Research On Algorithms For Survivable Virtual Network Embedding Problems

With the rapid development of the Internet,network traffic has grown rapidly,it is difficult for current network resources to meet people's needs.However,due to the unusually large and complex structure of the current network system,it is difficult to transform,the ossifycation of the Internet has become increasingly prominent.Network virtualization technology s physical resources without changing the current network structure.It is an effective way to solve the ossification problems of the current rigid Internet network.Virtual network embedding is the research focus of network virtualization,the traditional virtual network embedding algorithms mainly focus on improving the utilization of the underlying resources,and there are few studies on the survivability of virtual networks.Due to factors such as equipment aging,hacking technology improvement,etc,the failure of the underlying network becomes more frequent,and the research of survivable virtual network embedding algorithm becomes extremely important.Due to a single heuristic algorithm is easy to fall into local optimization and the convergence speed is slow,using a single heuristic algorithm to embed will cause waste of the underlying network resources.in response to this problem,this paper proposes an Adaptive Genetic Framework and Simulated Annealing Algorithm for virtual network embedding.Improvements:first,the algorithm uses simulated annealing algorithm instead of genetic algorithm mutation operation,improving the ability of the algorithm to jump out of the local optimal solution.Second,the algorithm designs a new parent selection operator based on the annealing idea,further reducing the probability of falling into a local optimum at the initial stage of the algorithm.Third,the algorithm designs the adaptive crossover operator to make the algorithm converge quickly and effectively.The experimental results show that the algorithm can effectively improve the evaluation indexes of virtual network embedding,such as the acceptance rate of virtual network requests,the revenue of physical network,and the physical network revenue cost ratio.Aiming at the problem of low virtual network request acceptance rate and low revenue of physical network in the survivable virtual network embedding algorithm,this paper proposes a hybrid mechanism survivable embedding algorithm,with specific improvements as follows:first,the algorithm designs a survivable embedding strategy based on protection mechanism.The core part is to design the objective TOPSIS method to predict the underlying unreliable nodes,and judge whether to give backup to the virtual network and reduce the consumption of backup resources.Second,the algorithm designs a survivable embedding strategy based on the recovery mechanism.The center of the strategy is to design survivable embedding constraints,which can effectively improve the recovery rate of virtual networks without backup.Third,the algorithm combines two strategies,the hybrid survivable embedding algorithm can ensure that the virtual network failure recovery rate and the physical network revenue can be improved without significantly reducing the virtual network request acceptance rate.