| Network anomalies threaten the network healthy,so network anomaly detection is required to maintain network security.In network anomaly detection methods,link abnormal detection,as a kind of branch,can help locate the source of the anomaly,thus providing a basis for subsequent analysis.The existing network link abnormal detection research methods have problem with large resource overhead,and no available measurement tools for users to customize.Based on the existing network link abnormal detection algorithm,this paper proposes an improved network link abnormal detection algorithm,which can reduce the number of required probes for detection and resource overhead.A network link abnormal detection tool is designed and developed,which can specifically recommend probes and complete anomaly detection for a certain detected path.The paper mainly includes the following two parts.First,this paper proposes an improved network link abnormal detection algorithm,which is as follows:design dynamic path weight algorithm to convert the path overlap times into the influence of path weight;further design independent path recommendation algorithm,select the probe measurement points without overlapping paths.Loop the two algorithms until the fault tolerance range is met,and finally the recommended probe set of the measurement target link is obtained.Secondly,base on the improved network link abnormal detection algorithm,a network link abnormal detection tool is designed and developed.Consider about the network situation and user defined fault tolerance rate,for a certain target link,this tool can recommend suitable measurement probes under the current network conditions,and detect link abnormal delay with minimum number of probes,thereby reduce the resource consumption.The experimental results show that compared with the existing algorithm,the improved algorithm proposed in this paper can effectively reduce the number of used probes by 60%-90%under the premise of ensuring the correct detection rate. |
PDF Full Text Request