CSO-based Efficient Resource Management for Sustainable Cloud Computing

Abstract

The pervasive need for cloud-hosted application services has resulted from the widespread use of cloud data centers. Not only that, but there has been a dramatic increase in the resource demands of current applications, especially in data-intensive businesses. This has resulted in an increase in the number of cloud servers made available, which has increased energy usage and prompted environmental concerns. Only partially do the difficulties of scalability and adaptability in cloud resource management get addressed by conventional heuristics and reinforcement learning-based techniques. Many existing works overlook the interdependencies between host temperature, task resource usage, and scheduling decisions. Especially in contexts with fluctuating resource demands, this results in poor scalability and an upsurge in computing resource requirements. The study recommended a holistic resource management strategy based on resource scheduling for enduring cloud computing as a solution to these restrictions. Energy, thermal, and cooling models are all taken into account in the proposed model, which expresses the optimization of data center energy efficiency as a multi-objective scheduling issue. To generate optimal scheduling decisions and approximate the quality of service (QoS) for a given system state, the model employs cat-based swarm optimization as a surrogate model. Using the China Ocean Shipping Company (COSCO) framework, we conducted experiments that demonstrate cloud service orchestrator (CSO)’s superior performance compared to state-of-the-art baselines in terms of energy ingesting, makespan, and execution overhead in both simulated and real-world cloud environments.