MIB Based Method For Detecting Anomaly Nodes In Large Scale Network

Network anomaly detection is to discover network failure and security issues and timely diagnosis of exclusion of the network anomaly. With the development of internet and increment of network nodes, there is a great challenge for network anomaly detection. How to detect network anomaly and locate anomaly quickly and efficiently has become a major research focus. By the real-time network data from MIB, we can find network anomaly quickly and efficiently for router nodes and application server nodes respectively, which are the main types of nodes in the large-scale network.The paper outlines the development of network management, analyses the current research of network anomaly detection and the advantages and disadvantages of the existing detection method for network anomaly nodes.Secondly, the paper introduces the basic concepts and general principles of SNMP and MIB, then detailed describes the real-time data acquisition method for the network data of MIB by SNMP.Thirdly, the paper outlines the router nodes detection method based on network traffic data in MIB. Through the analysis of the practical network data, the paper describes the router running situation and the detection standards of abnormal network. Then we do lots of experiments on detection method in real network, and do a performance analysis of the results. This method can detect anomalies in network through the router quickly and efficiently, choose the network anomaly detection sensitivity flexibly in accordance with customer demand for network applications and services. Finally, the paper introduces the application server nodes detection method based on performance indexes data in MIB. We do an in-depth research on various indexes of the covariance, building a covariance matrix, find failure by detection of matrix main Eigen value and the unusual circumstances of main Eigen value. This method only needs few historical data, and it has a high rate of accuracy and precision.