A kind of postpone teardown congestion control algorithm based on critical link prediction in multilayer intelligent optical networks

摘要

Congestion problem is an important factor affecting network performance, this paper studies the congestion control problem in multilayer intelligent optical networks (IONs), proposed a kind of postponed teardown congestion control algorithm (MinLP_PTD), which based on critical link prediction and using grooming strategy. Firstly, considering the expected congestion of the link, defined critical link. Secondly, when building a path for business, using the definition of critical link to judge whether the light-paths contain critical links, and then under the given network topology designed a new algorithm based on critical link forecast CLF-KSP. Thirdly, in order to save network resources and transport sub-wavelength business in multilayer IONs, using grooming strategy to make the newly arrived business to transport on the existing light-paths, which remaining bandwidth can meet the business. The minimum light-path dynamic traffic grooming algorithm is designed (MinLP), which uses CLF-KSP algorithm to establish light-path for existing light-path can not satisfy the business. Fourth, the detailed design of MinLP_PTD algorithm is presented. Fifth, simulation is carried out, and the results show: MinLP_PTD algorithm not only reduces the network blocking rate, but also improves network resource utilization, achieves an outstanding performance of the network.