Research And Simulation On The Queue Scheduling Algorithm Based On DiffServ

The study of Diffserv is the direction of the development of IP QoS. To implement the partition of operation, the queue scheduling arithmetic must be used. To satisfy the demand of application's expandability, Diffserv should be simple and efficient. So simpleness and feasibility of the arithmetic should be considered firstly in the study of the scheduling arithmetic. At the same time, the delay and the impartiality should be assured too. Currently in the Diffserv, how to group under the guarantee of QoS is the study hotspot.In this thesis, the advantages and disadvantages and the grope of applibility are obtained through analyzing the implement theory of Intserv and Diffserv. And the working theory of the network nodes and the queue scheduling arithmetic based on Diffserv are emphasized in this thesis. On the foundation of the comparison among the performance of the common scheduling arithmetic, a queue strategy which makes tocken-bucket as the flux adjuster of the network nodes and makes the combination of DWRR arithmetic and PQ arithmetic as the scheduler of the network nodes is presented according to performance index of the scheduling arithmetic and the technology standard of the Diffserv combined with the characteristics of the network nodes. After the basic idea of this strategy is set forth, the transmitting process of the packets is described and the academic reasoning of this strategy is made. At last, this scheduling strategy is tested with the OPNET Modeler and the feasibility of the queue scheduling using this strategy in the network nodes is ascertained with the result of the simulation.