VNF Mapping And Migration Method For Edge Multi-service Scenarios

With the development of Internet of Things(IoT)and edge computing in recent years,the types and number of IoT services are increasing rapidly,which put great pressure on the network's data transmission and processing.To alleviate the pressure,it is crucial to combine edge computing technology and network function virtualization(NFV)technology to sink computation capacity from the core network to the edge network and enhance the service processing capability of edge-side.The existing virtual resource allocation method could be divided into two parts:resource mapping and migration,which have problems such as the inability to optimize cost and load simultaneously,large energy consumption,or revenue loss of service,respectively.Consequently,for the sake of the high-quality and long-term operation of the edge network,it is of great significance to aim at the characteristics of edge network and IoT services and study the virtual resource VNF allocation mechanism.Aiming at the problem that the cost optimization and load balancing could not be realized simultaneously in current resource mapping method,a VNF balanced mapping method(VNFM-CBR)based on cost and reliability is proposed.The method includes two steps:initial primary link VNF mapping and redundant backup selection.In the mapping process,VNFM-CBR firstly establishes the importance evaluation VNF mapping model,then designs constraints and evaluation factors related to physical nodes and links,finally designs the mapping algorithm.Based on preprocessing the network view,the mapping algorithm advances the parent selection,new area discovery,and solution accepting mechanisms of the heuristic algorithm,which sufficiently improves the quality of feasible solutions.In the backup process,VNFM-CBR firstly builds a model for evaluating the effectiveness of redundant backups,then designs a redundant backup algorithm based on this model.VNFM-CBR designs a dual-mode mechani sm of link backup and VNF backup,then it adopts the backup sharing mechanism to effectively reduce backup cost and improve the reliability of the initial primary link.Simulation results show that the VNFM-CBR can effectively balance the load,optimize the cost,and achieve high reliability of the service.Aiming at the problems that imbalanced node load,massive revenue loss,or energy consumption in current resource migration,a VNF dynamic migration method based on an improved Monte Carlo algorithm was proposed(VNFM-EQA).VNFM-EQA firstly builds the edge network traffic model and VNF migration model,then establishes a multi-dimensional cost assessment VNF dynamic migration model based on traffic trend and migration scenario,finally proposes a corresponding delay-assisted migration algorithm.The migration algorithm redefines the exploration and expansion evaluation factors and introduces a taboo table to achieve the expansion of the solution space coverage and the improvement of the quality.Simulation results show that VNFM-EQA can effectively integrate virtual resources,reduce network energy consumption and revenue loss.