基于移动设备云迁移的节能决策算法

随着云计算的兴起,云迁移计算开始成为移动设备获取计算资源和降低功耗的有效方式.云迁移的主要想法是将移动终端的复杂任务经由无线网络迁移到云端执行,然后再接收计算结果.然而,无线网络的不稳定性和数据传输的高功耗限制了云迁移计算在移动设备中的应用.不同于已有工作,本文通过引入数据压缩的方法完善了云迁移计算决策模型,并且基于对未来时段网络期望的预测,提出了一种节能迁移计算决策算法——EPVAD.基于实际的3G网络带宽数据和开发测试平台,实验结果显示:EPVAD算法的节能效果较同类算法平均优14.9%,并且算法自身的系统开销可忽略.

An energy-efficient cloud-based offloading decision algorithm for mobile devices

With the advent of cloud computing,cloud-based offloading computing becomes an efficient method for the mobile device to enhance computing capability and to reduce energy consumption.However,the unreliability and the high data-transfer energy consumption of the wireless network affects its efficiency in mobile devices. Different from previous researches,we introduce data compression to the existing offloading decision model and present an efficient prediction method for the future network bandwidth expectation.Based on this prediction method,we present an energy-efficient offloading decision algorithm-EPVAD.Based on the practical 3G network bandwidth traces,we demonstrate the effiectiveness of the proposed algorithm through simulations and realistic platform evaluation.We show that the EPVAD algorithm can achieve averagely 14.9%of more energy reduction compared with the existing algorithm while its cost can be negligible.