Research On Optimal Deployment Of Service Function Chain In Cloud-Fog Networks

As two key technologies of next generation network,Network Function Virtualization(NFV)and fog computing are getting more and more attention.NFV provides a good paradigm for sharing the resources of the physical network.Virtual Network Functions(VNFs)compose the Service Function Chain(SFC)in a particular order,and network users can communicate with each other or service terminals through SFCs.Most studies on NFV or SFC are based on cloud computing environments.Moreover,as the resource demand for cloud computing increases,the abilities of centralized cloud computing has also been challenged.To solve the challenges of the centralized cloud computing,the researchers have proposed the distributed fog/edge computing.Fog/edge computing has been proposed as a supplement to cloud computing.Currently,in order to make better use of the advantages of fog/edge computing,there is a new research focus on computation offloading,which considers to offload some cloud computing services to the fog/edge network,so as to reduce the traffic load of cloud computing.Services offloading and fog/edge computing provide novel paradigms to solve the problem of resource consumption and the delay of cloud network.In addition,as two of the key technologies of 5G,Fog-based Radio Access Network(FRAN)and NFV have become the important direction for the radio network architecture evolution.Therefore,the autuor of this dissertation researches the problem of SFCs deploying in cloud-fog computing environments.When deploying an SFC,using the paradigm that combines cloud computing and fog computing can effectively reduce the resource pressure on cloud networks.The main contributions of this dissertation are as follows.(1)Since the SFCs deployment problem in mobile network is a NP-hard problem,for service providers,it is crucially important for efficient deploying SFCs into mobile network.So,this dissertation proposes SFC deployment algorithms for solving this challenge with two main design goals: i)minimizing the cost of link resource;ii)minimizing the cost of computing resources.The dissertation models the SFCs deployment problem as an optimization problem by using Linear Programming(LP),as well devises a heuristic algorithm to make a trade-off between these two conflicting design goals.The simulation results show that compared with the existing algorithms,the total SFC mapping cost of the proposed algorithm is reduced by about 14%,the total VNF mapping cost is reduced by about 40%,the total SFC link mapping cost is reduced by about 5%,and the SFC deployment success rate is increased by about 25%.(2)The problem of combining SFCs supporting live online services to reduce the network resource consumption need to be solved.To effectively take advantage of cloud-fog computing and thus achieve the goal of saving resources,this dissertation studies the SFC combination and deployment problem in cloud-fog computing environments.To solve this problem,the dissertation presents an efficient SFC combination and deployment algorithm(SFCM-CC).Finally,the dissertation conducts extensive simulations to evaluate the performance of the proposed algorithm.The results show that when the proportion of online live service increases by 10%,the total SFC mapping cost,total VNF mapping cost and total SFC link mapping cost will be reduced by about 10% respectively,and the SFC deployment success rate will be increased by about 10%.(3)Most of the current researches on services offloading consider offloading the running computing services from mobile phone or cloud network to fog/edge network.However,in the process of offloading,it will consume additional resources and affect the user's experience.Therefore,in this dissertation,when an SFC is deployed,the dissertation considers offloading and deploying a part of VNFs of each SFC into fog/edge network through effective online game decision,thereby reducing the traffic load of cloud network.The dissertation proposes a Delay-minimum SFC Deploying algorithm based on Services Offloading and Online Game(DSFCD-SOOG)that combines effective online game and services offloading decision to deploy the SFCs in cloud-fog computing networks,to reduce the end-to-end delay of SFC.Finally,the dissertation verifies the superiority of the DSFCD-SOOG algorithm in terms of resource efficiency,SFC delay and deployment success ratio in SFC deployment.Compared with the existing algorithms,the average VNF deployment cost of the proposed algorithm is reduced by about 5%,the average SFC link deployment cost is reduced by about 14%,the average SFC deployment cost is reduced by about 10%,the average SFC path delay is reduced by about 14%,and the SFC deployment success rate is increased by about40%.(4)When an online user's location changes,the SFC may be migrated.Therefore,the dissertation researches the problem of SFC migration/remapping caused by the user movement in cloud-fog environments.The dissertation firstly models the migration problem of SFCs as an LP;then the dissertation proposes two SFC migration strategies:the minimum number of VNFs migration strategy and the two-step migration strategy;and the dissertation has designed a two-step migration algorithm to migrate SFCs.The dissertation uses the cloud-fog computing environment to evaluate the proposed algorithm.The reconfiguration cost,the remapping success ratio,the migration time and downtime of proposed algorithms are more excellent than that of benchmark algorithm.Compared with the existing algorithms,the reconfiguration cost of the two-step migration algorithm is reduced by about 30%,the running time is reduced by about 66%,the downtime is reduced by about 85%,the migration time is reduced by about 55%,and the remapping success rate of SFC migration requests is greatly improved by about 34%.