Research On QoS Queue Scheduling Algorithm Based On IP Network

With the development of Internet technology and the increasing number of service types, thetraditional network which provides “best effort” service can’t fulfill the quality of service (QoS) andthen leads to network congestion.IP network has been changed to multi-services digital networkfrom the single digital network. Against this background, how to guarantee the network QoS is akey issue in the further development of the network. Queue scheduling is often used to achieve anefficient allocation of resources. For that reason whether network use appropriate schedulingalgorithm is of great importance to meet the requirement of QoS.In this paper, it firstly discusses the reason of network congestion and congestion controlmethod, then it analyses IntServ and DiffServ. The DiffServ model with excellent expansibilitycould satisfy the practical application. And it discusses the advantages and disadvantages of thecommon scheduling algorithm. The paper finally focuses on the WFQ algorithm, it exists twodefects: the shortcoming of distinguishing between the real-time services and ensuring the qualityof network under the condition of bursting. So it proposes the LS_WFQ algorithm based on WFQalgorithm. In the LS_WFQ algorithm it makes token-bucket as the flux adjuster of the network andadds an adaptive queue manager for managing the length of the queues. It also distinguishesbetween real-time services and non real-time services, and then adjusts weighted value according tothe length of the queue. In this paper, we simulate the WFQ algorithm and the LS_WFQ algorithmunder the same network topology. The simulation results shows LS_WFQ algorithm is better thanWFQ algorithm on the performances of transport delay, bandwidth allocation, etc and it could fulfillthe QoS under the condition of bursting.