Replication-Based Resource Provisioning and Constrained Aware Task Scheduling Framework for Cloud Workflows

Cloud computing (CC) is a popular distributed computing technology that is virtualized, scalable, and provides resources and services on demand through subscription-based models. Cloud Service Providers (CSPs) deliver these services, primarily catering to a vast range of business and scientific applications. CC is used to evaluate large-scale scientific applications, often implemented as scientific workflows (SWs). This paper highlights the wide range of applications using CC, including image processing, data mining, and healthcare systems, all of which require efficient and accessible cloud services. These applications often have strict deadlines and consume significant energy, emphasizing the need for low-energy resource utilization. The paper proposes a Replication-based Resource Provisioning and Constrained-aware Task Scheduling (R2PCT) Framework for Cloud Workflows to address these challenges. The framework aims to manage resources efficiently in a multilevel environment and schedule tasks using replication-based fault-tolerant techniques while considering deadline and budget constraints. To evaluate its effectiveness, simulations were conducted on WorkflowSim for Montage and CyberShake workflows. The proposed R2PCT strategy performs better than the existing state-of-the-art strategy, reducing execution time by an average of 52%, minimizing execution cost by 34%, and decreasing energy consumption by 35% compared to the existing R-DEAR strategy for Montage SWs. For CyberShake workflows, the proposed R2PCT strategy performs better than existing methods, reducing execution time by an average of 59%, minimizing execution cost by 38%, and decreasing energy consumption by 60% compared to the existing R-DEAR strategy.