I. Introduction
Evaluation is an integral part of network protocol design and protocol improvement. Since protocol properties become more complex to understand intuitively, it is more important than ever to systematically assess protocols. Complementing analytical and testbed-based evaluation methods, simulative approaches have proved to be a powerful tool for evaluating network protocols. The most common [1] network simulator class, discrete event simulation, combines several beneficial properties. By giving full control over otherwise external influences on simulation results, discrete event simulations allow to reproduce results easier than, e.g., using testbeds. Moreover, the decoupling of simulated time and real time allows to run single simulations faster. Finally, discrete event simulators support perfect repeatability, which means that repeating simulations with the same parameters results in exactly the same results. Thereby, discrete event simulators allow to scrutinize protocol behavior caused by rare network constellations, such as a certain sequence of packet losses.