5G及未来移动通信网络虚拟化的场景下,资源的管理与编排是实现虚拟化演进分组核心网(vEPC)业务高效部署的关键问题.vEPC中业务功能的载体是服务功能链(SFC),不同服务功能链的控制信令与多媒体数据流呈现负载差异化特点.该文针对传统算法资源定比分配所导致的负载不均问题,提出服务功能链的控制层和转发层解耦的映射模型,解耦后不同层的服务链可以独立地扩容和缩容,实现资源精准按需切片.拓扑映射阶段引入图相似理论,提出一种基于子图同构的虚拟网络映射算法VF2-H,依据节点的全局资源度优选映射子网,通过图特征分析设置剪枝条件,采用回溯机制进行启发式搜索.仿真结果表明,该算法在租户请求接收率、长期收益开销比等方面均具有良好的性能.%In 5G and the future mobile communication network, resource management and scheduling are the key issues to achieve efficient service deployment of virtual Evolved Packet Core (vEPC) nerwork. Service deployment in vEPC is based on Service Function Chain (SFC), in which signaling streams and forwarding streams have a big difference. On account of traffic differentiation of mobile network, the proposed model decouples the control layer and transfer layer of SFC. Different layers can make expansion and contraction independently to achieve accurate resources on-demand slice. Utilizing graph similarity theory, a virtual network embedding strategy called VF2-H is put forward in accordance with subgraph isomorphism. Firstly, candidate substrate subnet is preliminary selected on the basis of global resources capacity. Secondly, pruning condition is formulated based on the graph characteristics. Finally, collaborative search strategy is designed according to the characters of vEPC mapping. The simulation results validate the performance of the proposed algorithm in request accepting rate and long-term revenue-to-cost rate.
展开▼
