基于神经网络的NDN入侵检测方法

NDN Intrusion Detection Method based on Neural Network

为了从根本上解决现有互联网络中存在的问题,人们对未来网络架构进行了深入研究,其中命名数据网络(Named Data Networking,NDN)在可扩展性、安全性、数据分发等方面具有极大优势,得到了学术界和产业界的广泛重视,但同时也带来了新的安全隐患。一些新的攻击方式对NDN危害巨大,包括缓存污染攻击、兴趣泛洪攻击和内容污染攻击等。因特网服务提供商(Internet Service Provider,ISP)在实际部署NDN时,必然需要一个统一、智能、可扩展的入侵检测方法来为网络的安全性提供支持。BP神经网络具有一定的自适应与自组织能力,在分类预测方面具备良好的性能。因此,基于BP神经网络,设计并实现了命名数据网络的入侵检测方法。仿真结果表明,该方法对网络攻击的分类准确性和辨识效率较高。

In order to fundamentally solve the problems existing in the present Internet,people have conducted in-depth research on future network architectures.Among these architectures,NDN(Named Data Networking)has great advantages in terms of scalability,security,and data distribution,and has received extensive attention from academics and industry,and at the same time,it also brings new security risks.Some new attack methods are extremely harmful to NDN,including cache pollution attacks,interest flood attacks,and content pollution attacks.An ISP(Internet Service Provider)inevitably requires a unified,intelligent,and scalable intrusion detection method to provide support for network security in his actual deployment of NDN.BP neural network has certain self-adaptation and self-organization ability,and has good performance in classification prediction.Therefore,based on BP neural network,an intrusion detection method of named data network is designed and implemented.The simulation results indicate that this method has fairly high classification accuracy and identification efficiency for network attacks.