Print Email Facebook Twitter Performance analysis of ATM buffers with prioritized actual traffic Title Performance analysis of ATM buffers with prioritized actual traffic Author Van den Berg, R.E.M. Contributor Schoute, F. (mentor) Stavrov, B. (mentor) Faculty Electrical Engineering, Mathematics and Computer Science Department Telecommunicatie- en Verkeersbegeleidingssystemen Date 1997-03-07 Abstract The awareness that long-range dependence (or self-similarity) is present in data traffic and will have its effect on important traffic management issues is growing. The results of performance analysis and traffic engineering considerations drastically change when based on actual traffic instead of non-realistic Markovian models. In this paper, we use real -measured- data to examine the performance of a FIFO and an LDOLL (low delay or low loss) ATM buffer. We show that although LDOLL enhances the performance when loaded with normal distributed traffic, this effect is far less with more realistic traffic distributions. The effect of traffic modeling and parameters such as the amount of self-similarity on the performance of LDOLL are discussed. Subject Long- Range DependenceLDOLLATMWWWQueuingMeasurementPerformanceSelf-Similarity To reference this document use: http://resolver.tudelft.nl/uuid:b783974e-83af-4036-a7d4-e536012ace24 Part of collection Student theses Document type master thesis Rights (c) 1997 R.E.M. van den Berg Files PDF AA278942_8.pdf 859.75 KB Close viewer /islandora/object/uuid:b783974e-83af-4036-a7d4-e536012ace24/datastream/OBJ/view