Joint Optimization of VNF Placement and Scheduling Based on Pathfinder Algorithm
Virtual Network Function Placement and Virtual Network Function Scheduling are hot topic in recent years.They are closely connected and indispensable.Unified consideration of the two links and global optimization is not only the trend of tech-nology development,but also one of the difficult challenges facing the implementation of Network Function Virtualization.Research focusing only on a single problem lacks a holistic view.In this paper,a dynamic joint optimization model of VNF placement and scheduling is proposed,which minimizes the completion time of service function chain while satisfying relevant constraints.An im-proved pathfinder algorithm is proposed to solve this problem,and the mechanism ofgroup division is introduced to enhance the glob-al search ability and avoid falling into local optimum.Furthermore,the mechanism of individual crossover and mutation is designed to reduce the destruction of individual solutions and enhance the diversity of the population.In addition,elite reservation mecha-nism is introduced to accelerate the convergence of the algorithm.Experimental results demonstrate that,compared with typical evo-lutionary algorithms,the algorithm proposed in this paper has significant advantages inthe evaluation index of service function chain completion time.
network function virtualizationVNF Placement and Schedulingdynamic modelpathfinder algorithm
国家科技期刊平台
NSTL
万方数据
