Robust active queue management for uncertain TCP network congestion control in Internet

Congestion control is surely a mean to enhance network performances. Active queue management (AQM), which is a proactive approach, has been proposed as a solution to these problems [1]. One of the most well-known AQM policies is random early detection (RED). Although RED is an effective TCP congestion control [2], it can induce network instability and major traffic disruption if not properly configured. One of RED’s main weaknesses is that the average queue size varies with the level of congestion and with the parameter settings. Moreover, the average queuing delay from RED is sensitive to the traffic load and also to parameters setting and is therefore not predictable in advance. Delay is a major component of the quality of service delivered to the customers, so network operators would naturally like to roughly estimate the average delays in their congested routers.