Scholarlink

<p>Recent developments in the design of cyber-physical systems (CPS) have shifted towards the use of heterogeneous multi-core architectures integrated with cloud computing capabilities. Virtualization with resource sharing is a key element of cloud computing. The growing demand for cloud services has further emphasized the challenge of optimal resource allocation. In this paper, a network-on-chip-based approach is proposed, which optimizes the deployment and migration of virtual machines in cloud-supported CPS by combining reinforcement learning and metaheuristic algorithms. The proposed framework significantly reduces unnecessary migrations and enhances system stability, playing a key role in drastically reducing energy consumption. Evaluation results show that the proposed approach improves energy efficiency by approximately 2–3% under high-load conditions, compared to the studied methods. This improvement highlights the effectiveness of our approach in optimizing resource allocation and reducing energy consumption in cloud-supported CPS environment. This achievement makes the proposed framework a powerful option for enhancing efficiency and reducing operational costs in cloud environments. </p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The management of energy-aware virtual machine deployment and load balancing using mapping in network-on-chip for cyber-physical systems in cloud computing

  • 期刊论文

  • 发表日期：

    2025年9月15日

查看全文
  • Alireza Mahjoub,
  • Madjid Khalilian,
  • Javad Mohammadzadeh

摘要

Recent developments in the design of cyber-physical systems (CPS) have shifted towards the use of heterogeneous multi-core architectures integrated with cloud computing capabilities. Virtualization with resource sharing is a key element of cloud computing. The growing demand for cloud services has further emphasized the challenge of optimal resource allocation. In this paper, a network-on-chip-based approach is proposed, which optimizes the deployment and migration of virtual machines in cloud-supported CPS by combining reinforcement learning and metaheuristic algorithms. The proposed framework significantly reduces unnecessary migrations and enhances system stability, playing a key role in drastically reducing energy consumption. Evaluation results show that the proposed approach improves energy efficiency by approximately 2–3% under high-load conditions, compared to the studied methods. This improvement highlights the effectiveness of our approach in optimizing resource allocation and reducing energy consumption in cloud-supported CPS environment. This achievement makes the proposed framework a powerful option for enhancing efficiency and reducing operational costs in cloud environments.