基于蚁群平台的智能存储技术及应用  

Intelligent storage technology and application based on ant cluster

作　　者：钱磊 原昊 赵锦明 李祥 吴东 谢向辉 QIAN Lei YUAN Hao ZHAO Jin-ming LI Xiang WU Dong XIE Xiang-hui(State Key Laboratory of Mathematical Engineering and Advanced Computing, Wuxi 214083, China)

机构地区：[1]数学工程与先进计算国家重点实验室,江苏无锡214083

出　　处：《计算机工程与设计》2017年第5期1394-1401,共8页Computer Engineering and Design

基　　金：国家863高技术研究发展计划基金项目(2015AA01A301)

摘　　要：为提高大数据存储与分析能力,在蚁群平台基础上,提出一种面向领域应用的智能存储体DS-ISF(domain-specific intelligent storage framework)。将可重构的数据处理引擎与存储部件紧密耦合,减少数据移动开销,实现数据处理过程的硬件加速。在此基础上构建的大规模分布式原生XML数据库AntXDB,其数据查询吞吐率比x86平台上的传统XML数据库高出1-2个数量级,能效高出2-3个数量级。通过规模扩展,能支持PB级大规模XML数据存储,达到每秒上千并发查询能力,延迟不超过5s。To improve the ability of big data storage and analysis, a domain-specific intelligent storage framework （DS-ISF） was proposed based on ant cluster. A number of reconfigurable data processing engines were tightly coupled with data storage com-ponent, which helped reduce the overhead of data movement and provided hardware acceleration to data processing. A large-scale distributed native XML database named AntXDB was implemented based on DS-ISFs. Test results show that the throughput of data query on AntXDB is 1-2 orders of magnitude higher and its energy efficiency is 2-3 orders of magnitude higher than traditio-nal XML databases on x86 platform. By extending the scale of the system, it can support PB level XML data storing and thou-sands of concurrent queries per second with the delay less than 5 seconds.

关 键 词：蚁群平台 高能效 智能存储 可重构逻辑 XML数据库 

分 类 号：TP338.8[自动化与计算机技术—计算机系统结构]