To Secure IoT sensor nodes through Fog computing

Abstract

IoT is a network of smart devices with different technology that performs tasks assigned to it automatically with the help of the internet and enables people to access the devices from anywhere at any time. In several real-time applications of IoT, it provides efficient utilization of resources with less human intervention, and because of its flexibility, its usage is rapidly increasing and hence these cloud-based IoT devices are prone to many security threats. To mitigate threats and secure IoT nodes the fog computing similar to a cloud platform that is highly virtualized and decentralized which is an intermediate layer between cloud computing and IoT devices is used. Fog computing also reduces the latency issues of accessing the IoT node's data as it is located at the edge of the network. The fog computing is essential for IoT because of its characteristics that add benefits to IoT and as we know IoT devices are of resource constrained and have low computation power building an effective IDS for IoT nodes are essential such that the malicious IoT nodes are detected and isolated. To secure the IoT nodes the machine learning technique is deployed into the fog layer to detect the attacks and reduce the burden for IoT nodes as they have less computation power. The Kddcup99 dataset which contains data required to be audited and a wide variety of intrusions is trained with the machine learning algorithm that gives a satisfactory performance of detecting attacks with high accuracy is deployed in fog computing. Then the cooperative node mechanism is used to detect the malicious nodes in IoT which are caused due to insiders attack, other kinds of attacks, and such malicious nodes are isolated by the fog layer based on the alert sent by the common neighbor IoT node which detects the malicious IoT node.