Research On The Queue Scheduling Algorithm Based On DiffServ

With the rapid development of Internet technology, there is a greater demand for network quality of service(QoS).The traditional IP network provided the "Best-Effort" services, it mainly designed for the non real-time data without the requirement to ensure real-time data transferring, so it can’t satisfy the different quality of service in the complex network environment. In this context, IETF group proposed DiffServ Model, Queue scheduling algorithm for the model to become the core technology of IP QoS.Firstly,The thesis introduces the background and principle of IntServ and DiffServ module,describe two models of their strengths and weaknesses. The complexity of the network makes any kind of model cannot be used independently to achieve network QoS,so the IntServ and DiffServ service model combines two mutually complementary coordination.Secondly, analyses the causes of network congestion and several ways of congestion control, illustrate that queue scheduling algorithm is effective for solving network congestion,so the queue scheduling mechanism has been the focus of research, as the core content of the paper.On the basis of the DWRR proposed an improved algorithm VDWRR, and insert a new queue management thinking.With NS2 network simulator to verify the feasibility and effectiveness of VDWRR algorithm.Finally, In order to make up for the Round Robin algorithm and PRI algorithm(priority scheduling algorithm) respective defects, the PRI combined with VDWRR algorithm to put forward PRI+VDWRR algorithm, NS2 network simulator evaluated the algorithm,The scheduling policy in ensuring the flow of real-time while EF, provide a reliable non-guaranteed bandwidth while not EF.