The impact of correlation on delay performance of high speed networks

This paper focuses on the impact of correlation on the delay performance of high speed networks. Markov-modulated Poisson process (MMPP), which captures both burstiness and correlation between interarrival times, is used to model packet arrival processes in high speed networks. Only one hop of a high speed network is studied since queueing networks can be approximately analyzed by decomposing the network into several separate elements. For simplicity, the authors model this one hop of the network as an MMPP/M/1 queue. The delay performance of MMPP/M/1 queue is investigated in terms of the MMPP correlation. A comparison among correlation, burstiness, and utilization impacts on delay performance is presented.<
>