Automata-Based Dynamic Fault Tolerant Task Scheduling Approach in Fog Computing

Fog computing is an extension of cloud computing that offers computing, storage and communication resources near the network edge, which makes it an ideal platform for processing latency-sensitive and compute-intensive tasks. However, efficient task execution in a fog platform is a challenging problem since fog nodes are loosely interconnected, highly dynamic, heterogeneous, and prone to failures. Therefore, a task scheduling algorithm that ensures reliable execution of the tasks while optimising response time is paramount. To address this challenge, we propose a new Dynamic Fault Tolerant Learning Automata (DFTLA) task scheduling approach. DFTLA determines an efficient assignment of the tasks to the fog nodes based on variable-structure learning automata. We evaluated the proposed DFTLA scheduler and compared its performance with three baseline methods. The results of the experiments show that the proposed algorithm ensures reliable execution of the tasks while optimising response time and energy consumption. Moreover, the proposed approach outperforms the baseline algorithms in all performance evaluation criteria.