摘要
针对网络功能虚拟化(NFV)环境下核心网控制面资源分配问题,提出性能感知的资源分配算法.基于排队网络理论建立核心网控制面性能评估模型,推导出信令流程平均响应时间的近似表达式.为了确定核心网控制面虚拟网络功能(VNF)实例的最优配置数量,综合考虑处理性能和VNF实例部署成本,建立核心网控制面资源分配多目标优化模型,并提出改进的多目标遗传算法.仿真结果表明,该性能评估模型误差在10%以内,优于Jackson排队网络模型;与NSGA-II和HaD-MOEA相比,所提算法获得的近似Pareto前沿收敛性和多样性更好,更逼近真实Pareto前沿.
A performance-aware resource allocation algorithm was proposed aiming at the resource allocation problem of the core network control plane in network function virtualization(NFV)environment.Based on the queuing network theory,a performance evaluation model for the control plane was established,and an approximate expression for the average response time of the signaling procedures was derived.Further,considering both the processing performance and the deployment cost of virtual network function(VNF)instances,a multi-objective optimization model was developed for resource allocation of the control plane,and an improved multi-objective genetic algorithm was proposed.Simulation results showed that the error of the performance evaluation model was within 10%and the model was better than the Jackson network model.Compared with NSGA-II and HaD-MOEA,the approximate Pareto front obtained by the proposed algorithm was better in terms of convergence and diversity,and was closer to the real Pareto front.
作者
陈俊杰
李洪均
曹张华
CHEN Jun-jie;LI Hong-jun;CAO Zhang-hua(School of Information Science and Technology,Nantong University,Nantong 226019,China;Nantong Research Institute for Advanced Communication Technologies,Nantong 226019,China)
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2021年第9期1782-1787,共6页
Journal of Zhejiang University：Engineering Science
基金
南通市科技计划资助项目(JC2018025).
关键词
核心网
资源分配
排队网络
多目标优化
拥挤距离
core network
resource allocation
queueing network
multi-objective optimization
crowding distance
