An Implementation of a Distributed File Storage System and the Performance Analysis
BibTeX
@MISC{Xi_animplementation,
author = {Qian Xi and Wei Dong},
title = {An Implementation of a Distributed File Storage System and the Performance Analysis},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
File storage system is an interesting topic in distributed system research. Researchers are finding ways to guarantee a persistent available system. OceanStore is a global persistent data store designed to scale to billions of users. It provides a consistent, highly-available, and durable storage utility atop an infrastructure comprised of untrusted servers. [1] Instead of focusing on a large scale system, we starts with relatively small one that assumes no heavy request pressure at a very short time period. There’re two main principles of our design: Reed-Solomon Algorithm for fault tolerance and reservation time before retrieval for alleviating server’s burden. Design It is well-known that Reed-Solomon codes may be used to provide error correction for multiple failures in RAID like systems. [2] In our system, each submitted file is first divided into segments of the same size. And then each segment contains several data blocks of the same size. Certain checksum blocks are calculated from these data blocks. All the i th blocks (including both data and checksum blocks) of different segments in a single file are assigned to the same node. Assume that in each segment, the number of data blocks is
Keyphrases
performance analysis distributed file storage system data block th block reed-solomon algorithm heavy request pressure untrusted server fault tolerance certain checksum block large scale system several data block interesting topic persistent available system checksum block different segment file storage system reed-solomon code durable storage utility server burden error correction reservation time distributed system research single file multiple failure main principle global persistent data store short time period
