Comparison of two state reduction techniques for high speed scheduling

Today's internet infrastructure is designed to provide best-effort service which can cause unpredictable delays for sensitive application like voice and audio applications. Thus, today's internet applications require Quality of Service guarantees. In this thesis, we study various QoS techniques that provide packet scheduling algorithms with bounded delay and buffer guarantees. In particular, we are going to study the techniques that attempt to maintain less state of the network flow and still provide QoS guarantee. We emphasize on two of such techniques namely FIFO and Flow Aggregation and compare their performance by simulating their behavior in a network. However, FIFO is time slotted and Flow Aggregation is not. So we first adopt the Flow Aggregation to time slotted system and then do the comparison. The results of the simulation show that Flow Aggregation policy provide better performance than FIFO by accepting more flows with similar QoS requirement.