The Parallel Distributed Routing Technology Based On Traffic Identification

With the rapid development of network multimedia technology, Internet has developed from data-only transmission network to the integrated transmission network with data, voice, images and other multimedia information. The traditional network that provides best-effort services has been unable to meet the different requirements of network transmission quality. Emerging multimedia applications put forward higher requirements of network quality of service. Therefore, how to make full use of limited network resources and choose a route with different quality of service for different businesses become a hotspot of network researches.Now, service identification has been a mature technology, and it can effectively and accurately identify network services. Determining the classification of business accurately can help to judge the quality of service requirements exactly, therefore, the combination of service identification technology and routing technology will be a great breakthrough of quality of service research, which can provide better services for different applications.This paper firstly introduces the concepts related to the routing algorithm, and analyzes the strengths and weaknesses of the current routing algorithm. Then it describes the characteristics of a variety of business and their quality of service requirements, and introduces several main identification methods. Through the analysis, a QoS routing model based on service identification is proposed. Besides, the paper explains the design and overall framework of the model, and introduces the function of the service identification module, the DSCP classification module, the queue scheduling module and the QoS routing module function respectively, then focuses on the working process of QoS routing module.On this basis, a parallel distributed QoS routing algorithm based on traffic identification is proposed. The algorithm has the advantages of general distributed algorithms, considers several constraint parameters according to the characteristics of the business and improves the success rate of access requests by using parallel detection methods. To verify the improved algorithm, we run computer simulation through using the call acceptance ratio and traffic cost. The results reveal that the improved algorithm has good performance.