A state-space reduction method for computing the cell loss probability in ATM networks

Cell loss performance analysis in ATM networks has been considered as one of the most important issues in congestion control which is vital to the success of the ATM technique. In recent years, numerous approaches have been proposed for the computation of the cell loss probability in ATM networks. Among them are the stochastic fluid-flow (SFF) method and the Markov modulated deterministic process (MMDP) approach. The MMDP approach is basically a discrete version of the SFF method and it is numerical stable. Both approaches, however, require a considerable amount of computation time for problems of practical size. In this paper, we propose a state-space reduction method for both approaches. The idea is to find an appropriate tradeoff between the efficiency and accuracy. As a result, the proposed approach performs very well in terms of both accuracy and efficiency. For cases of practical interest (in which cell loss probabilities ranges from 10^-6 to 10^-10) its computation time is only a fraction of one CPU second