摘要
与公有云计算相比,针对数据与计算双重密集型任务的私有云计算系统对计算效率和系统管理效率提出了更高的要求,目前的公有云计算系统显得过于复杂和繁琐,因此需要一种简便易用的能够适应数据与计算密集型任务的私有云计算系统实现。借鉴公有云计算的相关理论和实现方法,提出了一种针对数据与计算双重密集型任务的私有云计算系统实现方案。该方案通过作业文件描述用户的计算任务,确定计算任务的计算模型和计算的输入输出文件;针对私有云的特点,简化Google云计算系统的MapReduce并行处理框架,得到更加直观的数据计算模型;自动连接计算数据流,使该方案更加精简和适应处理数据与计算双重密集型任务。实验结果表明:该方案能够减少额外的计算消耗,处理速度能得到显著提升,有很高的实用性。
Compared with public cloud computing systems,private cloud computing systems aiming at data and computing dual intensive tasks have higher demand in computing and management efficiency.The realization methods for public cloud computing system are too complicated for users to develop.A simplify and easy to use realization of private cloud computing system is requirement.To meet this requirement,this paper proposed an approach to build a private cloud computing system which was able to adapt both data and computing intensive tasks on basis of public cloud computing systems implementation.This approach used job files with aim of describing computing tasks and determined the input and output files of computing model.Computing model of data processing could be reflected more intuitively by simplify the Google MapReduce parallel computing framework;the use of connecting computing flow automatically made the approach more streamline and rapidly to process intensive tasks.Experiment results shows that this approach can reduce extra computation overhead and improve processing efficiency significantly.This approach offers a high practical value.
出处
《计算机应用研究》
CSCD
北大核心
2011年第2期621-624,共4页
Application Research of Computers
基金
国家自然科学基金重点资助项目(40839905)
现代通信国家重点实验室基金资助项目(9140C1104040904)
关键词
数据与计算双重密集型任务
私有云
云计算
并行计算
data and computing intensive task
private cloud
cloud computing
parallel computing
