基于粒子群优化的负载均衡的虚拟网络映射

发布时间：2018-03-25 02:13

本文选题：云计算　切入点：虚拟网映射　出处：《电子与信息学报》2013年07期

【摘要】：虚拟网络映射忽略通信路径上中间节点的资源消耗,以底层物理网络资源消耗最小或路径最短为目标,导致最优路径上中间节点资源不足而出现瓶颈,影响整个底层网络的性能及后续虚拟网络请求成功率。该文针对这一问题考虑物理路径上中间节点资源消耗,以节点负载和链路负载同时达到均衡为目标,将路径跳数限制作为约束,建立虚拟网络映射的数学模型,并采用多目标负载均衡粒子群优化算法求解,记作LB-PSO。实验证明该算法有效地消除了资源瓶颈,为后续虚拟网络请求提供一个更为均衡的底层物理网络,从而提高了虚拟网络构建成功率、网络资源利用率以及基础设施提供商的收益。
[Abstract]:Virtual network mapping neglects the resource consumption of the intermediate nodes on the communication path, aiming at the minimum resource consumption or the shortest path of the underlying physical network, which leads to the shortage of the intermediate node resources on the optimal path. The performance of the whole underlying network and the success rate of subsequent virtual network requests are affected. In view of this problem, the resource consumption of the intermediate node in the physical path is considered. The goal is to balance the node load and link load at the same time. A mathematical model of virtual network mapping is established with path hops restriction as constraint, and multi-objective load balancing particle swarm optimization algorithm is used to solve the problem, which is described as LB-PSO.Experimental results show that the algorithm effectively eliminates resource bottleneck. It provides a more balanced underlying physical network for subsequent virtual network requests, which improves the success rate of virtual network construction, the utilization of network resources and the benefits of infrastructure providers.
【作者单位】：北京邮电大学智能交换与控制国家重点实验室;
【基金】：国家自然科学基金(2009CB320406) 国家863计划项目(2011AA01A102) 国家创新研究群体科学基金(60821001)资助课题
【分类号】：TP393.01

【参考文献】