本文選題：CEPH + CRUSH�。� 參考：《微電子學(xué)與計(jì)算機(jī)》2017年06期

【摘要】：通過分析集群中單節(jié)點(diǎn)在不同網(wǎng)絡(luò)延時(shí)下的延時(shí)大小與集群存儲(chǔ)性能的關(guān)系,提出基于網(wǎng)絡(luò)時(shí)延的權(quán)重改進(jìn)方法.實(shí)驗(yàn)表明所提出的方法可根據(jù)存儲(chǔ)節(jié)點(diǎn)的網(wǎng)絡(luò)時(shí)延對(duì)權(quán)重進(jìn)行再調(diào)整,使數(shù)據(jù)較大概率存入網(wǎng)絡(luò)時(shí)延小的節(jié)點(diǎn).該方法降低了由單一節(jié)點(diǎn)網(wǎng)絡(luò)延時(shí)引起的存儲(chǔ)性能損耗,從而提高了CEPH分布式存儲(chǔ)系統(tǒng)的存儲(chǔ)性能.
[Abstract]:Through the relationship between the single node in the cluster analysis in different network delay delay under the cluster size and storage performance, put forward the improvement method of weighted network based on delay. Experiments show that the proposed method can be adjusted according to the weight of the network delay storage node, the data is stored in the nodes of the network delay probability. This method reduces the the storage performance loss caused by a single node of the network delay, so as to improve the storage performance of CEPH distributed storage system.

【作者單位】： 桂林電子科技大學(xué)信息與通信工程學(xué)院;桂林電子科技大學(xué)廣西高校云計(jì)算與復(fù)雜系統(tǒng)重點(diǎn)實(shí)驗(yàn)室;
【基金】：國(guó)家自然科學(xué)基金(61172053,61163058,61662018) 廣西云計(jì)算與復(fù)雜系統(tǒng)高校重點(diǎn)實(shí)驗(yàn)室(2015209,15207)
【分類號(hào)】：TP333
，

本文編號(hào)：1765754