A green proactive orchestration architecture for cloud resources

摘要

Intrinsically, cloud computing is the paradigm that is been envisioned to dynamically furnish endless computing services to pervasive worldwide clients. However, the massive energy consumption resulted from the underpinning data centers would drastically hinder the full potential from the dreamed paradigm. Accordingly, this work opts towards devising green proactive cloud resources orchestration architecture. Specifically, this work focuses on cloud-based web applications which considered one of the most paramount cloud service categories. The proposed framework hence geared for leveraging cloud system resources utilization, ensuring explicit guarantees on negotiated web services' obligations, meanwhile achieving power consumption minimization objectives. The proposed framework consolidates heuristic methodologies along with control theory approaches in a green resources orchestration hierarchical framework. Firstly, an energy-aware global controller is presented.The proposed global controller employs an online graph coloring-based approach for mapping the submitted virtual machines (VMs) to the appropriate cloud resources. The proposed global controller seeks energy-efficient configurations for the submitted VMs in a constrained environment without obligating the negotiated services SLAs. Secondly, a proactive fuzzy-logic-based local controller is proposed. The proposed local controller aimed at in confronting workloads' sustainable fluctuations via proactive amendment for the placement and provisioning schedules. Additionally, the proposed local controller oriented towards maintaining active power management policy, especially during transient peak of usage, thereby mitigating overall costs, and extending resources capacity and performance capabilities. Conducted simulations and comparisons highly demonstrate the effectiveness of the proposed architecture at outperforming counterpart approaches specifically in terms of scalability, total energy consumption, and maintaining the web applications SLAs objectives in spite of workload fluctuations.