面向政府治理大数据的高性能计算系统  

High performance big data computing systems for government governance

作　　者：吴维刚[1] 常亮 任江涛 古天龙 WU Weigang;CHANG Liang;REN Jiangtao;GU Tianlong(School of Data and Computer Science,Sun Yat-sen University,Guangzhou 510006,China;School of Computer Science and Information Security,Guilin University of Electronic Technology,Guilin 541004,China)

机构地区：[1]中山大学数据科学与计算机学院,广东广州510006 [2]桂林电子科技大学计算机与信息安全学院,广西桂林541004

出　　处：《大数据》2020年第2期41-56,共16页Big Data Research

基　　金：国家自然科学基金资助项目(No.U1711263)。

摘　　要：大数据处理系统是未来社会的基础设施之一。政府治理场景下的大数据处理任务具有多域异构、多主体等特点,因此需要针对性地进行研究设计。从应用需求出发,分析各类政府治理场景对大数据处理技术提出的挑战,梳理大数据分布并行处理的关键技术,包括数据存储管理、计算平台、关键算法等,调研总结相关技术的研究现状,并提出面向政府治理大数据的高性能计算系统的技术框架,分析讨论不同技术路线的优劣。最后展望相关技术的未来发展趋势。The big data processing system would be one of the major infrastructures of future society.The characteristics and requirements of government governance applications,including multi-domain,multi-entity,etc.,bring new challenges to big processing platform,and new design is desirable.Based on the requirements of applications,the challenges in big data processing for government governance scenarios were analyzed,and key techniques in distributed parallel processing for public governance were discussed,including data storage,computing platform,and key algorithms.The state-of-art techniques were investigated and reviewed,and a candidate framework for government governance data process system was proposed,the pros and cons of different technical methodologies were also discussed.Finally,the open problems and future directions were also discussed.

关 键 词：大数据处理 政府治理 分布式计算 计算框架 资源管理 

分 类 号：TP316[自动化与计算机技术—计算机软件与理论]