（北京航空航天大学软件学院北京市 100191）
摘要：企业信息化管理手段越来越丰富，大数据的处理却成为许多技术型企业的发展瓶颈，本文以某销售企业的进销存管理系统为蓝本，研究从传统单服务器系统升级为支持分布式并行处理的多服务器系统，即重构企业私有云平台的技术实现方式。该方式采用分治策略，调用系统预设的算法，将大数据任务均匀分解为若干规模较小的子任务集群，由中心控制节点根据系统动态负载均衡原理分发到对应的空闲子任务处理节点上，再将全部子任务计算结果统计汇总，实现大数据任务的分布式处理功能。该系统主要应用于解决企业对海量数据挖掘的需求。
关键词：大数据处理；分布式系统；云计算；私有云平台；数据挖掘。
Design and Implementation of DistributedComputing Systemand Performance Study
Jianchao Yang
（School of Software, Beijing University of Aeronautics and Astronautics, Beijing 100191）
vheaven@163.com
Abstract：Enterprise information management is more and more abundant; the processing of big data becomes the developmentbottleneckin many technology companies. This paper takes a clothing sales enterprise’sInvoicing management system as model, to upgrade it from the traditional single server system to multi servers system which supportparallelprocessing on distributed system, namely restructures the enterprise private cloud platform. The technology uses the divide-conquer strategy, calls the system default algorithm, decomposes a large task into several smaller subtasks and data clusters. According to the dynamic load balancing principle, a central control node distributes these subtasks to the remain free load subtask processing nodes, then summarizes all the subtasks’ results,solving the big data processing problem. The system is mainly used to meet an enterprise’s massive data mining needs.
Key words：big data processing; distributed system; cloud computing; private cloud platform; data mining.
参考文献
龚溪东，基于分布式服务的程序评测技术研究，北京航空航天大学，2011年5月
陆嘉恒，分布式系统及云计算概论，清华大学出版社，2011年5月
Gaston C. Hillar, Professional Parallel Programming，Jan 2012
Kai Hwang, Geoffrey C.Fox, Jack J.Dongarra, Distributed and Cloud Computing From Parallel Processing to the Internet of Things, Jan 2013
Qinyi Wu, Calton Pu, Danesh Irani, Cosmos A Wiki Data Management System, College of Computing Georgia Institute of Technology Atlanta, GA, 2010
Ronnie Chaiken, Bob Jenkins, Per-Åke Larson, SCOPE Easy and Efficient Parallel Processing of Massive Data Sets, Microsoft Corporation, ACM 978-1-60558-306-8/08/08, August 2008
Jingren Zhou, Nicolas Bruno, Mingchuan Wu, SCOPE parallel databases meet MapReduce, The VLDB Journal (2012) 21:611–636 DOI 10.1007/s00778-012-0280-z,June 2012
Jiawei Han & Micheline Kamber & Jian Pei, Data Mining Concepts and Techniques, America, 2011
作者简介：
杨剑超，北京航空航天大学软件学院硕士研究生，移动云计算专业，微软实习生，曾参与多个科研、商业项目研发，主要从事分布式计算，云计算，数据挖掘方向的研究。
