Abstract: Theoretical analysis of packet loss rate with multiple self-similar traffic input was proposed based on FBM model and the condition with Gaussian input. The results show that if the mean and variance of each traffic flow are approxmitately equal, the traffic flow with the largest Hurst parameter has a decisive influence on the packet loss rate of multiple traffic. This conclusion was verified by simulation and can be used for the analysis of network node performance, buffer design, and resource allocation.

Key words: multiple input, self-similar traffic, packet loss rate