Study of Stochastic Characteristics of Traffic on a High Speed Network

Abstract

With the ever increasing backbone bandwidth, the design and planning of networks are usually done by simulating the influence of various traffic type on the network. Accurate traffic classification is the key stone of numerous network activities. The research in this paper results from a study and analysis of packet trace obtained from a backbone link in NITNET. Millions of packets are passed through the link that makes the traffic measurement difficult or even impossible. An attempt has been made to study the online inbound and outbound packet trace analysis. Poisson based sampling methodology has been used to estimate the characteristics and distribution of packets.