面向虚拟化网络环境的网络安全态势要素提取及安全态势预测应用

The Application of Network Security Situation Elements Extraction and Security Situation Prediction in Virtual Network Environment

计算机与网络技术的发展,使得大部分行业已经进入深度的网络融合应用阶段。计算机网络的高性能和高效率在为人类社会带来福利的同时,也因网络安全问题给社会运转带来威胁。特别是随着诸如网络功能虚拟化(Network Functions Virtualization,NFV)、软件定义网络(Software Defined Network,SDN)等技术得以实施和应用的今天,迥异于传统物理网络结构的虚拟化网络,使得人们在应对网络安全问题时面临新的挑战,特别是在网络安全态势预测方面。基于此,针对NFV与SDN高度融合环境下的网络安全态势预测方案进行研究和探讨,构建编码网络模型,对该环境下的网络安全态势要素进行提取,并利用基于注意力机制的循环门控单元预测方法,建立网络模型,以实现网络安全态势预测。该方案弥补了传统网络安全态势预测方案在准确度、适应性等方面的不足,期望能够为网络安全态势预测的相关研究与应用提供帮助。

With the development of computer and network technology, most industries have entered the stage of deep network integration application. The high performance and high efficiency of computer network not only bring benefits to human society, but also threaten social operation due to network security problems. Especially with the implementation and application of technologies such as Network Functions Virtualization(NFV), Software Defined Network(SDN), etc. Today, the virtualized network, which is different from the traditional physical network structure, makes people face new challenges when dealing with network security problems, especially in network security situation prediction. In this paper, the network security situation prediction scheme under the highly integrated environment of NFV and SDN is studied and discussed. The network security situation elements under this environment are extracted by building a coding network model, and the network model is established by using the circular gating unit prediction method based on attention mechanism to achieve network security situation prediction. This scheme makes up for the shortcomings of traditional network security situation prediction schemes in accuracy, adaptability, etc. It is expected that the research in this paper can help the related research and application of network security situation prediction.