Research And Implementation Of Service Function Chain Management Methods In Network Function Virtualization

With the upcoming 5G networking technologies,massive and diverse demands from mobile users will pose significant challenges to the network design and management.To deal with this problem,edge computing has been proposed as a promising technology to provide computation and communication resources at the network edges.However,considering the time-varying characteristics of the application demands,it is infeasible to implement edge computing with traditional hardware-based middleboxes.Network Function Virtualization(NFV)provides a flexible architecture of service and resource management by transferring middleboxes on dedicated hardware to software running on commodity servers through service virtualization.In NFV,a full functional service is provided with a service function chain(SFC),which composes of a number of different virtual network functions(NFs).Through deploying SFCs into underlying network and steering packets passing through them by Software Defined Network(SDN)controller,the NFV-enabled networks will be able to provide services for various traffic demands.In NFV,network operators can arrange the order of network functions in SFCs and deploy them among multiple alternative servers to flexibly meet the needs of mobile users.But this flexibility also brings new challenges to the management of the SFCs.This paper has carried out a series of analysis and research on the above mentioned problems.Specifically,it sorts and classifies the SFCs management methods into two parts: SFCs design methods and SFCs update algorithms.The latest relevant theories and research work of each part are summarized,and their advantages and disadvantages are analyzed.On this basis,this paper mainly studies and solves the following three research work on SFCs management methods:(1)A network function scheduling strategy and SFCs design method with computing resources as the optimization goal;(2)An update mechanism for guaranteeing general consistencies based on time prediction model;(3)A fine-grained update sequence scheduling algorithm between SFCs based on consistent time slots.Finally,we conduct extensive simulations experiments to evaluate all the design methods mentioned above and test the correctness of the principle and the specific performance of aboved works.Then we analyze the results in detail.The experimental results show that our SFCs design method not only reduces the computational resource overhead for SFCs deployment,but also optimizes the link resource overhead to some extent.Besides,the network update rate will be improved greatly by leveraging our SFCs update mechanism and SFCs update sequence scheduling algorithm.