On Pollaczek-Khinchine Formula for Peer-to-Peer Networks

Abstract: The performance analysis of peer-to-peer (P2P) networks calls for a new kind of queueing model, in which jobs and service stations arrive randomly. Except in some simple special cases, in general, the queueing model with varying service rate is mathematically intractable. Motivated by the P-K formula for M/G/1 queue, we developed a limiting analysis approach based on the connection between the fluctuation of service rate and the mean queue length. Considering the two extreme service rates, we proved the conjecture on the lower bound and upper bound of mean queue length previously postulated. Furthermore, an approximate P-K formula to estimate the mean queue length is derived from the convex combination of these two bounds and the conditional mean queue length under the overload condition. We confirmed the accuracy of our approximation by extensive simulation studies with different system parameters. We also verified that all limiting cases of the system behavior are consistent with the predictions of our formula.