Research On Fault Diagnosis And Recovery Algorithm Of Service Function Chain In Network Function Virtualization

In order to solve the problem that the Operating Expense and Capital Expense are constantly rising due to the rigidity of special network hardware equipment and the difficulty of expanding new business in traditional network architecture,The network functions virtualization(NFV)is proposed.In NFV,network functions run on general servers in the form of software to form virtual network functions(VNF).Multiple VNFs form a Service Function Chain(SFC)in a certain order,providing users with flexible and efficient network services.However,while this virtualization technology improves network flexibility,it also loses the high reliability of dedicated hardware network devices.How to diagnose and recover faults is the two issues that need to be addressed in network self-healing research.This article conducts research from two aspects: SFC fault diagnosis and recovery.The specific work is as follows:Firstly,in order to meet the users requirement changes and more efficiently utilize network resources,SFC needs to be migrated and reconstructed,resulting in continuous dynamic changes in network topology;The mapping relationship between SFC and the underlying physical network,as well as the chain structure of SFC,leads to the characteristics of vertical and horizontal propagation of faults.For this reason,this paper designs a fault diagnosis model generation algorithm adaptive to network topology changes.When the network topology changes,the algorithm generates the corresponding fault diagnosis model according to the latest network topology after changes.Finally,based on the timing of dynamic Bayesian network,the problem of vertical and horizontal propagation of faults is solved.The experimental results show that the average diagnostic accuracy for all types of faults is as high as 97.8%.Secondly,SFC is an ordered chain structure.In the method of fault recovery by migrating VNF nodes,if the migration destination node is too far away from the physical node hops deployed by the VNF nodes before and after the VNF to be migrated,it will lead to link resource waste and increase SFC latency.Therefore,this article proposes to coordinate the migration of VNF nodes with the migration of their connected virtual links,At the same time,the optimization goal is to minimize the "migration overhead duration",which solves the problem of link resource waste and SFC delay increase caused by improper selection of migration destination node location.The process of finding migration destination node is transformed into a multi-objective optimization problem.This paper uses NSGA-Ⅱ algorithm to achieve multi-objective optimization.The experimental results show that the migration destination nodes obtained through the NSGA-Ⅱ algorithm have the shortest distance from the physical nodes deployed by the VNFs before and after the migration,while minimizing the "migration cost duration",and the "migration destination node resource occupancy rate" is the smallest.This method provides a technical foundation for accurately locating faults and quickly recovering faults in SFC after they occur,and has important practical significance for ensuring the quality of service in NFV networks.