Performance analyis of internet switch with variable length packet input traffic using Circulant Markov Modulated Poisson Process (CMMPP)

摘要

Many of the classical self-similar network traffic models are asymptotic in character, Hence, they are not as much of effective in queuing based performance estimate. Self-similar characteristic is most important for network traffic engineering as it causes degradation of network performance. In this paper, we investigate mean waiting time and packet loss probability of the switch under self similar variable of length in packet traffic can be modeling it as Queuing systems CMMPP / M /1 and CMMPP / M /1 / K, where as an examine time is exponential and Self-Similar traffic is emulated by Circulant Markov-Modulated Poisson Process (CMMPP). CMMPP model is already validated one to emulate the self-similar characteristics. We compare the analytical results with the simulation ones.