A Cluster-Based and Drop-aware Extension of RPL to Provide Reliability in IoT Applications

The standardized IPv6 Routing Protocol for Low-power and Lossy Networks (RPL) has enabled efficient communications between thousands of smart devices, sensors, and actuators in a bi-directional, and end-to-end manner, allowing the connection of resource constraint devices in multi-hop IoT infrastructures. RPL is designed to cope with the major challenges of Low-power and Lossy Networks (LLNs), specifically their energy-efficiency. However, RPL is facing with severe congestion and load balancing problems, leading to a low Packet Delivery Ratio (PDR) in the network. For the first time since the declaration of RPL, in this paper we explain that ignoring the specifications of the reception and transmission buffers in heterogeneous networks has caused these unbalanced traffic loads, leading to congestion, and consequently loss of packets in RPL. In order to resolve this problem, this paper introduces CBR-RPL; a light-weight RPL-based routing mechanism, which organizes the nodes into logical clusters and routes the packets through a novel drop-aware Objective Function (OF). The newly defined OF considers the queue occupancy of the nodes transceivers along with their drop rate simultaneously. According to an extensive set of experiments, which have been conducted via the Cooja simulator, it has been observed that the CBR-RPL improves the reliability in terms of PDR by 38.2%, and 75% compared to RPL and QURPL, respectively. In addition, CBR-RPL has also improved the amount of energy consumption in the nodes by up to $3\times$ compared to the state-of-the-art, mainly due to imposing fewer control packets to the network.