Analysis of Ethernet Traffic Statistical Properties

Abstract

Traffic profile, mainly its self-similarity properties, can have crucial impact on the network performance. In this regard, we evaluate traffic profile of Ethernet traffic. We have performed a measurement of the traffic on Ethernet network. Captured data has been analyzed from the protocol point of view with the stress on the self-similarity, LRD and SRD properties. To evaluate these characteristics, properties of wavelet transform (DWT) are deployed and, based on alpha parameter, scaling property of traffic is estimated. We show that self-similarity is present in analyzed data and that it depends on analyzed time scale and on analyzed protocol