Induced Burstiness in Generalized Processor Sharing Queues with Long-Tailed Traffic Flows

We analyze the asymptotic behavior of long-tailed traffic flows under the Generalized Processor Sharing (GPS) discipline. GPS-based scheduling algorithms, such as Weighted Fair Queueing, have emerged as important mechanisms for achieving differentiated quality-ofservice in integrated networks. We prove that, in certain scenarios, a flow may be strongly affected by the activity of `heavier'- tailed flows, and may inherit their traffic characteristics, causing induced burstiness. This phenomenon contrasts with previous results which show that, under certain conditions, an individual flow with long-tailed traffic characteristics is effectively served at a constant rate. In particular, the flow is then essentially immune from excessive activity of flows with `heavier'- tailed traffic characteristics. The stark contrast in qualitative behavior illustrates the crucial importance of the GPS weights in protecting individual flows. Appeared in: Proceedings of the 37th Annual Allerton Conference..