Streaming and Elastic Traffic in Closs Networks

Rapid progress in telecommunications which started in the end of 20th century raises new teletraffic problems. The result of the convergence of different networks stems that set of different traffic types should be served by the system. These types differ by their characteristics, in network bandwidth and Quality of Service (QoS) requirements. Diverse consumer needs, variety of on-line services and high-speed Internet access joined in the same system generate large amount of traffic most part of which includes elastic traffic. Multi-rate communication systems, developed at the beginning of 21th century as a result of engineering progress, are able to serve elastic traffic with variable rate depending on the number of prioritized streaming requests being processed in parallel. Specific characteristics of elastic traffic do not allow us to use Erlang multi-rate loss model for simulation of modern telecommunication systems. This paper concerns Erlang-like multi-service loss model of an adaptive multi-rate system for streaming and elastic traffic, where a set of bandwidth requirements corresponds to each type of elastic traffic. A bandwidth requirement is assigned to an entered elastic request depending on occupied bandwidth. The results obtained in the paper are applied to calculation of QoS parameters of the Closs network topology which is widely used in modern switching networks.

Erlang multi-rate loss model, elastic traffic, streaming (realtime) traffic, threshold admission control, blocking probability, network performance, Kaufman-Roberts recursion