Adaptive Method for Anomaly Detection Based on Kalman Filter
|School||Shanghai Jiaotong University|
|Course||Communication and Information System|
|Keywords||Kalman filter Single-node anomaly detection Multi-node network monitoring SNMP|
In today 's world , the rapid development of network technology makes the number of nodes on the Internet , network size and topology of the complexity of the increasingly growing in geometric progression . At the same time , however, the emerging network attacks , P2P applications the flooding and computer virus attendant . The abnormal traffic they generate often makes some important network node congestion occurs , application services lost response to serious and even cause paralysis of the entire network . Change the characteristics of the papers through the in-depth analysis the network anomalies occurred before and after the network equipment parameters performance , the improvement of traditional Kalman filter , adaptive exponential smoothing filter noise figure adjustment . And , on this basis , the further use of the advantages of the Kalman filter in real-time , scalability , and forward-looking , an adaptive network anomaly detection algorithm , and in order to establish a network anomaly detection model based on improved Kalman filter . The model for a single node or multiple nodes in a large network of key performance parameters corresponding to the detection, and has low complexity and good real-time performance . Finally , the experimental data collected in a real network environment using SNMP SNMP , verify and analyze the effectiveness of the algorithm and model .