On the Use of Economical Theories and Metrics to Design an Intelligent Active Queue Management in the Internet

Active Queue Management (AQM) plays an important role in the Internet congestion control. Random Early Detection (RED) is the most popular active queue management algorithm that is used in the in Internet routers. The effectiveness of RED algorithm highly depends on appropriate setting of its parameters. Moreover, RED uses only the average queue length as a congestion meter to trigger packet dropping as a congestion feedback. Since the average queue length considers only long-term behavior of any queue, this approach fails to see instantaneous changes of the queue and hence its reaction is not fast enough. This paper inspires from the economic principles to design a new congestion control algorithm. This algorithm, called ECORED, enhances the network performance by dynamically tuning of RED's parameters. ECORED introduces two meters i.e. queue length growth velocity and drop rate velocity to measure the congestion level and produce appropriate congestion feedbacks. Based on extensive simulations conducted using Network Simulator-2 (ns-2), we show that ECORED algorithm reduces the packet loss rate and achieves high utilization in compare with RED, ARED and NLRED.