Survivability For Traffic Grooming In Heterogenous WDM Networks

Traffic grooming is used to provide multiple low-rate connections for a high-rate wavelength channel in WDM networks. With network equipments updating, network nodes have different capacity of wavelength switching and traffic grooming. This heterogeneous network increases the complexity of traffic grooming. The network survivability refers to the capacity of maintaining the required quality of service after network faults, and this is one of key technologies in next generation networks.To solving above problems and improving the network survivability, this thesis addresses on traffic grooming in heterogeneous WDM mesh networks. The content of this thesis includes three issues: (1) analysis and modeling of heterogeneous WDM mesh networks; (2) efficient policies for traffic grooming; (3) dynamic restoration schemes for survivable traffic grooming.The contributions of this thesis include:(1) A transceiver saving auxiliary graph model (TSAG) is proposed for traffic grooming in heterogeneous WDM mesh networks. Different edges in the model present various network resources, such as wavelengths, transceivers and lightpaths. An algorithm of transceiver saving based on the model is also presented, which can determine whether a connection consume a grooming port or not when routed to the wavelength channel. Simulation results show that the model saves amount of transceiver resources, and have high network throughput and short running time, in the scenarios of static traffic and dynamic traffic.(2) A minimizing cost grooming policy is proposed for static traffic in WDM mesh networks based on TSAG model. The policy assigns the weight value of auxiliary edges according to the cost ratio between wavelength resources and transceiver resources. It can implement the object of minimizing the total network cost. (3) A resource-efficient grooming policy is proposed for dynamic traffic in WDM mesh networks base on TSAG model. The policy dynamically adjusts grooming schemes according to the ratio of available number between wavelength resources and transceivers. Simulation results show that the policy can achieve a low blocking probability and high resource utilization.(4) Based on TSAG model, two dynamic restoration schemes are proposed for survivable traffic grooming– dynamic restoration at lightpath level (DRAL) and dynamic restoration at connection level (DRAC). DRAL dynamically discovers resources to restore faults for lightpaths and DRAC dynamically discovers resources to restore faults for connections. They don't reserve any resources for backup paths. An evaluation model is presented to evaluate the performance of dynamic restoration schemes. Simulation results show that two schemes have low blocking probability and short processing time and their restoration probability are also above 90%.