POWer Adaptive Random Early Detection for Diff-Serv Assured Forwarding Service Classes

Abstract

With diverse performance requirements and ever- changing nature of Internet multimedia applications, differential treatment to these applications is inevitable. In this paper we evaluated the performances of popular Adaptive Random Early Detection (ARED) [7] algorithms against our POWer Adaptive Random Early Detection (POWARED) [8] algorithm in providing different throughput and delay assurances to various Differentiated Services (DiffServ) Assured Forwarding (AF) service classes. The results show that POWARED generally outperforms ARED in terms of loss rate and throughput with minimal tradeoffs in delay. DiffServ model is specifically chosen due to its wide market acceptance by industrial players in providing Quality of Service (QoS) across the Internet.