An Energy-Aware Multistages Hybrid Scheduling Approach for IoT Workflow Applications With Reliability Constraint in Cloud Computing Systems

With the rapid advancement of cloud computing, cloud services have been widely adopted for managing large-scale and complex IoT workflow applications due to their robust computational capabilities. However, efficiently scheduling and deploying these workflows while ensuring Quality-of-Service (QoS) for diverse users remains a significant challenge for cloud service providers. In this study, we propose a novel multistage workflow scheduling algorithm (RE-ACO) for energy-efficient management of reliability-constrained IoT applications in cloud environments. The algorithm operates in three key stages: 1) task ordering by ant colony optimization (ACO); 2) reliability constraint distribution with feedback information; and 3) energy-aware task assignment. The RE-ACO leverages ACO and an energy-aware task assignment strategy to optimize energy usage without compromising workflow reliability. First, the ACO algorithm determines the optimal task execution sequence. Next, a feedback-based reliability distribution method dynamically assigns subreliability constraints to individual tasks. Finally, each task is allocated to a virtual machine (VM) that minimizes energy consumption while meeting its subreliability requirement. Simulation results demonstrate that RE-ACO outperforms existing approaches, achieving the lowest energy consumption for reliability-constrained workflow scheduling compared to three benchmark algorithms.