Precedence-enabled per-hop behaviour: impact on TCP and UDP flows

Abstract

The author proposes and analyses an approach to support both precedence and preemption (P&P) and QoS handling in common packet transport infrastructures. Precedence has to do with the relative importance of the information content, whereas preemption has to do with mechanisms to deny lower precedence traffic access to network resources in favour of higher precedence traffic, when necessary. The author's approach to this duality is to enhance active queue management (AQM) techniques to provide P&P capabilities and rely on standard, well-studied QoS schedulers, for example, Weighted Round Robin, Class-Based Weighted Fair Queuing, etc., for handling QoS requirements. The present scheme allows low-order queues (within the context of QoS handling) to plead up to the next higher-order queue for help in alleviating queue congestion under periods of communication link overload. The scheme is referred as the cross-queue-AQM (CQ-ACM) scheme and can be extended to higher numbers of queues and any type of scheduler. Through extensive simulation studies, the performance of CQ-AQM scheme under heavy traffic limits is investigated, where preemption is required. The performance metrics of interest to this analysis are packet delay, packet loss and throughput as a function of the packet QoS class and P-L. Further, two new metrics of merit, Gain and System Efficiency, are defined, which are specific to P&P considerations. The studies concentrated on both flow-controlled and non-flow-controlled traffic, as well as mixed traffic conditions. Simulation results show that the algorithms perform extremely well for all cases studied.