a b s t r a c tThe Misra, Gong and Towsley (MGT) fluid model for Transmission Control Protocol (TCP) has been used extensively to guide designs of Active Queue Management (AQM) schemes aiming to alleviate adverse effects of Internet congestion. In this paper, we provide a new analysis of a TCP/AQM system that aims to improve the accuracy of the MGT fluid model especially in heavy traffic conditions. The analysis is based on the consideration of two extreme congestion scenarios that leads to the derivation of upper and lower bounds for the queue length and marking probability dynamics and showing that they approach each other in steady state. Both discrete and continuous time models are provided. Simulation results demonstrate that the new model achieves a significantly higher level of accuracy than a simplified version of the MGT fluid model.

This paper outlines a new multi-layer optimization approach that considers stochastic traffic composed of heavy tailed flows for optimizing traffic engineering and grooming, network design, dimensioning and planning, and evolution prediction.

No abstract