DocumentCode
466867
Title
OBSI: Object Based Storage System for Massive Image Databases
Author
Lihua Yu ; Gang Chen ; Yijun Bei
Author_Institution
Zhejiang Univ., Hangzhou
Volume
1
fYear
2007
fDate
July 30 2007-Aug. 1 2007
Firstpage
321
Lastpage
326
Abstract
The popularity of online free album services in the recent prevailing Web2.0 paradigm has posed great challenges to the existing image storage systems. In this paper, we analyze the characteristics of online image storage techniques from a system perspective. We present a novel framework called OBSI, which stands for an object based storage system for storing and retrieving images of online album services. In the design of OBSI we employ a novel divide and conquer approach which subdivides OBSI into two layers: an object layer and a chunk layer. The chunk layer handles the "dirty work", such as disk space management and reliability guarantee. The object layer provides a uniform object storage and retrieval interface to applications. OBSI reduces disk access by employing a simple and yet effective storage allocation algorithm which allocates storage objects continuously as clustered to their respective metadata. Experimental results show that OBSI provides high efficiency and throughput in image storage and retrieval services.
Keywords
Internet; divide and conquer methods; information retrieval; meta data; visual databases; Web2.0 paradigm; disk space management; divide and conquer approach; image storage systems; interface retrieval; massive image databases; metadata; object based storage system; online image storage techniques; uniform object storage; Artificial intelligence; Costs; Distributed computing; Educational institutions; File systems; Image databases; Image retrieval; Image storage; Security; Software engineering; Distributed File System; Image database; Object based; storage;
fLanguage
English
Publisher
ieee
Conference_Titel
Software Engineering, Artificial Intelligence, Networking, and Parallel/Distributed Computing, 2007. SNPD 2007. Eighth ACIS International Conference on
Conference_Location
Qingdao
Print_ISBN
978-0-7695-2909-7
Type
conf
DOI
10.1109/SNPD.2007.371
Filename
4287525
Link To Document