Modeling and simulation of congestion control algorithms

Consider an Internet traffic source sending packets into a single link connected to another source. The source's "window" is the maximum number of packets discharged without waiting for an acknowledgement, at any point in time. Such a window concept is the cornerstone of a number of congestion control algorithms, such as Tahoe, Reno, New Reno, Vegas and Sierra (partly rate-based). We develop novel mathematical methods for the performance levels of all these algorithms. It is shown that Sierra is by far, analytically, the most superior of these algorithms. We also then present simulation results that confirm this position experimentally.