Examining Self-Similarity Network Traffic Intervals

Abstract

Understanding network traffic pattern and its impact on the Internet provides valuable insights in designing new network protocols, particularly in designing one for applications with a tendency to generate bursty traffic of data, such as Voice over IP (VoIP). To capture the behavior, network traffic can be illustrated on many scales using the notation of self-similarity because network traffic is statistically self-similar. In this paper, we propose a study on analyzing the length of a traffic interval by self-similarity based on the difference between arrival times of packets. We examine the dependency between fast and slow interval as well as a study on the data transition between both intervals.