Analysis of a statistical multiplexer under a general input traffic model

The queueing behavior of a statistical multiplexer with a finite number of input lines is studied. Each input line is assumed to deliver fixed-length packets of information according to a generally distributed process whose parameters depend on the state of an underlying finite-state Markov chain associated with each of the input lines. A method for the derivation of the moments of the buffer occupancy is developed, and the first and second moments are derived. The mean packet delay introduced by the statistical multiplexer is derived through Little´s theorem. The applicability of the analyzed multiplexer in packet communication systems is illustrated through a simple example, and numerical results are provided for this case