Analysis of Self-Similar Traffic Based on the On/Off Model

Abstract

This paper analyzes the influence of self-similar traffic on Ethernet based on the On/Off model, which treats the self-similar process as the superposition of tremendous data sources with heavy-tailed distribution. Simulated in several scenarios with different H parameter, the result indicates that self-similar traffic has negative influence on the performance of networks at delay, throughput and other aspects. As a significant property, it has really brought great challenge to the deeper research of today’s network.