Packet Loss Probability Approximation in High-Speed Networks Based on Self-Similar Traffic

As the packet/cell loss ratio (PLR/CLR) is counted as one of the major quality of service (QoS) requirements; in this paper we proposed a linear upper bound expression for the loss probability approximation in high-speed communication network technologies. These technologies can support multiple classes of traffic with different QoS requirements and diverse traffic characteristics, by the use of traffic multiplexing. The proposed expression can be calculated in real-time, so it can be properly used in the network multiplexers. Based on the fact that in many cases, the real traffic in high-speed networks shows a self-similar behavior, we have used self-similar traffic with Pareto distribution as the input traffic model in our PLR approximation simulations. We have studied application of this expression in non-homogeneous as well as homogeneous cases.