Optimal Resource Allocation and Task Scheduling in Fog Computing for Internet of Medical Things Applications

Abstract

Fog computing evolved in 2012 and extended conventional cloud computing services to the Internet of Things (IoT) applications. Real-time applications require fast response to satisfy their quality-of-service requirements. However, cloud computing generates communication latency, which is unacceptable for real-time applications. Fog computing eliminates latency sensitivity by providing services at the edge to IoT users. However, the number of IoT users is increasing exponentially; thus, tasks are generated dynamically and stochastically. Fog computing is a resource-constrained paradigm, unlike the cloud; therefore, adequate resource utilization and task scheduling are challenging. This article proposes a novel framework for Internet of Medical Things (IoMT) applications based on load balancing and task scheduling to minimize overhead latency. To realize the proposed framework, we implement a modified particle swarm optimization (MPSO) technique for delay-sensitive IoMT applications. The proposed algorithm is implemented and evaluated using the iFogSim modeling toolkit. The evaluation is based on performance metrics of execution time delay, execution cost, energy consumption, and network bandwidth consumption as utility functions. Experimental results based on the proposed technique show significant improvements in the performance of IoMT applications (up to 20%, 30%, and 15% in terms of delay, cost, energy, and network, respectively), compared with their counterparts. Moreover, the proposed technique based on MPSO improves resource utilization by up to 80%.