Scholarlink

<p>The sudden surge in cloud services has led to enormous energy overheads and necessitated environmentally friendly, reconfigurable cloud infrastructures. Conventional energy conserving methods fail to optimize performance and use of resources for changing workloads. In this paper, an Energy-Adaptive Hierarchical Scheduling Architecture (EAHSA) is presented that adapts resource allocation dynamically in green cloud environments using a Thermocline Predictive Scheduler (TPS) and a bio-inspired Slime Mold Optimization Algorithm (SMA). The proposed design simulates the oceanic thermocline concept to predict workload spikes from past load gradients, facilitating anticipatory scaling and dynamic transfer between hierarchical compute nodes. TPS serves as the temporal prediction module, approximating energy-flow stratification levels based on time-dependent workload entropy. At the same time, SMA simulates the decentralized slime mold optimization behavior to find low-energy, high-throughput task paths of routing in a reconfigurable dynamic topology. The design employs a layered control logic comprising thermal-aware node profiling, prognostic entropy calculation, and swarm-based energy equalization. Empirical analysis with synthetic cloud traces and actual workloads (Google cluster data) shows up to 27.3% less energy consumption, 19.8% reduced SLA violation, and 22.5% quicker task convergence than baseline schedulers such as Round-Robin, Power-Aware Best-Fit, and Dynamic Voltage-Frequency Scaling (DVFS). Its hybrid predictive-optimization paradigm also facilitates adaptability in edge-centric fog-cloud environments. In general, EAHSA stands as a scalable and energy-aware scheduling paradigm that is appropriate for next-generation green cloud platforms.</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

Energy-adaptive hierarchical scheduling architecture (EAHSA) for reconfigurable green cloud systems

  • 期刊论文

  • 发表日期：

    2025年12月19日

查看全文
  • K. Punitha,
  • V. Thanikachalam,
  • S. Poornima

摘要

The sudden surge in cloud services has led to enormous energy overheads and necessitated environmentally friendly, reconfigurable cloud infrastructures. Conventional energy conserving methods fail to optimize performance and use of resources for changing workloads. In this paper, an Energy-Adaptive Hierarchical Scheduling Architecture (EAHSA) is presented that adapts resource allocation dynamically in green cloud environments using a Thermocline Predictive Scheduler (TPS) and a bio-inspired Slime Mold Optimization Algorithm (SMA). The proposed design simulates the oceanic thermocline concept to predict workload spikes from past load gradients, facilitating anticipatory scaling and dynamic transfer between hierarchical compute nodes. TPS serves as the temporal prediction module, approximating energy-flow stratification levels based on time-dependent workload entropy. At the same time, SMA simulates the decentralized slime mold optimization behavior to find low-energy, high-throughput task paths of routing in a reconfigurable dynamic topology. The design employs a layered control logic comprising thermal-aware node profiling, prognostic entropy calculation, and swarm-based energy equalization. Empirical analysis with synthetic cloud traces and actual workloads (Google cluster data) shows up to 27.3% less energy consumption, 19.8% reduced SLA violation, and 22.5% quicker task convergence than baseline schedulers such as Round-Robin, Power-Aware Best-Fit, and Dynamic Voltage-Frequency Scaling (DVFS). Its hybrid predictive-optimization paradigm also facilitates adaptability in edge-centric fog-cloud environments. In general, EAHSA stands as a scalable and energy-aware scheduling paradigm that is appropriate for next-generation green cloud platforms.