CN106100882A - A kind of construction method of the network fault diagnosis model based on flow value - Google Patents
A kind of construction method of the network fault diagnosis model based on flow value Download PDFInfo
- Publication number
- CN106100882A CN106100882A CN201610416606.2A CN201610416606A CN106100882A CN 106100882 A CN106100882 A CN 106100882A CN 201610416606 A CN201610416606 A CN 201610416606A CN 106100882 A CN106100882 A CN 106100882A
- Authority
- CN
- China
- Prior art keywords
- node
- path
- network
- flow value
- fault diagnosis
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/06—Management of faults, events, alarms or notifications
- H04L41/0677—Localisation of faults
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses the construction method of a kind of network fault diagnosis model based on flow value, the flow transmission mode of the construction method analog network of the described network fault diagnosis based on flow value, divide node state is thinner, use corresponding acquisition station choosing method and congested node, malfunctioning node location algorithm, though the congestion point positioning on the premise of choosing less website in network and trouble point.The flow transmission mode of the network fault diagnosis modeling network of the present invention, divide node state is thinner, more realistic；Newly constructed network model proposes corresponding acquisition station Algorithms of Selecting and congested node, malfunctioning node location algorithm；Improve the location efficiency in monitoring path by the present invention, even if experiment simulation proves that the present invention still can accurately and efficiently position the congestion point in network and trouble point on the premise of choosing less website.
Description
Technical field
The invention belongs to network failure management technical field, particularly relate to a kind of network fault diagnosis mould based on flow value
The construction method of type.
Background technology
Fault diagnosis, a core direction of network failure management, is to infer to lose from the failed sign that some observe
One process of the precise reason why losing.Fault diagnosis is generally divided into two steps: malfunction monitoring and fault location.With computer
Network is at ecommerce, electronic communication, and the extensive application of the aspects such as web service, fault diagnosis has had become as network manager
A vital task, substantial amounts of research focus is gathered in this, creates a large amount of fault diagnosis technology.But, owing to system becomes
Must become increasingly complex and provide many New function, the requirement to FLT is also constantly changing therewith.Although
Through there being these effort studied, the FLT of network remains the very promising of an opening and studies a question.
Detection method is one processing method end to end, is sent detection information bag by positive and reached to infer in network
The purpose of networking component health degree.Detection information is obtaining end-to-end information, and for example, delay, handling capacity, loss, path can
It is widely used by aspects such as situations.Existing fault diagnosis technology is mainly focused on discovery the fault positioning in network
Node or faulty link, the state of node or link is typically divided into two classes by these methods: healthy or fault.For example,
Maitreya Natu, Adarshpal S.Sethi etc., send detectable signal discovery by detective path selected in network
Disconnected link thus the purpose that reaches fault point；P é ter Babarczi, link is divided by Pin-Han Ho etc. equally
It is that 0 and 1 two states represents healthy and fault respectively, by disposing monitoring path (monitoring trail), obtain all
Possible chain circuit alarming code table, comparison alarm code can quickly position the single link failure of network, and in order to adapt to different nets
Network structure；And by building the method for Bayesian network, infer the faulty link group of IP network with probabilistic manner.
But the transmission quantity of information and data is often huge in network, the health status of each node, process and pass
The ability of transmission information is uneven.Some node is due to a variety of causes such as hardware aging or fault etc., when need information to be processed
Beyond its disposal ability, information will congestion, congested node can drag the transmission speed of slow network undoubtedly.And the appearance of congested node
Often influence whether again neighbor node about, present the pattern that one " infects ".Therefore only by node or chain k-path partition
It is inadequate for healthy and fault two states.How quickly to position and process these congested node be one be highly desirable to solve
Problem certainly.Fault diagnosis is detected by end-to-end transmission detectable signal, is by actively sending in a network
The health status of network components is inferred in detection.Detection is test assignment, and the success of test depends on the network components of detection
Health status.Detection is used for obtaining adding up end to end, such as delay, handling capacity, Loss Rate and route availability.Citing comes
Saying, detection can be regarded as a ping and collects with regard to Connectivity information.More complicated detection be a series of size not
The information with regard to total bandwidth or available bandwidth collected by same detection bag.For special server, detect also with regard to fault diagnosis
Can be that HTTP request is to collect the statistical information of response time.Obtain expected end except generating flow between Chosen Point
To the statistics of end, detection also provides other advantages.Detection is in terms of meeting special measurement, and detection flow design aspect is very clever
Live.For example, position ad hoc network point failure by average retardation amount on test path and Loss Rate size, can send preferably
Bottleneck and available bandwidth are identified in detection, or go to estimate background traffic.Other application of detection are by Detection location bottleneck
Node and send special detection further the reason that go further tracing trouble to occur.
Existing methodical acquisition station quantity is more, and fault diagnosis rate is relatively low.
Content of the invention
It is an object of the invention to provide the construction method of a kind of network fault diagnosis model based on flow value, it is intended to solve
Certainly existing methodical acquisition station quantity is more, the relatively low problem of fault diagnosis rate.
The present invention is achieved in that the construction method of a kind of network fault diagnosis model based on flow value, described base
In the flow transmission mode of construction method analog network of the network fault diagnosis of flow value, divide node state is thinner,
Use corresponding acquisition station choosing method and congested node, malfunctioning node location algorithm, even if before choosing less website
Put the congestion point in positioning network and trouble point.
Further, the described construction method based on the network fault diagnosis model of flow value comprises the following steps:
Step one, builds the network topological diagram of connection；
Step 2, initializes, and gives the flow value between each node Random assignment 0 to 0.5；
Step 3, randomly selects certain node, and the flow value of node is increased certain value, and flow value is more than appointment threshold value
The node of 0.8 is as preliminary congested node, and the node that flow value increases to 1 is considered as trouble point；
Step 4, the infection characteristic of simulation congested node, congested node peripherad neighbor node transport stream value, and will
Flow value is more than the node specifying threshold value as newly-increased congested node；
Step 5, newly-increased congested node continues to affect the neighbor node of oneself, transmits extra flow value, repeat this mistake
Journey.
Further, the choosing method of described acquisition station includes:
Input: network G=(V, E), maximum probe station number MAXpssize, fault is counted MAXfaults；
Output: acquisition station set of node PS；
Number of network node N=| V |, shaded nodes collection SN=V,
Select the node i of number of degrees maximum as first monitoring station PS=PS ∪ i, update shaded nodes and monitoring path；
Further, the monitoring path generating method that can use includes:
Input: network G=(V, E), acquisition station point set PS；
Output: available monitoring path Ap；
Further, the described detection method to remaining measuring point to be checked is as follows:
Input: available path Ap, point set Nn to be detected；
Output: the monitoring path Sp of available detection；
The unknown point set of the point set Un that cannot detect, Un；
Step 0.While
Step 1. is to each vi∈ Nn, finds out by the minimum v of path degree of covering in Nnmin, cover if there is no path
Lid viThen increase viIn Un, Nn=Nn-vmin, turn step 0；
Step 2. is at every pathj∈ Ap and covering vminIn, find out path path the longestmax, pathmaxIncrease to
In Sp；
Step 3.Ap=Ap-pathmax, Nn=Nn-vmin.
Further, described congestion point and fault point positioning method include:
Input: available path Ap, point set Nn to be detected；
Output: for the monitoring path Sp of detection；
The point set Un (unknown point set) that cannot detect；
Step 0.While
Step 1. is to each vi∈ Nn, finds out by the minimum v of path degree of covering in Nnmin, cover if there is no path
Lid viThen increase viIn Un, Nn=Nn-vmin, turn step 0；
Step 2. is at every pathj∈ Ap and covering vminIn, find out path path the longestmax, increase pathmaxTo Sp
In；
Step 3.AP=AP-pathmax, Nn=Nn-vmin.
The construction method of the network fault diagnosis model based on flow value that the present invention provides, the flow transmission of analog network
Pattern, divide node state is thinner, more realistic；Newly constructed network model proposes corresponding acquisition station and chooses calculation
Method and congested node, malfunctioning node location algorithm；Prove that the present invention has higher fault diagnosis precision by experiment simulation, shortly
The congestion point in network and trouble point still can be accurately and efficiently positioned on the premise of making to choose less website；Fault is examined
Disconnected is the process of detection failure, by positioning the cause of fault from the failure symptom observing；Owing to computer network is at electricity
Son commercial affairs, telecommuting, network service, being widely used of mesh services etc., fault diagnosis for network manager
Through becoming a vital task.Additionally, development key application performance improvement, service quality importance increase,
The development of large-scale complicated system makes quick detection and separation the visiting for robustness, reliability and system of fault
Asking property it is critical that.
Brief description
Fig. 1 is the construction method flow chart of the network fault diagnosis model based on flow value that the embodiment of the present invention provides.
Fig. 2 is the topological diagram diagram of the network fault diagnosis model based on flow value that the embodiment of the present invention provides.
Fig. 3 is that the different site location selection that the embodiment of the present invention provides affects schematic diagram to monitoring result.
Fig. 4 is the monitoring path schematic diagram being generated by website 1 that the embodiment of the present invention provides.
Fig. 5 be the embodiment of the present invention provide delete unnecessary monitoring path schematic diagram.
Fig. 6 is the path schematic diagram that the needs that the embodiment of the present invention provides additionally are chosen.
Fig. 7 is the diagnostic process flow chart that the embodiment of the present invention provides.
Fig. 8 is diagnosis and the non-diagnosis analysis schematic diagram of embodiment of the present invention offer；
In figure: diagnosis when (a) trouble point is different；Non-diagnosis when () trouble point is different b；C () average degree is examined when different
Disconnected rate；Non-diagnosis when () average degree is different d；Diagnosis when () network size is different e；F () network size does not diagnoses when different
Rate.
Detailed description of the invention
In order to make the purpose of the present invention, technical scheme and advantage clearer, below in conjunction with embodiment, to the present invention
It is further elaborated.It should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not used to
Limit the present invention.
Below in conjunction with the accompanying drawings the application principle of the present invention is explained in detail.
As it is shown in figure 1, the construction method based on the network fault diagnosis model of flow value of the embodiment of the present invention include with
Lower step:
S101: build the network topological diagram of connection；
S102: initialize, gives the flow value between each node Random assignment 0 to 0.5；
The flow value of node is increased certain value by S103: randomly select certain node, and flow value is more than appointment threshold value
The node of (taking 0.8) is as preliminary congested node, and flow value increases to the node of 1 and is considered as trouble point and (no longer has transmission energy
Power)；
S104: the infection characteristic of simulation congested node, congested node peripherad neighbor node transport stream value, and will stream
Value is more than the node specifying threshold value as newly-increased congested node；
S105: newly-increased congested node continues impact (infection) neighbor node of oneself, transmits extra flow to them
Value, repeats this process.
It is further described below in conjunction with the application principle to the present invention for the specific embodiment.
1st, the present invention proposes a kind of network model based on flow, the data-transmission mode of analog network.First give every
The node that flow value exceeds formulation scope is classified as congested node, " infects to simulate by the corresponding flow value of individual node distribution
Pattern " congestion point can will make its flow value increase to its neighbor node transmission certain flow, and the new congestion point of generation repeats to follow
Ring above procedure.After model construction completes, it is proposed that the Algorithms of Selecting of a kind of acquisition station and the detective path for fault location
Congested node in Algorithms of Selecting positioning network.The algorithm being proved by experiment simulation can be before choosing less acquisition station
Put the congested node positioning in network exactly.
Current invention assumes that 1) fixing route pattern, along fixing path transmission information between any two nodes.Do not consider standby
With path, if there is malfunctioning node between path, do not consider new transmission path.2) state or the event of nodes are only considered
Barrier, Link State and failure problems follow-up study in.3) fault of monitoring station is not considered.In fact, work as certain node
When being selected as monitoring station, can directly it be repaired if this node is malfunctioning node.
Building based on the network model of flow value, by the non-directed graph G (V, E) that net definitions is a connection, V represents network
Set of node, E represents link set, but node or link are simply classified as health or fault two kinds by original research mostly
Typically represent with the 0th, 1, and actual conditions are far from so, for example, have the disposal ability of a lot of node to weaken, hardware old and feeble, flow
Congestion, software fault etc. all may result in the health degree of node and be affected but be unlikely to work.The present invention will build
Based on the network model of flow, the node of these flow congestion being referred to as congested node, the existence of congestion point not only makes network
Service quality and transmittability be affected, and owing to cannot smooth and easy transmit towards the data message of these nodes, they
Neighbor node process task become weight, health degree also can be under some influence.Congestion point is made to have certain " infection " property,
As in figure 2 it is shown, congested node 1 cause surrounding neighbours point the 2nd, the 3rd, the 4th, 5 transmission pressure increase.In order to simulate this situation, this
The healthy degree of node is divided into more detailed situation based on flow value by bright network model.Each node starts have 0 to 1
Between flow value, reflection node health status.Two category nodes need to be found: the 1st, (flow value is more than necessarily for congested node
Threshold value).2nd, malfunctioning node.
Building process is as follows: 1) build the network topological diagram of connection.2) initialize, give each node Random assignment 0 to 0.5
Between flow value.3) randomly selecting certain node, the flow value of node being increased certain value, flow value is more than formulation threshold value
The node of (present invention takes 0.8) is as preliminary congested node, and flow value increases to the node of 1 and is considered as trouble point and (no longer has biography
Movement Capabilities).4) the infection characteristic of simulation congested node, congested node peripherad neighbor node transport stream value, and by flow
Value is more than the node specifying threshold value as newly-increased congested node.5) congested node increasing newly continues impact (infection) neighbour of oneself
Occupy node, transmit extra flow value to them, repeat this process.
According to the network model of design, due to the infection ability of congested node, the final congested node number producing can be big
In the initial congested node number setting, but concrete congested node number is unknown, and the situation in the middle of reality is also often
So.Even if in this case.The corresponding website Selection Strategy that the present invention proposes, and fault location algorithm, still may be used
Efficiently and accurately to position congested node and malfunctioning node present in network.
2nd, acquisition station Selection Strategy
In detection method, in order to obtain the information of the introductory path between node, need to dispose in a network for sending out
Send the acquisition station of detection bag acquisition approach information.According to theorem, when network exists k malfunctioning node, for ensureing network
In the health status of each node all can be detected, each node must is fulfilled for one of following two conditions 1) be detection site
Neighbor node (can be directly monitored).Or 2) there is the monitoring path of the k bar independence to this node.Therefore, fault in network
The arrangement number to detection site for the number of node directly affects.In view of the situation that detection site number is limited, each
The monitoring capability of acquisition station can be affected by website installation position, and appropriate website installation position can improve monitoring path pair
The coverage rate of network, and inappropriate website installation position can cause shaded nodes (cannot monitored station monitor node)
Exist, as it is shown on figure 3, the 1st, 5 for during website can fault point 3, and the 5th, the fault of 3 nodes 2 when being website cause the state of 1 without
Method is known.In order to maximize the coverage rate to network, minimize the cost disposed and safeguard, the distribution of acquisition station is disposed needs war
The slightly consideration of property.Heuristic greedy algorithm SNR, the thought of algorithm is, progressively chooses and shadow spots in network can be made to reduce most
Website.Algorithm assumes the number of maximum trouble point present in known network, but actual time often cannot the localization of faults
Number.Algorithm is intended to make monitoring rate reach absolutely, but sufficient amount of website sometimes need not be disposed also to obtain preferably
Monitoring rate.This is external in the case of do not select enough detection sites, and the website that SNR algorithm often walks chooses mode non-optimal plan
Slightly.Having done certain improvement for the network model building to SNR algorithm, each iteration selects to bring at most newly-increased monitoring path
Website.
False code is as follows:
First initialize, by the non-directed graph G=(V, E) that net definitions is a connection, all node original states
Being shadow spots SN=V, acquisition station point set is empty setFirst assume that network exists small number of malfunctioning node, this
Sample can cause selecting less acquisition station, but experiment simulation proves that reality but still can obtain higher fault diagnosis rate.
Select the node of number of degrees maximum as first detection site, because do so can provide most monitoring paths instinctively.
Algorithm calculates successively and selects each residue node as the new monitoring number of passes that can provide during monitoring station subsequently, and selects
The node at most new monitoring path can be provided as new website, and each iteration of algorithm selects a new website until network
In do not have shaded nodes or website number to reach default maximum to may select website number.
3rd, can use monitoring coordinates measurement
Website selects to complete, and can be sent corresponding detection information by website other nodes in network, is formed available
Monitoring path (according to shortest path first), such as Fig. 5, there are 1-2,1-2-3,1-2-4,1-in the available monitoring path that website 1 generates
5、1-5-6、1-5-7、1-5-7-8.Considering the situation of fixed route, the monitoring path of website to each node only has one.So
And, some monitoring paths are then unnecessary.Unnecessary monitoring path is deleted from available path, does not go to use them can
Spend with effective less detection, increase efficiency.In Fig. 4, can only use by the monitoring path of website to leaf node completely
1-2-3、1-2-4、1-5-6、1-5-7-8.If additionally, the detection site comprising other in the middle of a paths is also deleted, by
Website 1 can generate a monitoring path to node 5, but this path is unnecessary, because comprising another monitoring in path
Website 3, therefore, this monitoring path can be by 1 to 2, and 3 to 5 replace completely.Therefore the monitoring path of other monitoring stations will be comprised
Leave out from available monitoring path.Available monitoring coordinates measurement algorithm is as follows:
4th, congestion point and localization of fault
(1) in order to be accurately positioned problematic link or node present in network, need to collect and research and application path obtains
Monitoring information.After detection site selects to complete, other nodes in website to network all can be generated corresponding monitoring road
Footpath, these paths constitute a monitoring tree, sent monitoring information by website and can obtain on this path in these monitoring paths
Some nodes relevant information such as flow, process postpone, packet drop etc., but the existence of malfunctioning node can make this path
On trouble point after the health status of node cannot know, therefore extra monitoring path needs to continue to be sent to network
In.As shown in Figure 6, there is trouble point 2 in monitoring path 1 to 3 causes the state of 3 to obtain, and needs to continue along path 7 to 3,
Or 6 to 3 transmission monitoring information.Selecting to send detection information bag to all of monitoring path can cause bigger offered load to spend
And it is unnecessary often.But the monitoring path deficiency selecting may result in the congestion point (or trouble point) that cannot position
Too much situation, diagnosis reduces.For the capaciated flow network model building, use reasonably monitoring routing strategy, detection
The flow information of the node of path process, improves monitoring path utilization rate.It is possible not only to position the malfunctioning node in network, also may be used
Quick and precisely to position the node that flow value exceeds a prescribed threshold value---it is referred to as congested node.Algorithm steps is as follows:
Monitoring Path selection, first one group needs to be selected for the monitoring path monitoring the current state of network.First,
In addition to website, all node initializing are measuring point to be checked.For convenience of explanation, the relation square in a monitoring path and node is set up
System of battle formations A.Table 1, row represents monitoring path, and node is shown in list, and (i, j)=1 represents monitoring path i overlay node j to A.Can note
Only covered by a small amount of monitoring path to some node (node 1), from these paths, therefore at least to select (a path
C), at residue node 2 to be detected, in 3,4 (5 are covered by C), node 2 can be covered by minimum monitoring path, and next step selects
The path (A and B) of energy overlay node 2, B can cover more node, and therefore B is preferably to select.Iteration said process is straight
To network, arbitrary node is at least covered by a monitoring path.If node can not be covered by any monitoring path, call it as
Unknown point, and delete from measuring point to be checked.This step false code is as follows.
Table 1
(2) after detective path is selected, each node that corresponding detection information will be mail to by these paths in network.Path warp
The state (being flow value in the present invention) of the node crossed is acquired then, but the existence of malfunctioning node can cause fault on path
The state of the node after Dian cannot be acquired, i.e. remaining node to be detected.
(3) in order to obtain the state of these points, to remaining measuring point to be checked, and remaining detective path step again
1.So repeatedly until point set to be detected is empty set or does not has available detective path.Algorithm stops.The basic procedure of algorithm is such as
Fig. 7.
For what step (1) generated, it is impossible to be acquired the unknown point set of state, owing to its scope is less (or for empty
Collection), it is detected one by one, the congestion point can being accurately positioned in network in conjunction with the congestion point set being diagnosed to be and trouble point.
It is explained in detail below in conjunction with the application effect to the present invention for the emulation.
The present invention uses malfunction monitoring rate Pc=Nf/|Vf|, and do not monitor rate Pu=NuThe calculation of two criterion of/| V |
Method.NfFor the congested node number detecting, | Vf| it is in esse congestion point number, NuJoint for Status unknown in network
Point number.In the network of different average degrees, select different monitoring station numbers, the different feelings of initial congestion node number
The algorithm performing under condition obtains analogous diagram.Using Matlab as emulation tool, analogous diagram takes from the mean value of ten experiments.Figure
It is 20 that 8 (a), (b) are respectively nodes, when network average degree is 3, initial congestion nodes be respectively the 3rd, the 6th, 9 when corresponding examine
Disconnected rate and the result figure of non-diagnosis.The algorithm that can be seen that can obtain very high diagnosis, and relatively low non-diagnosis.Value
When obtaining it is to be noted that there is 6 initial congestion points, according to the infection mechanism of congestion point, the congestion point in network will be greater than 6, and
They can be accurately positioned when now choosing 3 websites.If initial congestion point be the 3rd, the 6th, 9 when according to SNR algorithm requirement,
At least need respectively the 3rd, the 6th, 9 monitoring stations could position all trouble points in network.Even and if choosing at website according to algorithm
When seldom, diagnosis has just reached high value.When initial congestion point is 9, owing to infection mode now congestion point number will be greater than
9, even so, choose 4 website diagnosis just already close to 0.9, and unknown point proportion is only 0.05.Unknown state saves
Point not only comprises the congestion point not being diagnosed to be, it is also possible to comprise owing to the existence of trouble point causes the normal joint of Status unknown
Point.In conjunction with non-diagnosis node collection, the scope of remaining congestion point position is substantially reduced, can examine one by one further
Look into.Fig. 8 (c), (d) are the asynchronous diagnosis of network average degree and non-diagnosis situation, it can be seen that with network average degree
Increasing diagnosis can improve therewith, and non-diagnosis can reduce, because the bigger network of average degree can carry when website number is identical
Improve monitoring efficiency for more can use monitoring path.And Fig. 8 (e), (f) be diagnosis under heterogeneous networks scale with not
Diagnosis contrasts.
The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all essences in the present invention
Any modification, equivalent and the improvement etc. made within god and principle, should be included within the scope of the present invention.
Claims (6)
1. the construction method based on the network fault diagnosis model of flow value, it is characterised in that described based on flow value
The flow transmission mode of the construction method analog network of network fault diagnosis, divide node state is thinner, uses corresponding
Acquisition station choosing method and congested node, malfunctioning node location algorithm, position in network on the premise of choosing less website
Congestion point and trouble point.
2. the construction method of the network fault diagnosis model based on flow value as claimed in claim 1, it is characterised in that described
Comprise the following steps based on the construction method of the network fault diagnosis model of flow value:
Step one, builds the network topological diagram of connection；
Step 2, initializes, and gives the flow value between each node Random assignment 0 to 0.5；
Step 3, randomly selects node, increases the flow value of node, and flow value is more than the node specifying threshold value 0.8 as just
The congested node of step, the node that flow value increases to 1 is considered as trouble point；
Step 4, the infection characteristic of simulation congested node, congested node peripherad neighbor node transport stream value, and by flow
Value is more than the node specifying threshold value as newly-increased congested node；
Step 5, newly-increased congested node continues to affect the neighbor node of oneself, transmits extra flow value, repeat this process.
3. the construction method of the network fault diagnosis model based on flow value as claimed in claim 1, it is characterised in that described
The choosing method of acquisition station includes:
Input: network G=(V, E), maximum probe station number MAXpssize, fault is counted MAXfaults；
Output: acquisition station set of node PS；
Number of network node N=| V |, shaded nodes collection SN=V,
Select the node i of number of degrees maximum as first monitoring station PS=PS ∪ i, update shaded nodes and monitoring path；
Selecting to produce the some c of maximum P (c), as next acquisition station, PS=PS ∪ c updates shaded nodes and monitoring path；
If | PS |=MAXpssize Break；
End。
4. the construction method of the network fault diagnosis model based on flow value as claimed in claim 3, it is characterised in that available
Monitoring path generating method include:
Input: network G=(V, E), acquisition station point set PS；
Output: available monitoring path Ap；
5. the construction method of the network fault diagnosis model based on flow value as claimed in claim 3, it is characterised in that described
As follows to the detection method of remaining measuring point to be checked:
Input: available path Ap, point set Nn to be detected；
Output: for the monitoring path Sp of detection；
The unknown point set of the point set Un that cannot detect, Un；
Step0.While
Step1. to each vi∈ Nn, finds out by the minimum v of path degree of covering in Nnmin, cover v if there is no pathiThen
ViJoin Un, Nn=Nn-vmin, turn step 0；
Step2. at every pathj∈ Ap and covering vminIn, find out path path the longestmax, pathmaxJoin in Sp；
Step3.Ap=Ap-pathmax；
Nn=Nn-vmin。
6. the construction method of the network fault diagnosis model based on flow value as claimed in claim 1, it is characterised in that described
Congestion point includes with fault point positioning method:
Input: available path Ap, point set Nn to be detected；
Output: for the monitoring path Sp of detection；
The point set Un (unknown point set) that cannot detect；
Step 0.While
Step 1. is to each vi∈ Nn, finds out by the minimum v of path degree of covering in Nnmin, cover v if there is no pathiThen
ViJoin in Un, Nn=Nn-vmin, turn step 0；
Step 2. is at every pathj∈ Ap and covering vminIn, find out path path the longestmax,pathmaxJoin in Sp；
Step 3.Ap=Ap-pathmax；
Nn=Nn-vmin。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610416606.2A CN106100882B (en) | 2016-06-14 | 2016-06-14 | A kind of construction method of the network fault diagnosis model based on flow value |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610416606.2A CN106100882B (en) | 2016-06-14 | 2016-06-14 | A kind of construction method of the network fault diagnosis model based on flow value |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106100882A true CN106100882A (en) | 2016-11-09 |
| CN106100882B CN106100882B (en) | 2019-04-30 |
Family
ID=57846228
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610416606.2A Active CN106100882B (en) | 2016-06-14 | 2016-06-14 | A kind of construction method of the network fault diagnosis model based on flow value |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106100882B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110557275A (en) * | 2019-07-12 | 2019-12-10 | 广东电网有限责任公司 | electric power communication network detection station selection algorithm based on network intrinsic characteristics |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101420331A (en) * | 2008-12-12 | 2009-04-29 | 北京邮电大学 | Fast fault locating method for ultra-long connection in T-MPLS network |
| CN101640622A (en) * | 2009-09-07 | 2010-02-03 | 中兴通讯股份有限公司 | Ethernet ring net protection method and transit node |
| JP5102804B2 (en) * | 2009-05-12 | 2012-12-19 | 日本電信電話株式会社 | Congestion impact evaluation apparatus, link traffic calculation method and program thereof |
| WO2014107901A1 (en) * | 2013-01-14 | 2014-07-17 | 华为技术有限公司 | Data storage method, database storage node failure processing method and apparatus |
-
2016
- 2016-06-14 CN CN201610416606.2A patent/CN106100882B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101420331A (en) * | 2008-12-12 | 2009-04-29 | 北京邮电大学 | Fast fault locating method for ultra-long connection in T-MPLS network |
| JP5102804B2 (en) * | 2009-05-12 | 2012-12-19 | 日本電信電話株式会社 | Congestion impact evaluation apparatus, link traffic calculation method and program thereof |
| CN101640622A (en) * | 2009-09-07 | 2010-02-03 | 中兴通讯股份有限公司 | Ethernet ring net protection method and transit node |
| WO2014107901A1 (en) * | 2013-01-14 | 2014-07-17 | 华为技术有限公司 | Data storage method, database storage node failure processing method and apparatus |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110557275A (en) * | 2019-07-12 | 2019-12-10 | 广东电网有限责任公司 | electric power communication network detection station selection algorithm based on network intrinsic characteristics |
| CN110557275B (en) * | 2019-07-12 | 2020-09-25 | 广东电网有限责任公司 | Electric power communication network detection station selection algorithm based on network intrinsic characteristics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106100882B (en) | 2019-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Self-diagnosis for large scale wireless sensor networks | |
| CN104796273B (en) | A kind of method and apparatus of network fault root diagnosis | |
| CN108933694B (en) | Data center network fault node diagnosis method and system based on dial testing data | |
| CN104270268A (en) | Network performance analysis and fault diagnosis method of distributed system | |
| CN102684902B (en) | Based on the network failure locating method of probe prediction | |
| Lu et al. | A new fault detection method for computer networks | |
| Liu et al. | Self-diagnosis for detecting system failures in large-scale wireless sensor networks | |
| Sitanayah et al. | Fault-tolerant relay deployment based on length-constrained connectivity and rerouting centrality in wireless sensor networks | |
| Ochoa-Aday et al. | Self-healing topology discovery protocol for software-defined networks | |
| Ma et al. | Link scanner: Faulty link detection for wireless sensor networks | |
| CN112383934A (en) | Multi-domain cooperation service fault diagnosis method under 5G network slice | |
| CN110557275B (en) | Electric power communication network detection station selection algorithm based on network intrinsic characteristics | |
| CN104486113A (en) | Fault link positioning method based on active greed and passive greed in sensor network | |
| CN111600805A (en) | Bayes-based power data network congestion link inference algorithm | |
| CN107947988A (en) | A kind of Real Time Communication Network analogue system | |
| CN107612772A (en) | The node state detection method and device of payment system | |
| Bermudez et al. | On the infiniband subnet discovery process | |
| CN106100882A (en) | A kind of construction method of the network fault diagnosis model based on flow value | |
| US7646729B2 (en) | Method and apparatus for determination of network topology | |
| Nanda et al. | Mesh-Mon: A multi-radio mesh monitoring and management system | |
| Duarte et al. | Non-broadcast network fault-monitoring based on system-level diagnosis | |
| Breitbart et al. | Effective network monitoring | |
| CN102868572A (en) | Path delay measuring method based on end-to-end detection | |
| Ciulla et al. | Damage detection via shortest-path network sampling | |
| Lee et al. | Diagnosis of TCP overlay connection failures using Bayesian networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |