Validating novel CAC algorithms on ATM testbeds

This paper is concerned with validating novel CAC algorithms developed in the framework of project COP579. The validation took place at the Expert Testing Facilities in Basel, Switzerland, where versatile source models and traffic traces were available from video, datafast and voice classes. The paper briefly summarizes the tested CAC algorithms then discusses mapping. The notion of mapping aims at finding an equivalent real environment in which all relevant aspects of the performance of different CAC algorithms can be tested without the validation process getting bogged down in side effects (e.g., unrealistically long measuring time to capture rare events etc.). After discussing the testing principles, different performance measures were introduced, based on how the CAC algorithms can be ranked. By analyzing the results a "toplist" of CAC algorithms was produced which can help when considering real implementation. Finally the reliability of the measurements and the inferred conclusions are addressed.