Research On IP Fast Recovery Based On Multiple Routing Configuration

With the constantly emerging of real time applications such as video telephony,electronic sports,on-line shopping and so on,a new challenge to the rapid recovery of network failures is presented.However,because the traditional routing protocol has a series of problems,such as long convergence time,packet loss,and even loop,congestion,and so on.When the network fails,it can not meet the needs of real-time business.Therefore,this paper studied the technology of IP network fast recovery.IP network fast recovery can achieve fast convergence through calculating the backup routing in advance.Based on active recovery technology,this paper mainly researches on algorithms of IP fast recovery based on multiple routing configuration.The main contents of this paper include the following aspects:First,there are fewer links in backup topology for multiple routing configuration,this paper proposes an improved backup topology design method.The main idea is to define the critical nodes according to the value of Betweenness Centrality and Closeness of nodes in backup topology,and then maximizes available links of Critical nodes to reduce the number of hops in the shortest path.The two methods are adopted to select the critical nodes: the first is Top K method,it is select K nodes with higher value of Betweenness and Closeness;the second method is Non-adjacent K method,when selecting the critical nodes,considering the location of the critical nodes,the adjacent nodes cannot be used as the critical nodes.The experimental results show that the proposed algorithm can achieve better results in the networks,and maximum reduce the number of reroute hops.Second,the paper analyses the problem of traffic congestion which may occur after the network fault recovery using multiple routing configuration and presents a load balancing method.The model first creates an objective function related to link load,and then uses heuristic method to obtain a series of weight values that can reach the optimal solution,through giving each link an appropriate weight to distribute network traffic.From the simulation results,we can know that the modified algorithm achieves more balanced load distribution of network traffic by reducing the maximum link utilization.