A Practical Self-Stabilizing Leader Election for Networks of Resource-Constrained IoT Devices

This paper presents a self-stabilizing and tunable leader election algorithm, called PraSLE, designed for real-world testbeds of resource-constrained devices in the Internet of Things (IoT). PraSLE is an extended version of the minimum finding algorithm that functions in a round-based asynchronous fashion. We provide a version of PraSLE for both reliable and unreliable networks. We show that PraSLE elects a unique leader in unreliable networks with probability 1. We also experimentally validate this result and demonstrate that PraSLE outperforms existing methods in terms of average convergence time for ring, line, and mesh topologies up to 40 nodes, and for the clique topology up to 80 nodes. The tunable nature of PraSLE enables engineers to optimize average convergence time, as well as communication and energy costs. PraSLE provides an important resource-efficient Distributed Computing Primitive (DCP) for unreliable networks of low-end IoT devices based on which other DCPs (e.g., Paxos) can be developed for such networks in the domains of IoT and Cyber Physical Systems (CPS).