摘要
传统编排算法分配网络功能时,对功能链所需资源、端到端时延的约束能力较差,导致功能编排资源开销较大、节点资源利用率和网络服务接收率较低。为此提出基于人工智能的虚拟化网络功能动态编排算法。将底层物理网络看作一个网络拓扑图,把网络功能放置到物理节点上,计算资源容量、功能链时延的约束条件和目标函数。通过人工智能技术训练出目标函数最优解,得到网络功能在物理链路的最优编排方案。最后将物理链路映射到虚拟化链路中,完成虚拟化网络功能编排。实验结果表明,上述算法相比传统算法,降低了网络功能编排的总体开销,能够充分保证网络负载均衡性,同时缩短了服务响应时间,提高了网络服务质量。
Generally, the traditional scheduling algorithm has the disadvantages of large functional scheduling resource overhead, low node resource utilization and network service reception rate. Therefore, this paper proposes a dynamic scheduling algorithm of virtual network function based on artificial intelligence. Firstly, the underlying physical network was regarded as a network topology, and the network functions were placed on the physical nodes to calculate the constraints and objective functions of resource capacity, function chain delay. Secondly, according to the artificial intelligence technology, the optimal solution of the objective function and the optimal arrangement scheme of the network function in the physical link were obtained. Finally, the physical link was mapped to the virtual link, thus achieving the virtual network function choreography. The results show that the algorithm is superior to the traditional algorithm, with low network function scheduling overhead, balanced network load and excellent network quality of service.
作者
刘世芳
孙艳秋
张柯欣
LIU Shi-fang;SUN Yan-qiu;ZHANG Ke-xin(School of Information Engineering,Liaoning University of Traditional Chinese Medicine,Shenyang Liaoning 110847,China)
出处
《计算机仿真》
北大核心
2021年第7期345-349,共5页
Computer Simulation
关键词
人工智能
虚拟化网络
服务功能链
编排算法
资源开销
Artificial intelligence
Virtualization network
Service function chain
Scheduling algorithm
Resource overhead
