發(fā)布時間：2018-03-03 13:22

  本文選題：蟻群平臺　切入點(diǎn)：高能效　出處：《計(jì)算機(jī)工程與設(shè)計(jì)》2017年05期 　論文類型：期刊論文

【摘要】：為提高大數(shù)據(jù)存儲與分析能力,在蟻群平臺基礎(chǔ)上,提出一種面向領(lǐng)域應(yīng)用的智能存儲體DS-ISF(domain-specific intelligent storage framework)。將可重構(gòu)的數(shù)據(jù)處理引擎與存儲部件緊密耦合,減少數(shù)據(jù)移動開銷,實(shí)現(xiàn)數(shù)據(jù)處理過程的硬件加速。在此基礎(chǔ)上構(gòu)建的大規(guī)模分布式原生XML數(shù)據(jù)庫AntXDB,其數(shù)據(jù)查詢吞吐率比x86平臺上的傳統(tǒng)XML數(shù)據(jù)庫高出1-2個數(shù)量級,能效高出2-3個數(shù)量級。通過規(guī)模擴(kuò)展,能支持PB級大規(guī)模XML數(shù)據(jù)存儲,達(dá)到每秒上千并發(fā)查詢能力,延遲不超過5s。
[Abstract]:In order to improve big data's ability of storage and analysis, an intelligent DS-ISF(domain-specific intelligent storage framework for domain application is proposed on the basis of ant colony platform. The reconfigurable data processing engine is tightly coupled with the storage unit to reduce the overhead of data migration. The data query throughput of the large-scale distributed native XML database AntXDB is 1-2 orders of magnitude higher than that of the traditional XML database on x86 platform. Energy efficiency is higher than 2-3 orders of magnitude. Through scale expansion, can support PB large-scale XML data storage, to thousands of concurrent queries per second, delay is not more than 5s.
【作者單位】： 數(shù)學(xué)工程與先進(jìn)計(jì)算國家重點(diǎn)實(shí)驗(yàn)室;
【基金】：國家863高技術(shù)研究發(fā)展計(jì)劃基金項(xiàng)目(2015AA01A301)
【分類號】：TP311.13;TP333
，

本文編號：1561206