Research on Self-Similar Network Traffic Prediction

Recent studies on network traffic have shown that self-similar is very popular,and the character will not be changed during buffering,switching and transmitting.The character self-similar must be considered in network traffic prediction.This paper analyzed and summarized the research results of self-similar network traffic prediction from the fields of self-similar modeling,parameter computing and performance prediction.An equivalent bandwidth algorithm of self-similar traffic prediction based on measurement was put forward.Our analysis has shown that the algorithm can effectively reduce computing and realizing complexities.