CN115242679A - Method for detecting node downtime in communication network - Google Patents
Method for detecting node downtime in communication network Download PDFInfo
- Publication number
- CN115242679A CN115242679A CN202210912314.3A CN202210912314A CN115242679A CN 115242679 A CN115242679 A CN 115242679A CN 202210912314 A CN202210912314 A CN 202210912314A CN 115242679 A CN115242679 A CN 115242679A
- Authority
- CN
- China
- Prior art keywords
- node
- downtime
- nodes
- model
- models
- 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.)
- Pending
Links
- 238000004891 communication Methods 0.000 title claims abstract description 31
- 238000000034 method Methods 0.000 title claims abstract description 18
- 238000004364 calculation method Methods 0.000 claims abstract description 4
- 239000013598 vector Substances 0.000 claims description 26
- 239000011159 matrix material Substances 0.000 claims description 10
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000000605 extraction Methods 0.000 description 3
- 230000010354 integration Effects 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0805—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability
- H04L43/0817—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability by checking functioning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Environmental & Geological Engineering (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a method for detecting the downtime of nodes in a communication network, which comprises the following steps: s1, preparing data, establishing a graph according to a communication logic link, wherein a node is a server, an edge is a logic communication link, determining a begin-time and an end-time of a time period, and adding a corresponding edge between nodes which have communication in the time period; s2, combining models, namely combining the models to be integrated to obtain a new model, and integrating the models by using GCN and graphsage models to form an integrated model; s3, model training is carried out, the integrated model obtained in the step S2 is trained, the loss is calculated according to parameters in the learning model, and after the loss is obtained through calculation, back propagation is carried out, and the parameters in the module are updated; and S4, a prediction stage. The method for detecting the downtime of the nodes in the communication network can better adapt to complex data and effectively improve the accuracy of node classification.
Description
Technical Field
The invention relates to the technical field of graph data analysis, in particular to detection of node downtime in a communication network.
Background
In a communication network, a server may be down, for example, to ensure normal operation of the communication network, which nodes may be down need to be predicted, and operations such as splitting are performed in time. Wherein a node is a server and an edge is a logical communication link of the node over a period of time. Thus, the problem is converted into a node classification problem on the graph, and a formal definition of the problem is given:
given a graph G, the graph G is a communication network, wherein the nodes in the graph G are servers. Each node has some characteristics, such as running time t, traffic a and downtime frequency b in the past day, and the characteristics are spliced to form a characteristic vector X = [ t, a, b ] of the node, and the characteristic vector X = [ X1, X2, \ 8230 ] of the whole graph is formed after the characteristic vectors of each node are sorted according to the id sequence. The downtime condition of the node needs to be judged so as to take corresponding measures to ensure smooth network communication.
Currently, graph neural networks are used for down node identification. However, the conventional graph neural network, such as GCN, usually only performs a convolution mode, and only can extract a single feature, and cannot perform diversified feature extraction for data of different backgrounds. When complex data is faced, features cannot be effectively extracted, and the classification precision is not high.
Disclosure of Invention
The invention aims to provide a method for detecting the downtime of nodes in a communication network, which can better adapt to complex data and effectively improve the accuracy of node classification.
In order to achieve the above object, the present invention provides a method for detecting a node downtime in a communication network, which comprises the following steps:
s1, preparing data, and establishing a graph according to a communication logic link
The node is a server, the edge is a logic communication link, a time period begin-time and an end-time are determined, and a corresponding edge is added between the nodes which have communication in the time period;
each node has some characteristics, the characteristics are spliced to form a characteristic vector X = [ id, t, a ] of the node, and the characteristic vector X = [ X1, X2, \8230 ] of the whole graph is formed by sorting the characteristic vectors of each node according to the id sequence;
s2, combined model
Combining the models to be integrated to obtain a new model, and integrating the models by using GCN and graphsage models to form an integrated model;
s3, performing model training
Training the integrated model obtained in the step S2, calculating loss of parameters in the learning model, performing back propagation after the loss is obtained through calculation, and updating the parameters in the module;
s4, prediction stage
And (4) obtaining an integrated model by using the training of the step (S3), inputting the characteristics of the nodes to be predicted and the topological structure of the whole network, and obtaining a final node classification result, namely the condition of the downtime of the nodes in the node classification part.
Preferably, in step S1, for the node downtime identification, corresponding historical downtime data needs to be prepared, and a corresponding downtime label is marked on each machine node.
Preferably, in step S2, the input of the integrated model is an adjacency matrix of the communication network structure and a feature matrix of the node, and the adjacency matrix and the feature matrix are respectively input into each network of the integrated model, and a vector of a final output node of each network represents embedding;
and finally splicing the vectors, expressing the vectors by using the aggregated nodes, finally outputting the characteristic expression of the nodes by the integrated model, and classifying the vectors through the full-connection module to obtain the final downtime condition.
Preferably, in step S3, loss is based on cross entropy, and the specific formula is as follows:
wherein p is ic Is the true label of the node corresponding to class c, y ic The node number is a real label of the node, M is a category number, and is 2 here, namely the node number corresponds to two categories of downtime and non-downtime, and N is the node number in the training data.
Therefore, the method for detecting the node downtime in the communication network is adopted, the characteristic extraction is carried out based on the multiple graph neural networks, the multiple-dimensional characteristics can be extracted by integrating the multiple graph neural network modules, the multi-dimensional diversified node vector representations are obtained, the vector representations are combined finally, the final result is obtained by classification finally, the complex data can be better adapted, and the accuracy of node classification can be effectively improved.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
FIG. 1 is a graph neural network framework based on ensemble learning;
FIG. 2 is an embodiment of a graph neural network framework based on ensemble learning.
Detailed Description
The technical scheme of the invention is further explained by the attached drawings and the embodiment.
Unless defined otherwise, technical or scientific terms used herein shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention belongs.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein, and any reference signs in the claims are not intended to be construed as limiting the claim concerned.
Furthermore, it should be understood that although the specification describes embodiments, not every embodiment includes only a single embodiment, and such description is for clarity purposes only, and it will be understood by those skilled in the art that the specification as a whole and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art. These other embodiments are also covered by the scope of the present invention.
It should be understood that the above-mentioned embodiments are only for explaining the present invention, and the protection scope of the present invention is not limited thereto, and any person skilled in the art should be able to cover the technical scope of the present invention and the equivalent replacement or change of the technical solution and the inventive concept thereof in the technical scope of the present invention.
The disclosures of the prior art documents cited in the present description are incorporated by reference in their entirety and are therefore part of the present disclosure.
Example one
A method for detecting the down of a node in a communication network is based on the architecture of an integrated learning graph neural network, and the framework flow of the method is shown in figure 1. And performing feature extraction based on the neural networks of various graphs to obtain diversified vector representations, finally merging the vector representations, and finally classifying to obtain a final result.
The method comprises the following specific steps:
s1, preparing data.
And (4) composition, namely, establishing a graph according to the communication logic link. The nodes are servers and the edges are logical communication links. A time period begin-time and end-time are determined during which corresponding edges are added between nodes that have communications.
Each node has some characteristics, such as node id, operation time t, traffic a of the past day and the like, the characteristics are spliced to form a characteristic vector X = [ id, t, a ] of the node, and the characteristic vector X = [ X1, X2, \8230 ] of the whole graph is formed by sorting the characteristic vectors of each node according to the id sequence.
For the node downtime identification, corresponding historical downtime data needs to be prepared, and each machine node is marked with a corresponding downtime label.
S2, combined model
And combining the models needing to be integrated to obtain a new model. And integrating by using GCN and graphsage models to form an integrated model.
The integration model inputs are the adjacency matrix and the feature matrix of the nodes of the communication network structure. Inputs are respectively input into each network of the integrated model, and the vector representation of the final output node of each network is embedding. And finally, splicing the vectors to obtain a summarized node representation. The final output of the integration model is a feature representation of the nodes. And then, connecting a full-connection module, and classifying to obtain the final downtime condition, as shown in fig. 2.
And S3, performing model training.
The integration model described above is trained. And (3) learning parameters in the model, wherein the optimization mode is to calculate loss and perform back propagation. loss is based on cross entropy, and the specific formula is as follows:
wherein p is ic Is the true label of the node corresponding to class c, y ic The node number is a real label of the node, M is a category number, and is 2 here, namely the node number corresponds to two categories of downtime and non-downtime, and N is the node number in the training data.
And after the loss is obtained through calculation, performing back propagation and updating the parameters in the module.
S4, prediction stage
And (3) obtaining an integrated model by using the training, inputting the characteristics of the nodes to be predicted and the topological structure of the whole network, and obtaining a final node classification result, namely the condition that the nodes are down in the node classification part.
Therefore, the method for detecting the node downtime in the communication network can better adapt to complex data and effectively improve the accuracy of node classification.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the disclosed embodiments without departing from the spirit and scope of the present invention.
Claims (4)
1. A method for detecting node downtime in a communication network is characterized by comprising the following steps:
s1, preparing data, and establishing a graph according to a communication logic link
The node is a server, the edge is a logic communication link, a time period begin-time and an end-time are determined, and a corresponding edge is added between the nodes which have communication in the time period;
each node has some characteristics, the characteristics are spliced to form a characteristic vector X = [ id, t, a ] of the node, and the characteristic vector X = [ X1, X2, \8230 ] of the whole graph is formed by sorting the characteristic vectors of each node according to the id sequence;
s2, combined model
Combining the models to be integrated to obtain a new model, and integrating the models by using GCN and graphsage models to form an integrated model;
s3, model training is carried out
Training the integrated model obtained in the step S2, calculating loss of parameters in the learning model, performing back propagation after the loss is obtained through calculation, and updating the parameters in the module;
s4, prediction stage
And (4) obtaining an integrated model by using the training in the step (S3), inputting the characteristics of the nodes to be predicted and the topological structure of the whole network, and obtaining a final node classification result, namely the condition that the nodes are down in the node classification part.
2. The method according to claim 1, wherein said method comprises: in step S1, for node downtime identification, corresponding historical downtime data needs to be prepared, and a corresponding downtime label is marked on each machine node.
3. The method of claim 1, wherein the method comprises: in the step S2, the input of the integrated model is an adjacent matrix of a communication network structure and a characteristic matrix of a node, the adjacent matrix and the characteristic matrix are respectively input into each network of the integrated model, and the vector of the final output node of each network represents embedding;
and finally splicing the vectors, expressing the vectors by using the aggregated nodes, finally outputting the characteristic expression of the nodes by the integrated model, and classifying the vectors through the full-connection module to obtain the final downtime condition.
4. The method according to claim 1, wherein said method comprises: in step S3, loss is based on the cross entropy L, and the specific formula is as follows:
wherein p is ic Is the probability that the node is class c, y ic The node is a real label corresponding to class c, M is a class number, and is 2 here, namely corresponding to two classes of downtime and non-downtime, and N is the number of nodes in the training data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210912314.3A CN115242679A (en) | 2022-07-30 | 2022-07-30 | Method for detecting node downtime in communication network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210912314.3A CN115242679A (en) | 2022-07-30 | 2022-07-30 | Method for detecting node downtime in communication network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115242679A true CN115242679A (en) | 2022-10-25 |
Family
ID=83676647
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210912314.3A Pending CN115242679A (en) | 2022-07-30 | 2022-07-30 | Method for detecting node downtime in communication network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115242679A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111966076A (en) * | 2020-08-11 | 2020-11-20 | 广东工业大学 | Fault positioning method based on finite-state machine and graph neural network |
| CN113162787A (en) * | 2020-01-23 | 2021-07-23 | 华为技术有限公司 | Method for fault location in a telecommunication network, node classification method and related device |
| CN114646839A (en) * | 2022-01-28 | 2022-06-21 | 国网河北省电力有限公司保定供电分公司 | Power distribution network fault section positioning method and device based on graph convolution neural network |
-
2022
- 2022-07-30 CN CN202210912314.3A patent/CN115242679A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113162787A (en) * | 2020-01-23 | 2021-07-23 | 华为技术有限公司 | Method for fault location in a telecommunication network, node classification method and related device |
| CN111966076A (en) * | 2020-08-11 | 2020-11-20 | 广东工业大学 | Fault positioning method based on finite-state machine and graph neural network |
| CN114646839A (en) * | 2022-01-28 | 2022-06-21 | 国网河北省电力有限公司保定供电分公司 | Power distribution network fault section positioning method and device based on graph convolution neural network |
Non-Patent Citations (1)
| Title |
|---|
| CHEN, ZHIWEN ET AL.: "《arXiv:2111.08185》" * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021164382A1 (en) | Method and apparatus for performing feature processing for user classification model | |
| CN109685110B (en) | Training method of image classification network, image classification method and device, and server | |
| CN112685504B (en) | Production process-oriented distributed migration chart learning method | |
| CN111611488B (en) | Information recommendation method and device based on artificial intelligence and electronic equipment | |
| CN105471647A (en) | Power communication network fault positioning method | |
| CN114816997B (en) | Defect prediction method based on graph neural network and bidirectional GRU feature extraction | |
| CN113254669B (en) | Knowledge graph-based power distribution network CIM model information completion method and system | |
| CN109787821B (en) | Intelligent prediction method for large-scale mobile client traffic consumption | |
| CN112217674A (en) | Alarm root cause identification method based on causal network mining and graph attention network | |
| CN110009045A (en) | The recognition methods of internet-of-things terminal and device | |
| CN113312494A (en) | Vertical domain knowledge graph construction method, system, equipment and storage medium | |
| CN107729939A (en) | A kind of CIM extended method and device towards newly-increased power network resources | |
| CN113779988A (en) | Method for extracting process knowledge events in communication field | |
| CN116402352A (en) | Enterprise risk prediction method and device, electronic equipment and medium | |
| CN114663002A (en) | Method and equipment for automatically matching performance assessment indexes | |
| CN113868537B (en) | Recommendation method based on multi-behavior session graph fusion | |
| CN116010813A (en) | Community detection method based on influence degree of fusion label nodes of graph neural network | |
| CN114219531A (en) | Waste mobile phone dynamic pricing method based on M-WU concept drift detection | |
| CN112184317A (en) | Waste mobile phone pricing method based on value-preserving rate and discrete neural network | |
| CN115238808A (en) | Method for detecting node downtime in communication network based on multiple tasks | |
| CN115438190B (en) | Power distribution network fault auxiliary decision knowledge extraction method and system | |
| CN115242679A (en) | Method for detecting node downtime in communication network | |
| CN116541792A (en) | Method for carrying out group partner identification based on graph neural network node classification | |
| CN115734274A (en) | Cellular network fault diagnosis method based on deep learning and knowledge graph | |
| CN115936763A (en) | Vehicle user loss prediction method based on feature derivation and feature screening |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221025 |