ABSTRACT

These days, fog-cloud based healthcare application partitioning techniques have been growing progressively. However, existing static fog-cloud based application partitioning methods are static and cannot adopt dynamic changes in the dynamic environment (e.g., where network and computing nodes have resource value variation) during the execution process. This study devises a Deep Neural Networks Energy Cost-Efficient Partitioning and Task Scheduling (DNNECTS) algorithm framework which consists of the following components: application partitioning, task sequencing, and scheduling. Experimental results show the suggested methods in terms of energy consumption and the applications' cost in the dynamic environment.

摘要

如今，基于雾云的医疗保健应用程序分区技术一直在逐步发展。然而，现有的基于静态雾云的应用划分方法是静态的，在执行过程中不能采用动态环境（例如，网络和计算节点具有资源价值变化的情况）的动态变化。本研究设计了一个深度神经网络能源成本效益分区和任务调度 (DNNECTS) 算法框架，该框架由以下组件组成：应用程序分区、任务排序和调度。实验结果表明了所建议的方法在动态环境中的能耗和应用成本方面。