Research on a low conflict flow matching hash algorithm

In order to solve the contradiction between the high-speed network traffic and computing resources, some processes like sampling and load balancing need to be made for IP flows. A good flow mapping algorithm can reduce the accumulation effectively and maximize the performance of the communication equipment. In this paper, a low-conflict hash matching algorithm is designed from the point of view of space performance and time performance. First we proposed an improved XOR hashing algorithm based on the characteristics of the quintuple of flows to improve the randomness of the key value. Then we proposed a dual mapping hash algorithm and a key value can be mapped to a memory address accurately. Through this matching algorithm, a searching element in an sequence can be positioned to a very small division away from the true position and it can eliminate the accumulation effectively.