This paper investigates an N server loss system, where the input is a superposition of two types of traffics, namely of a renewal process and a Poisson process. The holding times of the two customer types are exponentially distributed with different parameters. For this model, denoted by GI+Mℝ&bmacr;MℝNℝO, the paper derives a numerical algorithm for computing the individual blocking (loss) probabilities. The analysis is given by constructing a two-dimensional embedded Markov chain and by using the intensity conservation principle as well as point process arguments. The results generalize those of Kuczura and Willie. Finally, for the GI-GIℝ&bmacr;Mℝ1ℝ0 loss system the paper gives a system of partial differential equations for the densities of the steady state distribution and discuss a special case.
