The Research And Implementation Of Self Adaptive Traffic Classification Control Technology

In the fixed network, the router can provide reliable data forwarding service, but it has no ability of changing its service automatically as the network changed. It can only achieve optimal performance in special network environment and can not provide reliable data forwarding services to ensure QoS(Quality of Service), therefore it is necessary to study the self adaptive traffic classification control technology to rationalize the allocation of resources.This paper focused on solving the problems mentioned above and the main contribution includes four parts:First, the research on IP QoS mechanism and the key technologies of IP QoS; Second, analyzed the mechanism of Linux traffic control in detail and the code of the traffic control module in Linux; Third, the research on a self-adaptive queue scheduling technology which supported fast switching between scheduling policies and its realization; Fourth, the research on two self adaptive traffic classification control technologies and its realization.In the research of traffic control technology and the way of its realization in Linux system, this paper involves in how to support the QoS, Qdisc, queue scheduling algorithms, TBF algorithms, traffic shaping, congestion control, especially the code of traffic control moduleIn the research of self-adaptive queue scheduling technology, this paper proposed and implemented a new queue discipline in Linux platform, which called PDQ(Priority DRR Queuing), and it contains two algorithms PQ (Priority Queue) and DRR (Deficit Round Robin), and it can provide fast switching between the algorithms according to the change of the network environment.In the research of self-adaptive traffic classification control technology, two methods of Implementation were proposed:One is based on link monitoring, the other is based on traffic Monitoring, and the testing results of each methods were also given and be analyzed in this paper, which provides a highly value in development Self Adaptive Traffic Classification Control Technology.