An efficient reduced load heuristic for computing call blocking in large multirate loss networks

We develop an efficient reduced load fixed-point technique for obtaining estimates of the blocking probabilities in large multirate loss systems, which are commonly used to model the call setup phase in ATM networks. The point of departure is that for large systems we exploit a single-link approximation based on large deviation techniques whose computational complexity do not scale with the system size. The explicit characterization then allows us to prove the convergence of the reduced load fixed-point algorithm in the regime of interest for ATM networks applications. The accuracy of the approach is demonstrated on several examples. We compare our results to those obtained by simulation and the stochastic knapsack method which clearly demonstrate the accuracy and effectiveness of the approach.