Scholarlink

<p>In this research, an optimised deep learning-based energy-efficient method for migration of data in a heterogeneous cloud is presented. Virtual machines (VM), Containers, and Physical Machines (PM) are utilised in cloud simulations.The data migration is facilitates by a presented approach known as the adaptive dragonfly optimization (ADrO) algorithm, which employs a Levy flight strategy to prevent local optima trapping and maintain population variation.In the meantime, an algorithm called the actor critic neural network (ACNN) predicts the load. Additionally, predicted load, transmission cost, demand, resource capacity, agility, reputation, migration time, and energy consumption are all taken into account as objective functions. Experimental results show that the performance of data migration method outperforms state-of-the-art approaches in key performance metrics: memory usage, energy consumption, migration cost, and migration time of 475&#xa0;mb, 9367&#xa0;mj, 38.1245$, and 14,412&#xa0;ms respectively.</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

Multi-objective data migration in container-based heterogeneous cloud environments using deep adaptive dragonfly optimization

  • 期刊论文

  • 发表日期：

    2025年5月9日

查看全文
  • S. Durgaprasad,
  • Arif Mohammad Abdul

摘要

In this research, an optimised deep learning-based energy-efficient method for migration of data in a heterogeneous cloud is presented. Virtual machines (VM), Containers, and Physical Machines (PM) are utilised in cloud simulations.The data migration is facilitates by a presented approach known as the adaptive dragonfly optimization (ADrO) algorithm, which employs a Levy flight strategy to prevent local optima trapping and maintain population variation.In the meantime, an algorithm called the actor critic neural network (ACNN) predicts the load. Additionally, predicted load, transmission cost, demand, resource capacity, agility, reputation, migration time, and energy consumption are all taken into account as objective functions. Experimental results show that the performance of data migration method outperforms state-of-the-art approaches in key performance metrics: memory usage, energy consumption, migration cost, and migration time of 475 mb, 9367 mj, 38.1245$, and 14,412 ms respectively.