Paper Info
Title
Hdstore: An Ssd/Hdd Hybrid Distributed Storage Scheme For Large-Scale Data
Abstract
Traditional data storage schemes are primarily based upon Hard Disk Drives (HDD). However, with the appearance of large amount of data on the Web, the read/write performance based on HDD has reached a bottleneck. Thus the emerging of Solid State Drives (SSD) has provided an opportunity for the storage of the Web of data. In this paper, we propose an SSD/HDD hybrid distributed storage scheme, called HDStore, for large-scale data, in which the single fix-sized journal file using the append-only mode is stored on SSD to support efficient read and write, while several segment files focusing on read are stored on HDD. Through a series of operations build, split, move, and merge between the journal and segment files, we constructed HDStore storage scheme based on JS-model. The experimental results show that HDStore obtains an efficient optimization of data read/write, especially the write performance has increased by 15 % compared to the traditional HDD-based scheme.
Year
DOI
Venue
2014
10.1007/978-3-319-11538-2_20
WEB-AGE INFORMATION MANAGEMENT: WAIM 2014 INTERNATIONAL WORKSHOPS
Keywords
Field
DocType
JS-model, HDstore, SSD, Hybrid distributed storage scheme
Data mining,Bottleneck,Computer data storage,Computer science,Distributed data store,Merge (version control),Solid-state,Operating system
Conference
Volume
ISSN
Citations 
8597
0302-9743
1
PageRank 
References 
Authors
0.35
10
6
Name
Order
Citations
PageRank
Zhijie Feng110.35
Zhiyong Feng2794167.21
Xin Wang36524.90
Guozheng Rao4309.79
Yazhou Wei510.35
Zhiyuan Li61380155.70