A strategy for comparing reliable multicast protocols applied to RMNP and CTES

An increasing number of networked multimedia applications do require a reliable multicast service to support the distribution of data to a potentially large number of receivers, where the receivers may be located in a sparse mode over a WAN network. Designing a reliable multicast protocol is certainly a challenging task, and the literature is abundant in proposals with different approaches to the problem. Because of the large number of proposals, and considering that none of them appears to be clearly superior to the others, a quantitative performance comparison is required to find out their strengths and drawbacks. As the complexity of the algorithms and protocols is quite high, analytical approaches may be ruled out. A simulation approach may provide acceptably reliable results as long as it is built following the more or less well established guidelines. However, building up the model of the system to be simulated in a way that it assures fairness in the comparison of "competing" reliable multicast protocols requires a detailed study. A strategy to compare different approaches to reliable multicast is presented. This strategy has been applied to a concrete case, by comparing two state-of-the-art protocols: reliable multicast network protocol (RMNP) and control tree based on end systems (CTES).