拓扑综合评估与权值自适应的虚拟网络映射算法

Virtual Network Embedding Algorithm Based on Topology Comprehensive Evaluation and Weight Adaptation

针对现有虚拟网络映射算法对节点拓扑特征考虑得不够全面、节点评价方式较为单一且指标权值不能根据网络环境自适应调整等问题,提出一种拓扑综合评估与权值自适应的虚拟网络映射算法。文中在节点映射阶段综合考虑节点中心度、就近度与邻近聚集度等拓扑属性,结合节点CPU与邻接带宽和等资源属性对节点进行多指标重要度排序,根据网络环境的变化利用熵权法自适应调整指标权值。仿真结果表明,相较于最新的和经典的虚拟网络映射算法,所提算法的映射成功率提高了2%~23%,长期平均收益开销比提升了3%~17%,且该算法对不同资源需求类型的虚拟网络请求都能保持良好性能。

The existing virtual network embedding algorithms do not consider the topological features of nodes comprehensively,the evaluation method of nodes is relative simple and the weights cannot be adaptively adjusted according to the network.To solve these problems,a virtual network embedding algorithm based on topology comprehensive evaluation and weight adaptation is proposed.In the node embedding stage,by considering the centrality,proximity and adjacent aggregation of nodes,this paper establishes a node multi-metric evaluation model combined with the node resource properties such as the node CPU and the sum of adjacent bandwidth.The weights are adjusted adaptively according to the change of network environment by using the entropy weight method.Simulation results show that compared with the latest and classical virtual network embedding algorithms,the acceptance ratio of the proposed algorithm is improved by 2%~23%,and the long-term average revenue-to-cost ratio is increased by 3%~17%.Moreover,the proposed algorithm can maintain good performance for different types of virtual network requests with different resource requirements.