【摘要】：如何有效地將海量數(shù)據(jù)分布到存儲(chǔ)節(jié)點(diǎn),是存儲(chǔ)系統(tǒng)首要解決的問題.提出的MJHAR(matrix-based jump hash algorithm for replication data)對(duì)象分布算法簡(jiǎn)潔、高效,支持權(quán)值和數(shù)據(jù)冗余機(jī)制.該算法創(chuàng)造性地將節(jié)點(diǎn)映射到二維矩陣,對(duì)象的分布、定位只需從矩陣的行內(nèi)、行間計(jì)算目標(biāo)節(jié)點(diǎn)的行號(hào)和列號(hào)即可.理論研究表明,該算法滿足公平性、自適應(yīng)性、緊湊性、節(jié)點(diǎn)變化對(duì)象遷移量較小的特點(diǎn).實(shí)驗(yàn)結(jié)果表明,該算法的計(jì)算時(shí)間比一致性hash算法快40%,比跳躍hash算法快23%,極大地縮短了計(jì)算時(shí)間,且比一致性hash算法對(duì)象分布更加均勻.
[Abstract]:MJHAR (matrix-based jump hash algorithm for replication data) object distribution algorithm is simple, efficient, and supports weight and data redundancy mechanism. The algorithm creatively maps nodes to two-dimensional matrix, object distribution and location. Theoretical research shows that the algorithm satisfies the characteristics of fairness, adaptability, compactness and less object migration. Experimental results show that the algorithm is 40% faster than consistent hash algorithm, 23% faster than jump hash algorithm, and greatly shortens the computation time. The computation time is more uniform than that of the consistent hash algorithm.
【作者單位】： 西安交通大學(xué)電子與信息工程學(xué)院;
【基金】：國(guó)家重點(diǎn)研發(fā)計(jì)劃(2016YFB1000303) 國(guó)家自然科學(xué)基金(61672423)~~
【分類號(hào)】：TP333

【相似文獻(xiàn)】

相關(guān)期刊論文 前2條

1 徐鳳剛;許俊奎;潘清;;可擴(kuò)展Hash方法的一種改進(jìn)算法[J];計(jì)算機(jī)工程與應(yīng)用;2006年04期

2 陳濤;肖儂;劉芳;付長(zhǎng)勝;;基于聚類和一致Hash的數(shù)據(jù)布局算法[J];軟件學(xué)報(bào);2010年12期

，

本文編號(hào)：2185300