云计算的核心技术之一是MapReduce技术,目前国内外云计算研究机构都对MapReduce技术特别关注,并进行了基于不同体系结构上的实现研究,尤其是在对其所做的研究是基于开源hadoop 平台,这些都为进一步研究提供了机遇；文章对传统MapReduce模型的处理流程进行了介绍,分析了传统处理过程中目前存在的一些问题,最后针对这些问题详细阐述了模型的改进方案,改进的方案保证了 ReduceTask均衡和控制了 ReduceTask大小.%MapReduce technology is one of the core technology of Cloud computing, cloud computing research institutions have great concern for the MapReduce technology at home and abroad, and different architectures have carried out to achieve, especially in the open-source platform for its hadoop research done more research opportunities. This paper describes the traditional model of MapReduce processing and analysis of the treatment process may now be some problems, and finally provides a model improvement program. It can ensure Reduc-eTask balanced and control ReduceTask size.
展开▼
