LESP:A fault-aware internet of things service placement in fog computing

Abstract

The rapid advancement of 5G networks enables increase adoption of Industrial Internet of Things (IIoT) devices which introduces variety of time-sensitive applications requires low-latency, fault-tolerant, and energy-efficient computing environments. Fog computing infrastructure that extends cloud computing capabilities at the network edge to provide computation, communication, and storage resources. Due to the limited computing capacity of the Fog node, it restricts the number of tasks executed. The other key challenges are the risk of hardware and software failure during task execution. These failures tend to disrupt the configuration of fog computing nodes, affecting the reliability and availability of services. As a result, this can negatively impact the overall performance and service level objectives. The fault-tolerant-based IoT service placement problem in the fog computing environment primarily focuses on optimal placement of IoT services on fog and cloud resources with the objective of maximizing fault tolerance while satisfying network and storage capacity constraints. In this study, we compared different community-based techniques Girvan-Newman and Louvain with Integer Linear Programming (ILP) for fault tolerance in fog computing using the Albert-Barabási network model. In addition, it proposed a novel Louvian based on eigenvector centrality service placement (LESP) to improve conventional Louvian methods. The proposed algorithm is simulated in iFogSim2 simulator under three different scenario such as under 100, 200 and 300 nodes. The simulation results exemplify that LESP improves fault tolerance and energy efficiency, with an average improvement of approximately 20% over Girvan-Newman, 25% over ILP, and 12.33% over Louvain. These improvements underscore LESP’s strong efficiency and capability in improving service availability across a wide range of network configurations.