An Exploration Into Secure IoT Networks Using Deep Learning Methodologies

With Internet of Things (IoT) emerging as a new technology connecting few devices at home to billions of devices at corporate, the prices of such connected devices have become significantly low over the years and therefore more prone to attacks like Denial of Service (DoS) and Distributed Denial of Service (DDoS). To enable resilience-continuous monitoring is needed alongside adaptive decision making. To address these challenges, Software Defined Networks (SDN) can help handling security threats in IoT networks dynamically and in an adaptive manner. Various research work has been done in the field of Cyber Security and the Deep learning methodologies to apply the techniques to secure the network by watermarking [4]. Honeypots show great results in overcoming the threat of DDoS attacks and provide dynamic protection for SDNs, however anti-honeypot software can identify the honeypots and nullify their actions. Pseudo-Honeypots are proposed to protect from such anti-honeypot attacks by proving several groups of Bayesian-Nash Equilibrium (BNE) as a part of the strategy. SDNs have been used in many of the proposed techniques that relies on the Long Short-Term Memory (LSTM) models for predicting attacks or anomalies in the traffic data or even classify malicious IoT nodes.