A fault tolerance aware green IoT workflow scheduling algorithm for multi-dimensional resource utilization in sustainable cloud computing

Abstract

In recent times, the distributed cloud computing landscape has witnessed a remarkable surge in processing vast amounts of data and the crucial need to maintain service level agreements (SLAs) between providers and consumers. In this dynamic environment, cloud resources are inherently multidimensional, encompassing computing machines and communication connections susceptible to failures and energy-related considerations. However, given two-objective energy and reliability optimization, existing time allocation policies focus primarily on optimizing a single intent, which leads to system degradation in environments that generate problematic constraints from executable processing units. To handle the shortcomings of the application placement policies, we suggest a solution in the form of a three-phase bi-objective workflow scheduling issue called (Bi-OWSP). This three-phase approach aims to optimize workflow scheduling by simultaneously considering energy and reliability as two vital objectives. To achieve this, we employ two distinct algorithms. First is the stepwise dynamic voltage and frequency scale algorithm ensures energy efficiency by calculating optimal frequencies, reducing energy usage, and minimizing task mapping time. The second algorithm is the reliability-conscious heterogeneous fault tolerance approach, which emphasizes avoiding high-deficit servers and communication links to enhance system reliability. Furthermore, we introduce an energy-aware stepwise reliability maximization algorithm, which intelligently selects the best combination of task-server pairs to achieve energy minimization and reliability maximization. Through extensive simulation experiments on artificial and real-world workflow applications, we demonstrate the significance of Bi-OWSP in providing superior energy-reliability compensation solutions compared to competing algorithms.