摘要
SDN(软件定义网络)核心技术是通过将网络设备控制层与数据转发层分离,实现网络流量的灵活控制。目前针对SDN网络架构的恶意应用研究还较少。针对这一问题,在总结分析现有恶意应用检测方法的基础上,基于深度学习技术提出一种面向SDN恶意应用的检测方法,将恶意应用转化为图片,在TensorFlow深度学习框架下对32个SDN恶意样本进行学习和检测,实验数据表明,该方法对恶意应用检测率可以达到89%,验证了方案的可行性。
The core technology SDN (software defined network) is to realize the flexible control of network traffic by separating the network device control layer from the data forwarding layer.At present,there are few researches on malicious application of SDN network architecture.To solve this problem,based on the analysis of existing malicious application detection methods,a detection method for SDN malicious application was proposed based on deep learning technology.The malicious application was transformed into a picture,and 32 SDN malicious samples were learned and detected under the TensorFlow depth learning framework.Experimental data show that the detection rate of malicious application can reach 89%,which verifies the feasibility of the proposed scheme.
作者
池亚平
余宇舟
杨建喜
CHI Ya-ping;YU Yu-zhou;YANG Jian-xi(Network Space Security Department,Beijing Electronics Science and Technology Institute,Beijing 100070,China)
出处
《计算机工程与设计》
北大核心
2019年第8期2134-2139,共6页
Computer Engineering and Design
基金
国家发改委信息安全专项基金项目(发改办高技[2015]289号)
国家863高技术研究发展计划基金项目(2015AA017202)
国家重点研发计划基金项目(2018YF1004101)