MapReduce并行编程模型研究综述_李建江.pdfVIP

11 Vol. 39 No. 11 2011 11 ACTA ELECTRONICA SINICA Nov. 2011 MapReduce 1 1 1 1 2 李建江, 崔 健 , 王 聃 , 严 林 , 黄义双 ( 1. , 100083; 2. , 610041) : MapReduce , , , . MapReduce , MapReduce , MapReduce ( : ) .MapReduce . : MapReduce; ; ; : TP3164 : A : 0372-2112 (2011) 11-2635-08 Survey of MapReduce Parallel Programming Model 1 1 1 1 2 LI Jia-jia g , CUI Jia , WANG Da , YAN Li , HUANG Yi-shua g ( 1. Department of Computer Science and Technology, School of Computer and Communication Engineering, University of Science and Technology Beij ing, Bejiing 100083, China; 2. esearch Institute of Exploration Southern Division Company, SINOPEC, Chengdu, Sichuan 610041, China) Abstract: Through well-defi ed i terfaces a d ru time support library, MapReduce parallel programmi g model ca auto- matically perform the large-scale computi g tasks i parallel, hide the u derlyi g impleme tatio details, a d reduce the difficulty of parallel programmi g.This paper reviews the domestic a d overseasresearch of the MapReduce, describes a d a alyzesthe charac- teristics a d lack of the typical research achieveme ts about MapReduce at home a d abroad.The this paper focus o the i-depth a alysis of the key tech ologies about MapReduce ( i cludi g: model optimizatio , model impleme tatio accordi g to the differe t platforms, task scheduli g, load bala ci g, a d fault tolera ce) . Fi ally, this paper prospects the MapReduce for the future tre d. Key words: MapReduce; parallel programmi g model; ru time support library;