面向移动设备的温度感知的任务调度算法

由于志愿者分布式计算可以为计算量庞大的科研项目提供足够的计算能力,甚至比超级计算机的计算能力还要强大,因此,志愿者分布式计算技术受到了很多研究人员的关注,很多不同的志愿者分布式计算架构被广泛应用。以往的很多志愿者分布式计算架构通常考虑的志愿者主机是PC电脑,或者单纯地把移动设备当作PC电脑一样进行处理。由于移动设备的很多特性跟PC电脑存在着很大的差异,所以很多时候这些志愿者分布式计算架构并不能高效地处理同时拥有PC电脑和移动设备志愿者的志愿计算项目。针对志愿者分布式计算系统上两个主流的志愿者分布式计算任务调度方法——迭代计算的任务调度算法和先来先服务的调度算法FCFS在处理移动设备志愿者计算上存在着的不足,为了提高志愿者分布式计算平台的执行效率,提出了一个面向移动设备的温度感知的任务调度算法TATSA。实验结果表明,TATSA比主流的任务调度算法ISA和FCFS在移动设备志愿者计算时效率明显更高。

A temperature aware task scheduling algorithm for mobile devices

Because the volunteer distributed computing can provide sufficient computing power for the research projects needing massive computing, it is even more powerful than supercomputing. As a result, the volunteer distributed computing technology has attracted a lot of attentions so that many architectures of different volunteer distributed computing are widely used. Most of those architectures usually consider the PC computers as volunteers, or simply treat the mobile devices as PC computers. Because many characteristics of mobile devices are very different from PC computers, those architectures cannot efficiently handle the volunteer computing projects that have both PC computers and mobile devices as volunteers. In order to solve the shortcomings of two popular task scheduling methods of the volunteer distributed computing: the iteration scheduling algorithm (ISA) and the first come first serve algorithm (FCFS)) in dealing with the calculation of mobile device volunteers, and improve the efficiency of the volunteer distributed computing platform with mobile devices volunteers, we propose a temperature aware task scheduling algorithm (TATSA). Experimental results show that the TATSA is more efficient than the two mainstream task scheduling algorithms ISA and FCFS in mobile device volunteer computing.