基于拓扑预配置的公平虚拟网络映射算法

发布时间：2018-02-24 20:02

本文关键词： 网络虚拟化虚拟网络映射节点可重用虚拟拓扑预配置离散粒子群优化算法　出处：《计算机研究与发展》2017年01期 　论文类型：期刊论文

【摘要】：虚拟网络映射是实现云环境下资源多租赁运营及弹性计算资源服务的关键基础环节,其目的是在满足虚拟网络资源需求的前提下将虚拟网络植入到合适的底层物理节点和链路.现有虚拟网络映射算法的研究成果大多以极大化物理资源利用率为目标,对虚拟网络请求排队中的公平性问题考虑较少.为此提出了一种基于虚拟拓扑预配置及可重用技术的虚拟网络映射算法以提高映射公平性.将虚拟网路映射过程分为2步骤:拓扑预配置过程和映射过程.1)对在线队列中较大的虚拟网络拓扑进行等价变换,将其变换为节点及链路数目更小的拓扑,减少虚拟网络请求在拓扑上的差异从而提高公平性;2)建立形式化的虚拟网络映射模型,并利用离散粒子群算法对优化模型进行求解;为了充分利用可重用技术能在求解过程中节省带宽资源的特性,引入粒子位置分配增强机制以提高物理网络资源利用率.仿真实验结果表明:提出的算法在物理网络资源利用率、收益/成本比及虚拟网络接受公平性等方面均优于已有同类算法.
[Abstract]:Virtual network mapping is the key link to realize multi-lease operation and flexible computing resource service in cloud environment. The purpose is to implant the virtual network into the appropriate physical nodes and links on the premise of satisfying the requirements of the virtual network resources. Most of the research results of the existing virtual network mapping algorithms aim at maximizing the utilization ratio of physical resources. In this paper, a virtual network mapping algorithm based on virtual topology preconfiguration and reusable technology is proposed to improve the mapping fairness. It is divided into two steps: topology preconfiguration process and mapping process. 1) equivalent transformation of large virtual network topology in online queue. Transform it into a topology with smaller number of nodes and links, reduce the difference in topology between the requests of virtual network and improve the fairness. (2) establish a formal virtual network mapping model, and use discrete particle swarm optimization algorithm to solve the optimization model; In order to make full use of reusable technology to save bandwidth resources in the solution process, Particle location assignment enhancement mechanism is introduced to improve the utilization ratio of physical network resources. The simulation results show that the proposed algorithm can improve the utilization ratio of physical network resources. The income / cost ratio and the acceptance fairness of the virtual network are better than the existing algorithms.
【作者单位】：东北大学秦皇岛分校计算机与通信工程学院;广东外语外贸大学思科信息学院;东北大学软件学院;
【基金】：国家杰出青年科学基金项目(61225012,71325002) 国家自然科学基金项目(61300195,61379041) 河北省自然科学基金项目(F2014501078,F2016501079)~~
【分类号】：TP393.01

【参考文献】