Research On Alternate Routing Algorithm And Adaptive Routing Algorithm Based On Cost Updating For Underwater Backbone Optical Networks

Recently,the submarine optical cable is growing rapidly.Aim to meet the need of the future underwater optical network at the aspect of sensing,access,transmission,routing switch and process,scholars from different countries focus on the basic theory and key technologies of the underwater integrated optical network which supports multiple information sensing and transmission.The ultimate objective is to construct an underwater-integrated optical network to meet the requirements of modern communication system capacity,and fulfill the requirements of sensing,transmission and signal processing.Firstly,this paper introduces the history and research status of underwater optical network,Dijkstra algorithm,and Routing and Wavelength Assignment algorithm.Then this paper analyzes the differences between underwater backbone optical network and terrestrial backbone optical network.Several underwater backbone optical network topologies are analyzed in terms of cost,network delay,survivability,and load status.Finally,the initial planning scheme of underwater backbone optical network topology is determined.This paper points out that the traditional alternate routing algorithm cannot solve the optimal use of main route and alternate route on the initial planning scheme topology.An alternate routing algorithm which based on the KSP algorithm is proposed to solve this problem.This algorithm use bidirectional search strategy,and can effectively solve the optimal alternate routing problem.In the future,the underwater backbone optical network may be developed in accordance with the current terrestrial backbone network which developed toward dense and meshed directions.The traditional fixed routing algorithm,always cause high blocking probability and unbalanced load when making a routing decision.Consequently,the algorithm called CU-RWA algorithm is proposed to optimize route selection in optical networks.The algorithm is able to train the network and constantly update the link cost depending on the load status of link.Simulation results obtained indicate that the proposed algorithm outperforms Dijkstra algorithm in load balancing,blocking probability and link utilization.