Proposing a Hybrid RPL Protocol for Rank and Wormhole Attack Mitigation using Machine Learning

Internet of Things have profoundly transformed the way technology is deployed today in different domains of life. However, its widescale implementation has also caused major security concerns in context of data communication because of escalating interconnectivity of resource-constrained smart devices. Due to the exacerbating security attack vulnerability, it has become necessary to address the issue of insecure routing in these devices. Low-power and lossy IoT networks on which they run commonly use RPL for routing due to its lightweight nature and compatibility for data transmission. However, RPL is prone to both WSN-inherited and RPL-specific attacks. Several existing solutions have addressed the detection of some of them. However, lack of mitigation techniques is observed which can extenuate attacks of both types such as wormhole as well as rank attack; when they are launched on an RPL-based network. Therefore, the aim of this study is to introduce RPL, its vulnerability to the two attacks, and the proposition that machine learning techniques like support vector machines can be effectively used to develop a secure and improved version of RPL for mitigation of both WSN-inherited and RPL-specific attacks in an RPL-based IoT network.