Research On Service Function Chain Orchestration Technologies For Resource Optimization In Network Function Virtualization

Network Function Virtualization(NFV)decouples network functions from dedicated hardware and operates them as Virtual Network Functions(VNFs)in commodity servers.The Service Function Chain(SFC)in NFV environment has the advantages of low operating cost,strong portability,and fast service update,which provides a good solution for promoting the development of emerging network services.Appropriate SFC orchestrations method can help operators save basic resources needed to provide network services,cut down the operating costs.Thus,SFC orchestration technology for resource optimization is an important research field.This dissertation focuses on the SFC orchestration technologies for resource optimization in NFV.Aiming at improving the SFC orchestration mechanism for resource optimization,the dissertation has proposed solutions for the problem of optimizing the delay and the resource allocation of network services jointly,improving the elastic scaling ability of virtual resource,and enhancing the survivability of service function chains.The proposed solutions can help operators provide network services with good quality,high adaptability and survivability,improve resource utilization and reduce operating costs.The main innovations are as follows:In the existing SFC orchestration methods,the optimization process of resource and quality of service are separated.Solution spaces are limited by each other.To solve this problem,we propose a spatial and temporal optimal method of SFC orchestration based on an overlay network structure.The resource cost and related delay of SFC can be abstracted into the links' weight of overlay networks,which helps to convert the joint optimization problem into the shortest path problem that can be easily solved.As for the SFC requests set requiring batch processing,we design an overlay network based iterative optimal orchestration algorithm.The simulation results demonstrate that the proposed orchestration method can reduce the end-to-end delay by 21.6%,the utilization ratio of link bandwidth resource by 12.4%,and the operational expenditure by 15.2%.The acceptance ratio of requests set can be improved by 22.3%.Most existing virtual resource scaling schemes are limited to reactive policy,lacking online elastic deployment methods.We propose an Online Elastic Orchestration Policy(OEOP)based on the perception of flow rate evolution to solve this problem.OEOP predetermines the fine-grained VNF scaling demands by the evolution of flow rate.The amount of VNF instances can be adjusted adaptively before the service arrives.The impact of delay which is caused by the process of VNF deployment on the service continuity can be avoided.In addition,the online elastic VNF deployment is achieved according to the real-time update information of SFC paths and the load of nodes.The simulation results demonstrate that OEOP can enhance the matching between virtual resource supply and workload demand.Comparing with other algorithms,the throughput of VNF is improved by 10.2%~24.8%,and the operational expenditure can be reduced by 20.7% on average.To maintain the survivability of SFC,the existing methods face to the problems of high backup redundancy and high resource overhead.So an efficient backup sharing mechanism of survivable SFC is proposed.This mechanism adopts redundant backup strategy,which needs to deploy backup VNF instances in physical network to ensure the survivability of the SFC.The survivable SFC backup problem is modeled as facility location problem,and an efficient backup deployment algorithm is designed to compress the redundancy of backup resource in order to reducing the operating costs.Then,we use coalitional game strategy to adjust the mapping relations between primary and backup VNF instances,which can ensure the synchronization by optimizing the delay of links.The experiments show that this mechanism effectively implements backup redundancy compression under the premise of ensuring the survivability of SFC.The timeliness of synchronization information between primary and backup VNF instances can be improved.The operating cost of our mechanism can be reduced by 12.3%.