A Simulation Study on Scheduling Real-Time and Non-Real-Time Traffic

Abstract The Queue Length Threshold (QLT) algorithm prioritizes real-time tasks until the number of accumulated non-real-time tasks exceeds a threshold. On the other hand, with the Minimum Laxity Threshold (MLT) algorithm non-real-time tasks are given priority until the minimum laxity of the real-time tasks exceeds a threshold. However, to the best of our knowledge, the optimal setting of these two thresholds has not been previously studied. Nor has a performance study been carried out on these two algorithms in comparison with other scheduling algorithms. This paper determines, by using sensitivity analysis, the optimum values of the two thresholds and then uses these optimum values to carry out the performance study on the MLT and QLT algorithms in comparison with other scheduling policies. Systems such as multimedia systems, ATM networks and communication systems can then use these optimum values to schedule both their real-time tasks and non-real-time tasks in a most efficient way.