基于NFV的边缘流量分析增强物联网安全方法

Edge traffic analysis method based on NFV to protect IoT security

智能设备安全功能的减弱或缺失导致用户信息容易被攻击者非法窃取。针对传统的集中式IT安全措施在可扩展性和成本方面存在的局限性问题,设计了一种基于网络功能虚拟化(NFV)的边缘流量分析方法,该方法引入一种新的基于Docker的轻量级架构用于网络功能虚拟化,每一个网络功能都可以从云中获取并立即部署在IoT网关上。并通过特征提取、多数投票类型检测的方法,对3种模型中给定的数据点进行异常检测。实验结果表明基于Docker架构在存储空间、内存使用、延迟、吞吐量和可扩展性方面比标准化NFV体系结构性能更加优越,而且该方法能在1 s内以95%的准确率检测到攻击。

The weakening or lack of security functions of smart devices can be easily exploited by malicious users to steal user information. Traditional centralized IT security measures have limitations in terms of scalability and cost. Therefore an edge traffic analysis method based on NFV is proposed to protect the security of smart devices. This method introduces a new lightweight architecture based on Docker for network function virtualization, each network function can be fetched from cloud and deployed immediately onto the IoT gateway. Then propose a majority vote type detection method in three models to detect a given datapoint. Results show the advantages of the architecture over the standardized NFV architecture in terms of storage space, memory usage, latency, throughput and scalability. Finally, experiments show that the method can detect attacks with 95% accuracy within 1 second.