Optimized rule-based delay proportion adjustment for proportional differentiated services

We present a novel method to adjust output queue delay proportion fairly among traffic classes of different priorities in relative differentiated services. The delay proportion adjustment is based on acceleration of incoming traffic in each class. It aims to reduce the undesirable effects of queue-delay propagation toward higher priority classes, caused by the introduction of bursty data into lower priority classes. We use a fuzzy controller to make the decision regarding the amount of proportion adjustment, as it is very flexible and adjustable. We suggest an efficient extension to the particle swarm optimization algorithm for the purpose of optimizing the fuzzy system. The simulation shows that the dependency of high-priority-class delay, which is a value that indicates quality-of-service of the traffic, on lower priority classes is significantly reduced by the proposed delay proportion adjustment.