Design of active queue management based on the correlations in internet traffic

The implementation of Active Queue Management (AQM) in Internet routers is a recommendation of the Internet Engineering Task Force. However, none of the current AQM algorithms directly makes use of the correlations present in Internet traffic, which are related to Long Range Dependence (LRD) and Self-Similarity. This is somewhat surprising in that it is well known that LRD is exhibited by Internet traffic, particularly at the edges of the network. In this paper, an algorithm called Modified Adaptive Random Early Detection is presented. This algorithm makes use of the correlations in the mean queue length in order to predict likely congestion scenarios in the not too distant future. These can then be acted on anticipatively. The results are extremely encouraging and demonstrate a superior performance when benchmarked against current mechanisms. The proposed algorithm could also be non-intrusively implemented in current routers.