Simulations for Blocking Probability Evaluation of a Loss System Loaded by Optical Flow Switching Traffic

Abstract

How to achieve the balance between utilization and Quality of Service (QoS) is an important issue for Optical Flow Switching (OFS) network. For a single router, reallocating capacity every time period by predicting the traffic later on is a more efficient and cost-effective way than using fixed number of circuits because the previous one requires less capacity guaranteeing the same blocking probability. However, the analytical solution for arranging suitable number of wavelengths based on the condition of the network is difficult. Thus, in this project, a simulation approach is introduced as to evaluate the blocking probability of a loss system loaded by optical flow switching traffic. We tried a new way of simulation that obtains the blocking probabilities in a certain time intervals given that there have already been some circuits in use at the beginning. Various scenarios with different numbers of circuits allocated and a wide range of time intervals were tested. The results reveal the impact of length of time interval and number of circuits upon blocking probability while other conditions are the same. Furthermore, we used different mean arrival rates, mean flow lengths and traffic load to see if they will influence blocking probability.