基于软件定义网络技术实现人工智能网络体系架构

An artificial intelligence network architecture based on software defined network

软件定义网络(Software Defined Network,SDN)拥有着集中控制、分布式路由和控制平面与转发平面相分离的特点,相比传统的TCP/IP网络,提高了网络的灵活性,并提供了网络可编程的能力。假设控制器能够获取底层网络的特征,并加以分析应用到网络管理中,将会有很好的应用前景。鉴于此,提出了一种基于SDN的人工智能网络架构。该系统使用SDN控制器从底层网络中获取到大量的各类信息,并利用大数据以及人工智能技术加以分析,得到底层网络的特征数据。而这些特征数据会被用来辅助控制器对于下层网络的管理的决策。

Software Defined Network （SDN）, which is centralized control and distributed routing and decouples the control plane and the forward plane, compared with traditional TCP/IP network, is more flexible and makes the network programmable. It has good prospects for application if the features of the bottom network were obtained by the controller and applied in the management of network. Thus, in this paper, an artificial intelligence network architecture based on software defined network was proposed. The system obtains enormous information of the bottom network by the means of SDN controller, and analysis them with the big data and artificial intelligence technology to extract the features of the bottom network. These features will be used to aid the controller for the management of network.