Research On Differentiated Protection Mechanism With Traffic Aware For WDM Networks

In order to satisfy the trend of next generation optical network, which ishigh-capacity, intelligence and highly reliablity, network can not only evaluate thefeatures of traffic automatically, but also provide the appropriate reliable qualityassurance to the services of different priority. At the same time, the bandwidth of awavelength is huge, each traffic occupies a wavelength will cause great waste ofnetwork resources, as a result, effective traffic grooming is necessary. However, if thereis no appropriate protection after grooming, when the network failure occurs, muchmore data will lose. Moreover, there are lots of services with different priority in onewave after grooming, a differentiated protection should be taken for better protection.Therefore, differentiated protection based on traffic-aware and differentiated trafficgrooming protection technology is not only indispensable in the development of opticalnetworks, but also the focus of this paper.To classify network traffic type accurately and provide differentiate protectionefficiently based on traffic priority model, a Differentiated Protection Strategy withBayesian based Traffic Aware Classification(DPMBTAC) is presented. This strategydefines information significance to evaluate the features of traffic. Combined withBayesian Network theory, a traffic priority model is established. Meanwhile, a targetedprotection against network failure is provided according to the quality of service(QoS)parameters of different priority traffic. The simulation results show that this strategy canachieve accurate traffic classification, and get better performance in resource utilizationrate and blocking rate than the traditional differentiated protection. But the blocking rategets worse fast with the increasing traffic load. Therefore, it is necessary to studyefficient routing and differentiate protection in heavy traffic load.In order to make the wavelength carry more traffic, solve the problem of blockingrate getting worse fast in heavy load, and provide a reasonable smart differentiatedprotection, a Traffic Aware Differentiated Protection Strategy based on ClusteringTraffic Grooming(TADPM-CTG) is presented. It makes the whole network topologyinto clusters efficiently, to reduce the routing area and speed up grooming for thedynamic traffic. Meanwhile, it uses layered auxiliary graph, residual capacity matrixand cluster Aggregation layer to cooperate with each other for grooming in inter-andintra-clusters, achieving the reasonable planning of resources and the higher efficiency of grooming. After grooming, according to the proportion of different priority servicesin one wavelength, evaluate the link importance, and design a smart P-Cycle to givedifferentiated protection to the links. When the service arrives or leaves, the P-Cyclewill extend or constrict dynamically. The simulation results show this strategy can makea better use of network resource. Meanwhile the blocking rate increases slowly in theheavy load, and shows good performance.