Spectral analysis of packet loss rate at statistic multiplexer for multimedia services

A novel concept of spectral representation of packet loss rate is introduced to characterize the transient behavior of information loss during system congestion. It is found that essential characteristics of loss rate spectrum in congestion are captured by the inherent excess input rate spectrum of the arriving process, less dependent on buffer size. In principle, each individual traffic stream can have substantially different requirements for its loss rate spectrum for services. It is shown through selective packet discarding that one can properly redistribute the overall loss rate spectrum among individual sources to significantly improve the service quality of each. The spectral representation of packet loss provides a great convenience to measure the network transmission quality of multimedia services. With the design of multilevel buffer overload control, one can tune the loss rate steady-state distribution of different priority streams to a piecewise step function, which minimizes the damage of packet loss on service qualities