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

Currently, preferential packet handling in transport networks is based solely upon application-level Quality of Service (QoS) requirements. No preferential packet handling is based upon the importance of the information being carried by the network. In future transport networks, packet handling should provide preferential transport to important, high precedence traffic, specifically during conditions of resource scarcity, e.g., network overload conditions, while simultaneously satisfying packet scheduling required to meet application QoS needs. We propose and analyze an approach to supporting both Precedence and Preemption (P&P) and QoS handling in common transport infrastructures. Precedence has to do with the relative importance of the information content while preemption has to do with mechanisms to deny lower precedence traffic access to network resources in favor of higher precedence traffic, when necessary. Our approach to this duality is to enhance Active Queue Management (AQM) techniques to provide P&P capabilities and rely upon standard, well studied QoS schedulers, e.g., Weighted Round Robin, Class-Based Fair Queuing, etc., for handling QoS requirements. We refer to this combination as a Precedence-Enabled Per Hop Behavior (PHB). In this way, when operating under engineered loads, the well known scheduling algorithms support high quality QoS for applications. Under network congestion situations, the enhanced AQM layer provides the necessary P&P preferential packet handling favoring high Precedence-Level (P-L) information. Our 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. We refer to our scheme as the Cross Queue-AQM (CQ-ACM) Scheme. Our scheme can be extended to higher numbers of queues and any type of scheduler in a straightforward manner. Through extensive simulation studies, we investigate the performance of our CQ-AQM scheme under heavy traffic limits, where preemption is required. The performance metrics of interest to our analysis are packet delay, packet loss and throughput as a function of the packet QoS class and P-L. We also define here two new metrics of merit, which are specific to P&P considerations. These are the Gain, which measures the benefit of the P-L handling to the high P-L traffic, and the System Efficiency, which measures the lost total system capacity due to implementation of P-L handling. Our studies concentrated on both flow controlled and non-flow controlled traffic, as well as mixed traffic conditions. Our simulation results show that our algorithms performed extremely well. We find that the application of our CQ-AQM scheme on top of standard QoS scheduling is effective in simultaneously supporting QoS and P&P transport and assures the delivery of high precedence traffic under all ranges of traffic loads and traffic types studied. So our local CQ-AQM scheme, which is extremely simple to implement and extends to all types of existing QoS schedulers, is an extremely useful tool in the overall architecture for packet-based, precedence-enabled, future transport services.