INTERNET TRAFFIC AND PACKET ROUND TRIP DELAY SELF-SIMILARITY

Abstract

Statistical analysis was carried out on the Internet
traffic data grabbed at Internet Service Provider
gateway. The analysis reveals the "bursty" nature,
which was explained through SelfSimilarity and
Long Range Dependence(LRD). In this work, it has
been shown that the Internet traffic and Packet
Round Trip Delay visuaiized as a time series are
statistically selfsimilar. The autocorrelation
function decays reveals the data under
investigation exhibits Long Memory or Long Range
Dependence. The degree of self similarity of Packet
Round Trip Delay as measured by Hurst Parameter
is presented and was found that it is direct~v
related to network traffic load.