Dynamic optimization of packet classification rules organization based on Internet traffic statistical characters

Aimed at the natural weakness of the commonly used linear search packet classification algorithm in the aspect of time efficiency,this paper proposed a novel algorithm which dynamically optimized the rules organization based on the Internet traffic statistical characteristics.This algorithm included calculating the weights of rules,optimizing the order of rules,dynamically triggered optimizing and so on.Because of the minimized resource consumption,this algorithm can practically improve the system performance and is shown to be efficient in the actual tests.