Impact Analysis of Rank Attack on RPL-Based 6LoWPAN Networks in Internet of Things and Aftermaths

Routing protocol for low power and lossy networks (RPL) serves the routing process in Internet Protocol version 6 over low-power wireless personal area network protocol stack. RPL protocol is deployed on sensor nodes and is vulnerable to many attacks due to the resource-constrained and non-tamper-resistant nature of these nodes. Many attacks on RPL-based networks can be carried out by an attacker using either an insider or an outsider attack method. The concerns associated with the RPL protocol regarding security and privacy may restrict its global adaption. Therefore, it is necessary to conduct a thorough investigation of RPL-specific attacks and their effects on the underlying network. To this end, our work provides a detailed investigation of the rank attack, a type of denial-of-service attack and its effect on the RPL network. We made an in-depth experimental study on four different RPL network topologies and analysed the impact of rank attack. To analyse the effect of rank attack, we use both grid and random variants of the network topologies. Based on the location of the sink node, we classify the grid topology as grid corner and grid centre, whereas, based on mobility, we classify the random topology as random topology with static nodes and random topology with mobile nodes. Under the attack scenario, the average number of packets received in grid topology with sink node at centre degrades by 10%. For grid topology having the sink node at the corner, the average number of packets received degraded by 13.44%. The percentage degradation of the average number of packets received at the sink node with random topology is nearly 8%. Random topology with mobile nodes has 7.4% lesser number of packets received at the sink node. Next, the value of average inter-packet time with random mobile topology has the highest increase of about 19% under attack. At the same time, we notice an increase in average inter-packet time of about 6.4% with random topology, 3.7% with grid topology having the sink node at the corner, and 0.7% with grid topology having the sink node at the centre. Further, when the rank attack is mounted, we notice a hike of 24% in message count with random mobile and 28% with grid corner topology. Moreover, both random topology and grid topology with sink at the centre have a hike of about 13% in message count. Furthermore, when average power consumption is considered under attack scenario, random topology with mobile nodes and grid topology with the sink node placed at the corner demonstrate an increase of 11% under attack. At the same time, random and grid topologies with the sink node at the centre have a rise of nearly 2% as compared to normal RPL. The results of our experiments show that the rank attack considerably degrades network performance in terms of packets received at the sink node, inter-packet time, control message count, and average power consumption. Further, our investigation also confirms that the position of the attacker node in the network topology has a significant impact on the effect of the rank attack. Furthermore, we also suggest some potential solutions to mitigate this attack.