学海网 文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > Erasure Coding vs Replication A QuantitativeComparison

Erasure Coding vs Replication A QuantitativeComparison

Erasure Coding vs Replication A Quantitative Comparison

Erasure Coding vs.Replication:A Quantitative

Comparison

Hakim Weatherspoon and John D.Kubiatowicz

Computer Science Division

University of California,Berkeley

hweather,kubitron@cs.berkeley.edu

Abstract.Peer-to-peer systems are positioned to take advantage of gains in net-

work bandwidth,storage capacity,and computational resources to provide long-

term durable storage infrastructures.In this paper,we quantitatively compare

building a distributed storage infrastructure that is self-repairing and resilient to

faults using either a replicated system or an erasure-resilient system.We show

that systems employing erasure codes have mean time to failures many orders

of magnitude higher than replicated systems with similar storage and bandwidth

requirements.More importantly,erasure-resilient systems use an order of magni-

tude less bandwidth and storage to provide similar system durability as replicated

systems.

1Introduction

Today’s exponential growth in network bandwidth,storage capacity,and com-putational resources has inspired a whole new class of distributed,peer-to-peer storage infrastructures.Systems such as Farsite[2],Freenet[4],Intermemory[3], OceanStore[8],CFS[5],and PAST[7]seek to capitalize on the rapid growth of resources to provide inexpensive,highly-available storage without centralized servers.The designers of these systems propose to achieve high availability and long-term durability,in the face of individual component failures,through repli-cation and coding techniques.

Although wide-scale replication has the potential to increase availability and durability,it introduces two important challenges to system architects.First, system architects must increase the number of replicas to achieve high durability for large systems.Second,the increase in the number of replicas increases the bandwidth and storage requirements of the system.

This paper makes the following contributions:First,we briefly quantify the availability gained using erasure codes.Second,we show that erasure-resilient codes use an order of magnitude less bandwidth and storage than replication for systems with similar mean time to failure(MTTF).Third,we show that employing erasure-resilient codes increase the MTTF of the system by many orders of magnitude over simple replication with the same storage overhead and

第1页

TOP相关主题

我要评论

站点地图 | 文档上传 | 侵权投诉 | 手机版
新浪认证  诚信网站  绿色网站  可信网站   非经营性网站备案
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.xuehai.net All Rights Reserved.  email
返回顶部