculates a test that is an optimal solution or an approximate optimal solution to the problem of maximizing the mutual information amount between the probability variable representing the execution result of the test and the probability variable representing the state of the target network. Additional A network fault location identification device that calculates a test is provided.

Advantages of the Invention

[0011] According to the disclosed technology, it becomes possible to identify the fault location as accurately as possible with a limited number of path measurements under probabilistic routing.

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Mode for Carrying Out the Invention

[0013] Hereinafter, embodiments (these embodiments) of the present invention will be described with reference to the drawings. The embodiments described below are merely examples, and the embodiments to which the present invention is applied are not limited to the following embodiments.

[0014] (Overall System Configuration, Operation Outline) Fig. 1 shows an overall configuration example of the system in this embodiment. As shown in Fig. 1, this system has a configuration in which a network failure location identification device 100 is connected to a target network 200, which is a network to be identified for network failure locations. The network failure location identification device 100 identifies the failure location by repeating one or more times the process of having the path measurement device perform a test related to path measurement and obtaining the result based on the routing topology information of the target network 200, the routing information that can change dynamically, the constraint information of the path measurement device, and the like.

[0015] The target network 200 has a plurality of nodes and a plurality of links, and data transmission and reception are performed by probabilistic routing. Note that a link may also be referred to as a branch or an edge.

[0016] In this embodiment, the network failure location identification device 100 identifies the failure location as accurately as possible with a limited number of path measurements under probabilistic routing. The general operation of the network failure location identification device 100 for this purpose is as follows.

[0017] In this embodiment, when the network failure location identification device 100 narrows down the failure location, the effectiveness of path measurement is expressed using "mutual information". Mutual information can be naturally defined even under probabilistic routing. Path measurements with a large mutual information, that is, those that are expected to narrow down the failure location the most, are preferentially performed to obtain measurement data.

[0018] The network failure location identification device 100 updates the probability distribution representing the possibility of the failure location in the framework of Bayesian estimation according to this measurement data. By repeating these procedures, the failure location is gradually narrowed down.

[0019] In this embodiment, the network failure location identification device 100 solves the above problems by using highly effective path measurements and sequentially determining the measurements to be performed according to Bayesian estimation, thereby efficiently narrowing down the failure location with a small number of measurements.

[0020] Hereinafter, the processing content executed by the network failure location identification device 100 will be described in detail.

[0021] (Problem setting) First, the formulation of the problem in this embodiment will be described. However, the technology according to this embodiment can also be applied in situations that do not strictly follow the following definitions. For example, although the following formulates the identification of link failures, the technology according to this embodiment can also be similarly applied to the identification of node failures.

[0022] Let the target network be an undirected graph G(V, E). V = {v i} i is the vertex set, and E = {e j}j is a set of branches. The "network state" or simply "state" is represented by a binary vector s = (s1, ···, s |E| ) ∈ S ⊂ {0, 1} |E| . Here, s j = 1 (s j = 0) indicates that branch e j is abnormal (normal), and S represents the set of all possible states.

[0023] The set of monitoring paths A = {a1, ···, a|A|} is a set of "monitoring paths" represented by a binary vector a j = (a j1 , ···, a j|E| ) ∈ {0, 1} |E| . If a jl = 1, it means that the monitoring path a j includes branch e l . Since a monitoring path can be regarded simply as a set of branches, paths with loops are also allowed.

[0024] When the test of the monitoring path a j is executed, if any branch in a j is abnormal, the result is 1; if all branches are normal, the result is 0. That is, the result 1 indicates that the monitoring path is blocked, and the result 0 indicates that the monitoring path can pass through. In this embodiment, since probabilistic routing is assumed, a j cannot be directly specified, and only a "group" of multiple monitoring paths with the same source vertex (S node) and destination vertex (D node) can be specified.

[0025] When a group c i ∈ C is specified, it is assumed that the tests of the monitoring paths a j are executed independently with probability p ij (i = 1, ···, |C|, j = 1, ···, |A|). Here

[0026]

Number

[0027]

Number

[0028]

Number

[0029] Considering such a situation, a "probe test" or simply a "test" ξ i ∈X={ξ1,···,ξ |X|} is defined. In this embodiment,

[0030]

Number

[0031] The problem in this embodiment is as follows.

[0032] Now, assume that the state s ∈ S is unknown. Also assume that the prior distribution of s and the number N of feasible tests ξ are given. At this time, in order to identify the true state with as high a probability as possible by performing N tests, what kind of strategy should be used to perform the tests (how many times and at what timing should each test be performed) to be efficient? Also, how should the true state be estimated according to the test results? Note that in the following explanation, it is assumed that the state s does not change dynamically, but the technology according to this embodiment is applicable even if it changes midway.

[0033] (Details of the processing content by the network failure location identification device 100) In this embodiment, an "adaptive measurement" approach is taken. This means that the network failure location identification device 100 sequentially determines the next test to be performed according to the test results obtained at present. Adaptive measurement is more complex to implement than a non - adaptive approach that determines all N tests at once, but since the information available for determining the optimal test increases step by step, it is expected that high - precision state identification can be achieved with a small number of test times as a result.

[0034] Specifically, it is formulated as the following batch processing.

[0035] Divide the N tests into B batches of size N B That is, N B ×B = N holds. In the b - th batch ((b ∈ [1, B] ∩ Z + ), the test design

[0036]

Number

[0037]

Number

[0038]

Number

[0039]

Number

[0040] When designing M (the subscript b of M b is omitted as appropriate), the "goodness" of M must be quantitatively represented. Since the purpose is to identify the true state, it is a natural strategy to sharpen the probability distribution of the state as much as possible, that is, to reduce the entropy.

[0041] Therefore, in this embodiment, as an index of the effectiveness brought by y M obtained by the execution of M, the mutual information amount I(S; Y M between S and Y M ) will be used (when regarding y M and s as random variables, capital letters such as Y M , S are used). I(S; Y M ) is the entropy of the state distribution when Y MIt represents how much it decreases on average by observing this. Thus, the problem considered in this embodiment can be described as follows.

[0042] [Problem]: When the prior distribution Pr(s) and the measurement results obtained up to the (b - 1)-th batch

[0043]

Number

[0044]

Number

[0045] Here, I(S;Y M |D b-1 ) is the mutual information amount between Y b-1 and S when D M is given, and is given as in the following Equation 1.

[0046]

Number

[0047]

Number

[0048]

Number

[0049] However, the technology according to this embodiment is not limited to the greedy method, and other approximation optimization methods and optimization methods can also be used. Also, although the case where the routing probability does not change dynamically is described here, if information can be obtained even when it changes, the technology according to this embodiment can be applied.

[0050] The procedure (algorithm) of the process executed by the network failure location identification device 100 is shown in Algorithm 1 of FIG. 2. First, there is a StateSpaceReduction process for reducing the possible states that can be on the second line, which will be described later.

[0051] In the batch process starting from the fifth line, M is created based on the greedy method from lines 6 to 10. Inside the while loop, ξ is sequentially selected such that the increment of I(S;Y M |D b-1 ) is the largest, and the i-th i_max component of M is incremented. After creating M, it is executed to obtain y max , and the posterior distribution Pr(s|D M ) is updated in lines 11 to 13. b )

[0052] Subsequently, the calculation method of the mutual information amount in line 8 will be described. The mutual information amount is based on Procedure 2 shown in FIG. 3. Basically, it is based on the formula of I(S;Y M |D b-1 ) in Equation 1, but since the sum with respect to y M in Equation 3 is a sum of an exponential number of terms, exact execution is difficult. Therefore, Monte Carlo sampling is performed as follows.

[0053] First, N M of y yObtain the individual samplings according to the binomial distribution in the 4th row. For each sample, in the 6th to 7th rows, calculate the posterior distribution by the method described later and calculate its entropy. Then take the average of them in the 8th row. In Procedure2, it is assumed that all possible states s can be enumerated, but when |S| is large, the sum over s may be replaced by the sample average.

[0054] Next, regarding the calculation of the posterior distributions Pr(s|D b )(Pr(s|D b-1 ) and so on) that appear in the 13th row of Algorithm1 and the 2nd, 6th, and 8th rows of Procedure2, an explanation will be given. Pr(s|D b ) is

[0055]

Number

[0056]

Number

[0057]

Number

[0058]

Number

[0059] Now, although Algorithm1 is an approximation algorithm based on the greedy method, it can be shown that it has a constant approximation degree in terms of maximizing the mutual information as follows. That is, even in the worst case, it is guaranteed that the mutual information is above a certain value.

[0060] [Theorem]: N y When N is sufficiently large, let M obtained by Algorithm1 be b denoted as ~ M b and let the optimal M be b denoted as

[0061] [Number] Then,

[0062] [Number] holds.

[0063] (Sketch of proof) Using the fact that the routing probabilities are independent of each other, it can be shown that the mutual information I(S; Y M | D b-1 ) has the property of a monotone DR-submodular function [Non-Patent Document 4] as a function of

[0064] [Number] . In general, it is known that the maximization problem of a monotone DR-submodular function can achieve a (1 - 1 / e)-approximation by the greedy method [Non-Patent Document 5]. (Proof ends) Finally, the process of StateSpaceReduction in the second line of Algorithm1 will be explained. This process is to remove impossible states from S in advance to shorten the execution time of Algorithm1. Generally, the matrix U contains many components with values of 0 or 1.

[0065] In fact, since one group c usually contains only a very small number of branches in the entire network, it often has many 0 components. Also, since the monitoring paths within the same group often pass through some common branches, there are many 1 components. Additionally, the following obvious proposition also holds: "Let u i (s)=1(0). At this time, if group c i is executed and the result is 1(0), then the true state is not s." Based on the above, the following definition is introduced.

[0066] [Definition] (Removable state): When group c i is executed and the result 1(0) is obtained, a state s that satisfies u i (s)=0(1) is removable.

[0067] Obviously, removable states can be excluded from the state space S. Also, the following lemma holds.

[0068] [Lemma]: When the test ξ is executed, the expected value R(U, ξ) of the number of removable states is given by Equation 5 below.

[0069] [Equation] Here, δ d (x)=1(x = d), δ d (x)=0 (otherwise), where (d = 0, 1), and 0 0 = 1. In particular, when ξ = e l (a unit vector with only the l-th component being 1), it becomes Equation 6 below.

[0070] [Equation] Here,

[0071] [Equation] is as follows.

[0072] (Sketch of proof) From the linearity of the expected value

[0073] [Number] it becomes. Considering that the execution of each group is a Bernoulli trial,

[0074] [Number] it can be described as, so take the complementary event of this and sum over i. Equation 6 is ξ = e l Substitute and calculate, and the inequality is

[0075] [Number] is obtained from. (Proof completed) Based on the above lemma, Procedure 3 in Figure 4 is the summary of the StateSpaceReduction process. Calculate Equation 6 for each group c l and obtain c l_max that maximizes it. Next, arbitrarily select and execute a test ξ that includes c l_max . Exclude the removable states from S according to the execution result, and also reduce the matrix U. Repeat the above steps N iter times.

[0076] Note that, instead of Equation 6, ξ that maximizes Equation 5 can be arbitrarily selected and executed. Since Equations 5 and 6 are lighter than the calculation of mutual information, the total calculation time of Algorithm 1 is consequently reduced.

[0077] (Example) As an example of the above processing, a configuration example of the network failure location identification device 100 and an example of a processing procedure using the configuration will be described.

[0078] Fig. 5 shows a configuration example of the network failure location identification device 100. As shown in Fig. 5, the network failure location identification device 100 includes an input UI 110, a state number reduction unit 120, a test execution unit 130, a mutual information amount maximization unit 140, a posterior distribution calculation unit 150, and an output UI 160. The state number reduction unit 120, the test execution unit 130, and the posterior distribution calculation unit 150 are connected to the target network 200 as shown. Note that "state number reduction unit 120 + mutual information amount maximization unit 140" may be called a test optimization unit. Also, the posterior distribution calculation unit 150 may be called a test result analysis unit.

[0079] The processing procedure of the network failure location identification device 100 having the above configuration will be described with reference to the flowchart of Fig. 6.

[0080] In S101, first, data and parameters (prior distribution Pr(s), U, N B , B, etc.) necessary for algorithm execution are input to the input UI 110. In S102, based on these, the state number reduction unit 120 determines ξ according to Procedure 3 and passes it to the test execution unit 130.

[0081] In S103, the test execution unit 130 executes a test on the target network 200 using a connectivity confirmation program such as ping. Again in S102, the state number reduction unit 120 determines the next ξ according to Procedure 3 based on the obtained result.

[0082] After performing S102 to S103 a determined number of times, next, in S104, the mutual information amount maximization unit 140 creates M according to the greedy method of Algorithm 1. This is passed to the test execution unit 130, and in S105, a test is executed on the target network 200.

[0083] The result is passed to the posterior distribution calculation unit 150, and in S106, the posterior distribution is calculated according to the framework of Bayesian estimation. The obtained posterior distribution is passed to the mutual information maximization unit 140, and the above steps (S104 to S106) are repeated according to the loop of Algorithm1. After performing the determined number of times, the estimated state is output as the maximum likelihood value from the final posterior distribution to the output UI 160. In S107, the output UI outputs the estimated state.

[0084] (Regarding the effect) According to the technology related to the present embodiment described above, it is possible to identify the failure location as accurately as possible with a limited number of path measurements under probabilistic routing. In this technology, since path measurements effective for failure identification are preferentially performed, the number of path measurements required until failure identification is suppressed to a small number, and it is possible to expect shortening of failure identification and reduction of network load.

[0085] Evaluation was performed using the three network data [Non-Patent Document 6] in FIG. 7. #failures in FIG. 7 represents the number of branches that fail simultaneously, and |S k | is the total number of possible states. Consider the following settings for each network.

[0086] The number of groups is |C| = 3|V|. For each node, when it is set as the S node, three randomly selected nodes are set as the D nodes, and the node pairs (groups) are determined. For each node pair, the shortest path, the second shortest path, and the third shortest path are regarded as monitoring paths (|A| = 9|V|). For the i-th shortest path for each group, the probability

[0087] [Number] is selected as such (l i is the length of the path). One probe test ξ is regarded as executing three groups with the same S node once, and a total of |X| = |V| patterns of tests are considered. The initial state distribution Pr(s) is uniform, and N ySet it to 30. Also, the state reduction by Procedure3 was carried out with N iter = 10.

[0088] In addition to the technology (denoted as PM) according to this embodiment, as a comparison, Random, LS [Non-Patent Document 2], and LASSO [Non-Patent Document 3] were implemented. Random is a method of randomly selecting ξ without using the mutual information amount in the method of this embodiment.

[0089] LS and LASSO are existing methods based on non-adaptive approaches proposed as network tomography in probabilistic routing. Regarding LASSO, it was implemented when the hyperparameter λ was changed to 0.0001, 0.001, and 0.01.

[0090] The correct answer rate was used as the evaluation index. For a state with two failure branches, it was regarded as correct only when both were specified. Regarding Missouri, experiments were conducted for all true state patterns, and for ION and Ntelos, experiments were conducted for 50 true state patterns (randomly selected), and the correct answer rate was calculated.

[0091] The results of the experiments are shown in FIGS. 8 to 10. In each figure, the horizontal axis is the total number of tests N, and the vertical axis is the correct answer rate. It can be seen that the method PM according to this embodiment shows higher performance than the existing methods LS and LASSO. For example, to exceed a correct answer rate of 0.96, the best existing methods require 160 (Missouri), 320 (ION), and 640 (Ntelos) test numbers, while PM only requires 22, 34, and 34 test numbers respectively. Also, PM exceeds the results of Random, indicating the usefulness of using the mutual information amount.

[0092] (Hardware configuration example) The network failure location identification device 100 can be realized, for example, by causing a computer to execute a program. This computer may be a physical computer or a virtual machine on the cloud.

[0093] That is, the network failure location identification device 100 can be realized by using hardware resources such as a CPU and a memory built in a computer to execute a program corresponding to the processing performed by the network failure location identification device 100. The above program can be recorded on a computer-readable recording medium (such as a portable memory), saved, or distributed. Further, it is also possible to provide the above program through a network such as the Internet or e-mail.

[0094] FIG. 11 is a diagram showing an example of the hardware configuration of the computer. The computers in FIG. 11 include a drive device 1000, an auxiliary storage device 1002, a memory device 1003, a CPU 1004, an interface device 1005, a display device 1006, an input device 1007, an output device 1008, etc., which are mutually connected by a bus BS.

[0095] A program for realizing the processing on the computer is provided by a recording medium 1001 such as a CD-ROM or a memory card, for example. When the recording medium 1001 storing the program is set in the drive device 1000, the program is installed from the recording medium 1001 via the drive device 1000 into the auxiliary storage device 1002. However, the installation of the program does not necessarily have to be performed from the recording medium 1001, and it may be downloaded from another computer via a network. The auxiliary storage device 1002 stores the installed program and also stores necessary files, data, etc.

[0096] When there is an instruction to start a program, the memory device 1003 reads and stores the program from the auxiliary storage device 1002. The CPU 1004 realizes the functions related to the light touch maintenance device 100 according to the program stored in the memory device 1003. The interface device 1005 is used as an interface for connecting to a network or the like. The display device 1006 displays a GUI (Graphical User Interface) or the like according to the program. The input device 1007 is composed of a keyboard, a mouse, buttons, a touch panel, or the like, and is used to input various operation instructions. The output device 1008 outputs the calculation result.

[0097] (Appendix) This specification discloses at least a network failure location specifying device, a network failure location specifying method, and a program according to the following respective items. (Item 1) A network failure location specifying device for specifying a failure location of a target network, a test optimization unit that calculates an optimal path measurement test to be executed on the target network; a test execution unit that executes the test calculated by the test optimization unit on the target network; and a test result analysis unit that narrows down the state of the network according to the result of the test by the test execution unit, wherein the test optimization unit calculates a test using the analysis result obtained by the test result analysis unit, the test execution unit executes the test, and the test result analysis unit repeats the process of narrowing down the state of the network according to the result of the test one or more times Network failure location specifying device. (Item 2) In the routing information in the target network, the paths connecting the source and destination nodes are not uniquely determined, but are determined probabilistically, and it is impossible to observe which path is selected in the test, and only the probability distribution thereof is available The network failure location specifying device according to Item 1. (Item 3) The test optimization unit selects a test that is an optimal solution or an approximate optimal solution to a problem of maximizing the mutual information amount between a random variable representing the execution result of a test and a random variable representing the state of the target network. The network fault location identification device according to claim 1 or 2. (Item 4) Based on the execution result of the test, the test optimization unit selects a test such that the expected value of the number of network states that can be excluded as candidates increases. The network fault location identification device according to any one of claims 1 to 3. (Item 5) The test result analysis unit updates the probability distribution representing the network state according to the framework of Bayesian estimation based on the execution result of the test. The network fault location identification device according to any one of claims 1 to 3. (Item 6) A network fault location identification method executed by a network fault location identification device for identifying a fault location of a target network, a test optimization step of calculating an optimal path measurement test to be executed on the target network; a test execution step of executing the test calculated in the test optimization step on the target network; a test result analysis step of narrowing down the state of the network according to the result of the test in the test execution step, and using the analysis result obtained in the test result analysis step, calculating a test in the test optimization step, executing the test in the test execution step, and narrowing down the state of the network according to the result of the test in the test result analysis step, and repeating the process one or more times. Network fault location identification method. (Item 7) A program for causing a computer to function as each unit in the network fault location identification device according to any one of claims 1 to 5.

[0098] As described above, the present embodiment has been explained. However, the present invention is not limited to such a specific embodiment, and various modifications and changes are possible within the scope of the gist of the present invention described in the claims.

Explanation of Reference Numerals

[0099] 100 Network fault location identification device 110 Input UI 120 State reduction unit 130 Test execution unit 140 Mutual information maximization unit 150 Posterior distribution calculation unit 160 Output UI 1000 Drive device 1001 Recording medium 1002 Auxiliary storage device 1003 Memory device 1004 CPU 1005 Interface device 1006 Display device 1007 Input device 1008 Output device

Claims

1. A network fault location identifying device for identifying a fault location in a target network, comprising: a test optimization unit that calculates a test for path measurement to be executed on the target network; a test execution unit that executes the test calculated by the test optimization unit on the target network; a test result analysis unit that narrows down the state of the target network according to the result of the test by the test execution unit, wherein the test optimization unit further calculates a test using the analysis result obtained by the test result analysis unit, the test execution unit executes the further calculated test, and the test result analysis unit repeats the process of narrowing down the state of the target network according to the result of the further calculated test one or more times. The network fault location identifying device is characterized in that the test optimization unit calculates a further test that is an optimal solution or an approximate optimal solution to a problem of maximizing the mutual information amount between a probability variable representing the execution result of the test and a probability variable representing the state of the target network. Network fault location identifying device.

2. A network fault location identifying device for identifying a fault location in a target network, comprising: a test optimization unit that calculates a test for path measurement to be executed on the target network; a test execution unit that executes the test calculated by the test optimization unit on the target network; a test result analysis unit that narrows down the state of the target network according to the result of the test by the test execution unit, wherein the test optimization unit further calculates a test using the analysis result obtained by the test result analysis unit, the test execution unit executes the further calculated test, and the test result analysis unit repeats the process of narrowing down the state of the target network according to the result of the further calculated test one or more times. The network fault location identifying device is characterized in that the test optimization unit calculates a further test such that the expected value of the number of states of the target network that can be excluded as candidates increases based on the execution result of the test. Network fault location identifying device.

3. A network fault location identifying device for identifying a fault location in a target network, comprising: a test optimization unit that calculates a test for path measurement to be executed on the target network; A test execution unit that executes the test calculated by the test optimization unit on the target network, A test result analysis unit that narrows down the state of the target network according to the result of the test by the test execution unit, and The test optimization unit further calculates a test using the analysis result obtained by the test result analysis unit, the test execution unit executes the further calculated test, and the test result analysis unit narrows down the state of the target network according to the result of the further calculated test. A network fault location identification device that repeats the process one or more times, The test result analysis unit updates the probability distribution representing the state of the target network according to the execution result of the test according to the framework of Bayesian estimation, Network fault location identification device.

4. A network fault location identification method executed by a network fault location identification device for identifying a fault location of a target network, A test optimization step of calculating a test for path measurement to be executed on the target network, A test execution step of executing the test calculated by the test optimization step on the target network, A test result analysis step of narrowing down the state of the target network according to the result of the test by the test execution step, and A network fault location identification method that further calculates a test by the test optimization step using the analysis result obtained by the test result analysis step, executes the further calculated test by the test execution step, and according to the result of the further calculated test by the test result analysis step. The process of narrowing down the state of the target network is repeated one or more times, In the test optimization step, a further test that is an optimal solution or an approximate optimal solution to the problem of maximizing the mutual information amount between the probability variable representing the execution result of the test and the probability variable representing the state of the target network is calculated, Network fault location identification method.

5. A network fault location identification method executed by a network fault location identification device for identifying a fault location of a target network, A test optimization step of calculating a test for path measurement to be executed on the target network, A test execution step of executing the test calculated by the test optimization step on the target network; A test result analysis step of narrowing down the state of the target network according to the result of the test in the test execution step; and A network fault location identification method that repeats, one or more times, a process of calculating a further test by the test optimization step using the analysis result obtained by the test result analysis step, executing the further calculated test by the test execution step, and narrowing down the state of the target network according to the result of the further calculated test by the test result analysis step; In the test optimization step, calculate a further test such that the expected value of the number of states of the target network that can be excluded as candidates increases based on the execution result of the test Network fault location identification method.

6. A network fault location identification method executed by a network fault location identification device for identifying a fault location of a target network, comprising: A test optimization step of calculating a test for path measurement to be executed on the target network; A test execution step of executing the test calculated by the test optimization step on the target network; A test result analysis step of narrowing down the state of the target network according to the result of the test in the test execution step; and A network fault location identification method that repeats, one or more times, a process of calculating a further test by the test optimization step using the analysis result obtained by the test result analysis step, executing the further calculated test by the test execution step, and narrowing down the state of the target network according to the result of the further calculated test by the test result analysis step; In the test result analysis step, update the probability distribution representing the state of the target network according to the execution result of the test according to the framework of Bayesian estimation Network fault location identification method.

7. A program for causing a computer to function as each part in the network fault location identification device according to any one of Claims 1 to 3.

Images