本文已被:浏览 3234次 下载 6895次
Received:September 06, 2013 Revised:October 12, 2013
Received:September 06, 2013 Revised:October 12, 2013
中文摘要: 经过几年的发展,并行编程模型MapReduce产生了若干个改进框架,它们都是针对传统MapReduce的不足进行的修正或重写. 本文阐述和分析了这些研究成果,包括: 以HaLoop为代表的迭代计算框架、以Twitter Storm为代表的实时计算框架、以Apache Hama为代表的图计算框架以及以Apache YARN为代表的框架管理平台. 这些专用系统在大数据领域发挥着越来越重要的作用.
Abstract:With the rapid development of recent years, some improved framework of MapReduce parallel programming model appeared. They are correction and recoding against lack of MRv1. This paper describes and analyzes this research achievements, including iterative computing framework as represented by HaLoop, real-time computing framework as represented by Twitter Storm, graph computing framework as represented by Apache Hama, computing resources negotiation platform as represented by Apache YARN. These special systems play a vital role in BigData fields.
keywords: MapReduce Hadoop parallel computing big data processing
文章编号: 中图分类号: 文献标志码:
基金项目:江苏省卓越工程师(软件类)计划试点专业(苏教高函[2012]17号)
引用文本:
应毅,刘亚军.MapReduce并行计算技术发展综述.计算机系统应用,2014,23(4):1-6,11
YING Yi,LIU Ya-Jun.Survey of Developments of MapReduce Parallel Computing Technology.COMPUTER SYSTEMS APPLICATIONS,2014,23(4):1-6,11
应毅,刘亚军.MapReduce并行计算技术发展综述.计算机系统应用,2014,23(4):1-6,11
YING Yi,LIU Ya-Jun.Survey of Developments of MapReduce Parallel Computing Technology.COMPUTER SYSTEMS APPLICATIONS,2014,23(4):1-6,11