A Dynamic Forecast Load-balancing Algorithm for High-speed Network Instruction Detection System

The high-speed data flow network always makes a serious performance bottleneck on intrusion detection system, leading to the detector of parallel intrusion detection system load unevenly. Packets cannot be fast through the detection system, the processing time tends to make the network more congested and packet loss rate surged high. To solve this problem, a dynamic forecast load balancing scheme for highspeed network intrusion detection system is bring forward. The program calculate the load of detector at some point by dynamic forecast algorithm and the detector can migrate spontaneously or through the control center to ensure that the load between the various detectors are basically the same. In the meanwhile, the load balancing based on connection strategy can effectively maintain the attack evidence. The simulation results proved the feasibility and effectiveness of the program. It can effectively balance the load, reducing the system loss rate and node processing delay.