Fault Tolerant Routing Of Data Center Optical Interconnection Network Based On SDN

The optical interconnect data center network has the advantages of high bandwidth,high flexibility,low overhead,etc.It is a very promising data center network architecture.By introducing software-defined network technology to the optical interconnect data center,the control layer is separated from the forwarding layer.Operators can use logically centralized software to control the entire network,optimize network resources from a global perspective,and enhance data center network flexibility.The data center optical interconnection network carries a large number of services.If the network fails,there will be significant losses and impacts.Therefore,whether the data center network service transmission is not affected will become a crucial issue.The fault tolerance of the network is the ability to guarantee successful transmission after a failure in the network.In this thesis,the problem is studied in detail,proposed a path risk coefficient and congestion perception fault-tolerant mechanism and a optimized ant colony algorithm recovery method.This thesis analyzes from the perspectives of protection strategy and recovery strategy.Aiming at the problem that the protection strategy is prone to congestion under multi-fault conditions and the failure recovery success rate is low,this thesis proposes a fault-tolerant mechanism based on path risk coefficient and congestion perception.By calculating the number of flows on the network link and the bandwidth usage rate,the risk coefficient of the link is obtained,and the path with the higher risk coefficient is more likely to be congested or faulty in the network fault state.The algorithm in this thesis avoids the path with higher risk level when selecting the protection path.In the fault recovery process,the congestion awareness adjustment mechanism is added.If the protection path is found to be congested,a path is recalculated for the damaged service.The simulation analysis of the scheme with PRS algorithm,segmentation routing algorithm and fault backtracking algorithm proves that the scheme has lower fault recovery delay,fewer flow entries and improve the success rate of failure recovery under higher fault recovery.The recovery strategy rerouting algorithm has a long calculation time and is prone to congestion and load imbalance.In this thesis,a recovery method based on optimized ant colony algorithm is proposed.By the advantage of heuristic algorithm find approximate optimal solution fast.In the iterative process of ant colony algorithm,bandwidthcorrelation factor and load balancing factor are added to update pheromone concentration.The optimization improves the iterative speed of the ant colony algorithm,so that the fault recovery service can be allocated to the link with the appropriate bandwidth capacity and load.The load balancing in the fault recovery process is realized.The simulated analysis of the scheme,the ant colony algorithm based on genetic algorithm and adaptive ant colony algorithm proves that the scheme can effectively reduce the fault recovery delay and link bandwidth occupancy,and can achieve load balancing under fault recovery.