The Deployment And Migration Of Virtual Network Function Chain In NFV Networks

With the rapid development of communication network,both the number of users and the traffic demand of users have increased dramatically,which makes the current networks face many huge challenges such as the complex network deployment,the vertical and closed network architecture,the mismatch between traffic and telecom revenue and so on.The emergence and development of NFV(Network Function Virtualization)technology has greatly promoted the evolution of the communication network architecture.In the network based on NFV,virtualization and cloud computing technologies are employed to manage network functions via software as opposed to having to rely on hardware to handle these functions.By shifting dedicated,hardware-based network functions could be processed to software running on commercial hardware platform.The application of SFC(Service Function Chain)is an important use of NFV technology,which could manage the flow of specific services and applications flexibly.At the same time,SFC supports very granular and elastic service delivery,the modification of network functions,and the movement of application workloads in the existing network,so the study on SFC has been widespread concern both at home and abroad.In this dissertation,we investigated the related technologies of SFC for the network based on NFV technology,and we discussed deployment and migration of the virtual service function chain.Specifically,our research work is focused on the following two issues.First of all,we note that there is such a phenomenon in the network,which is due to the limitations of the physical resources generated by the debris,resulting in each host has insufficient resources to deploy the service function chain while there are enough resources in the whole network,reducing the number of service requests accepted in the network.For the sake of this problem,we aim to maximize the number of SFCs which are satisfied with the end-to-end delay requirement of the network service requests in the resource-constrained network.We establish the mathematical model for the problem of virtual service function chains' deployment and propose a deployment scheme of SFCs based on heuristic greedy algorithm.The simulation results show that our algorithm could effectively ensure maximized the number of SFCs in the resource-constrained network,taking into account both making full use of the physical host and physical switch resources in the network and the network performance.Next,it is necessary to study the migration of virtual service function chains because of the mobility of the users and the load balancing in the network,to realize the effective allocation of resources and the flexibility of the network.However,there are only a few researches on the migration of virtual SFCs.In this thesis,especially considering the characteristics of actual network function,the two important factors of network performance and the rational utilization of resources in the network are taken into account for the migration and reconfiguration problem of SFCs.We build a mathematical model,and present a time-delay optimization of the service function chain reassignment strategy,which is based on improved genetic algorithm.By comparison with the greedy migration algorithm and random migration algorithm,the simulation results show that the virtual SFCs reconfiguration scheme based on the improved genetic algorithm could effectively guarantee the network performance and take into account the utilization of network physical resources at the same time.Based on the analysis of NFV technology and service function chain,this dissertation designs the deployment and migration reconfiguration scheme of virtual service function chain in NFV networks,which provides a new idea for the deployment and migration of service function chain in NFV networks.