Characterizing Internet backbone traffic from macro to micro

Abstract

The continuous growth in both commercial and public network traffic with various quality-of-service (QoS) requirements is calling for better service than the Internet's best effort mechanism. One of the challenging issues is to predict the overall behavior of aggregate network traffic. While network traffic characterization has been studied extensively due to its importance in network scheduling and throughput, an accurate characterization of network traffic still remains elusive. In this paper, in addition to characterizing the aggregate network traffic, we classify the traffic into different categories, e.g., P2P, VOIP, and provide insight to each of them in terms of their traffic pattern and impact to the overall traffic. Our study verifies that like many works reported in literature, majority Internet backbone traffic is contributed by a small portion of users. A major contribution of this paper is that we found a linear equation which could be used to approximate the impact of each user to the overall traffic. Many new applications appear recently and become more and more popular as Internet evolves. We show that in current Internet backbone of China, 9 applications, which can be classified into three categories, web browsing, P2P service, and gaming, contribute 95 percent of the total traffic. It is also demonstrated the P2P applications are the dominant traffic contributor among the three categories.