Multi-Scale Queuing Analysis of Long-Range Dependent Traffic and Variable Service Rates

This paper proposes a multi-scale queuing (MSQ) analysis for long-range dependent traffic inputs and both fixed and variable service rates. Based on the observation that the queue length actually shapes both the traffic and service distributions at each time scale, the proposed queue-constrained (QC)-MSQ framework includes this inter-scale coupling effect in the queuing analytical models in order to enhance the estimation accuracy. Simulations are used to verify the accuracy of the analysis. The results show that the proposed analytical technique provides a more accurate estimation of the queue length and delay distributions, and that this accuracy is less affected by the traffic loading, correlation level, and variation.