MapReduce实现平台的分析与比较
首发时间:2011-11-25
摘要:MapReduce是由Google公司提出的一种并行编程模型。近几年随着云计算技术的蓬勃发展,作为云计算核心技术之一的MapReduce也受到了较多的关注,获得了很大的发展。MapReduce技术具有简洁的模型、良好的扩展性、容错性和并行性,随着其性能的不断改进和分析能力的不断增强,采用MapReduce编程模型的平台也越来越多。本文首先介绍了传统的MapReduce编程模型的原理及核心思想,其次分析了Hadoop、Phoenix和Mars三种MapReduce的实现平台,并对其进行了系统架构及性能的对比,最后总结出各自的特点和适用范围。
关键词: 并行处理 MapReduce Hadoop Phoenix Mars
For information in English, please click here
the Analysis and Comparison between MapReduce Implementations
Abstract:In this paper MapReduce model was introduced.The MapReduce is a parallel programming model created by Google.In recent years as the development of cloud computing ,MapReduce,one of the core technology of cloud,also obtained very big improvement. MapReduce is a simple model with good scalability, and fault tolerance . Due to the improvement of its performance and the enhancement of its analysis ability ,there are more and more implementations which adopt this progremming model.This paper first introduced the traditional MapReduce programming model and the principle of the core ideas, then analyzes the realization of the three MapReduce platform:Hadoop, Phoenix and Mars .Their system structures are also introduced.In the end,by making a comparision between their mechanical properties of contrast, we summarized their characteristics and application scope.
Keywords: parallel process MapReduce Hadoop Phoenix Mars
基金:
论文图表:
引用
No.****
同行评议
勘误表
MapReduce实现平台的分析与比较
评论
全部评论0/1000