Detrended fluctuation analysis of traffic data

TL;DR

This paper applies detrended fluctuation analysis to internet traffic data under different routing strategies, revealing three correlation regimes and proposing a three-phase traffic model for better understanding network behavior.

Contribution

It introduces a novel application of DFA to classify traffic correlation regimes and suggests a three-phase traffic model, enhancing previous two-phase models.

Findings

Traffic correlation can be categorized into weak, medium, and strong regimes.

DFA scalings are constant in weak and strong regimes, vary monotonically in medium.

Traffic should be modeled as three phases: free, buffer, and congestion.

Abstract

Different routing strategies may result in different behaviors of traffic on internet. We analyze the correlation of traffic data for three typical routing strategies by the detrended fluctuation analysis (DFA) and find that the degree of correlation of the data can be divided into three regions, i.e., weak, medium, and strong correlation. The DFA scalings are constants in both the regions of weak and strong correlation but monotonously increase in the region of medium correlation. We suggest that it is better to consider the traffic on complex network as three phases, i.e., the free, buffer, and congestion phase, than just as two phases believed before, i.e., the free and congestion phase.