| 基于再生码的云存储系统——Ustor |
| |
| 引用本文： | 柳 青,冯 丹,李 白.基于再生码的云存储系统——Ustor[J].通信学报,2014,35(4):19-173. |
| |
| 作者姓名： | 柳 青 冯 丹 李 白 |
| |
| 作者单位： | 华中科技大学 计算机科学与技术学院，湖北 武汉 430074 |
| |
| 基金项目： | 国家重点基础研究发展计划(“973”计划)基金资助项目(2011CB302301)；国家自然科学基金资助项目(61025008, 61232004)；国家高技术研究发展计划(“863”计划)基金资助项目(2013AA013203)；国家科技支撑计划基金资助项目(2011BAH04B02)；中央高校基本科研业务费基金资助项目(2013TS043)；电子信息产业发展基金资助项目 |
| |
| 摘 要： | 摘 要：Ustor是一个构建在多个商业云存储服务之上的云存储系统，它旨在保证数据可靠性的同时减少单点失效时占用的修复带宽。不同于将所有数据存储在单个云中，Ustor将数据编码后分布在多个云存储系统中保证可靠性。Ustor的编码模块部署了包括Reed-Solomon码和功能性修复再生码（FRC）在内的多种纠删码，是第一个将功能性修复再生码应用于多个异构的、真实的云存储系统中的应用。与传统的冗余编码比较，FRC显著地减少了单个云存储发生数据丢失时需要从网络上传输的数据量。实验表明：与不编码比较，冗余编码给系统增加了5%~10%的响应时间开销，但可保障节点失效；FRC码编、解码和修复速度与Reed-Solomon码基本相当，256 MB大小文件编码时间差距在0.5 s以内；FRC码修复时与传统的Reed-Solomon码相比减少了25%以上需要下载的数据量。
|
| 关 键 词： | 关键词：云存储 再生码 纠删码 修复带宽 |
Ustor: cloud storage system based on regenerating codes |
| |
| Qing LIU,Dan FENG,Bai LI.Ustor: cloud storage system based on regenerating codes[J].Journal on Communications,2014,35(4):19-173. |
| |
| Authors: | Qing LIU Dan FENG Bai LI |
| |
| Affiliation: | School of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan 430074, China |
| |
| Abstract: | Abstract: Ustor is a cloud storage system built on multiple commercial cloud storage services, while it aims to reduce the repair bandwidth for a single failure and keeps the data reliability in the meantime. Different from storing all data in one cloud, Ustor encodes data and distributes them into several clouds for reliability. Coding module of Ustor deployed kinds of erasure codes, including Functional repair Regenerating Codes(FRC) and Reed-Solomon codes. Ustor is the first application that adopts functional repair regenerating codes over several heterogeneous, real cloud systems. Compared to traditional erasure codes, FRC markedly reduces the amount of data downloaded through the network when data loss happens in a single cloud. Observations can be drawn from the experimental results: First, erasure coding guarantees data reliability, increasing 5%~10% response time compared to non-coding; Second, speed of FRC encode, decode and repair are a little slower than Reed-Solomon code, and the time gap of encoding 256MB by two codes is less than 0.5 seconds; Third, compared with tranditional RS codes, FRC reduces more than 25% data amount needed to download during a repair in our experiment. |
| |
| Keywords: | Key words: cloud storage regenerating codes erasure codes repair bandwidth |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《通信学报》浏览原始摘要信息 |
|
点击此处可从《通信学报》下载全文 |
|
