Congestion probabilities in CDMA-based networks supporting batched Poisson traffic

We propose a new multirate teletraffic loss model for the calculation of time and call\ud congestion probabilities in CDMA-based networks that accommodate calls of different serviceclasses\ud whose arrival follows a batched Poisson process. The latter is more “peaked” and\ud “bursty” than the ordinary Poisson process. The acceptance of calls in the system is based on the\ud partial batch blocking discipline. This policy accepts a part of the batch (one or more calls) and\ud discards the rest if the available resources are not enough to accept the whole batch. The\ud proposed model takes into account the multiple access interference, the notion of local (soft)\ud blocking, user’s activity and the interference cancellation. Although the analysis of the model\ud does not lead to a product form solution of the steady state probabilities, we show that the\ud calculation of the call-level performance metrics, time and call congestion probabilities, can be\ud based on approximate but recursive formulas. The accuracy of the proposed formulas are\ud verified through simulation and found to be quite satisfactory