CSR-PDOS: Exploring and Developing Improved Erasure Codes for Storage Systems

CSR-PDOS：探索和开发改进的存储系统纠删码

• 批准号: 0615221
• 负责人: James Plank
• 金额: --
• 依托单位: University of Tennessee Knoxville
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2006
• 资助国家: 美国
• 起止时间: 2006-08-01 至 2010-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0615221&HistoricalAwards=false
• 关键词: CSR; PDOS; Exploring; Developing; Improved

Storage systems are exploding in size, complexity and popularity, andas they grow, they expose themselves to a variety of failures. Assuch, the software that supports these systems must be able totolerate a growing number of failures. Erasure coding is a techniquethat allows storage nodes to be enriched with extra codinginformation so that if one or more nodes fail, the original data maybe recovered from the survivors. While erasure codes have existedfor decades, there is a considerable gap in understandingconstructions, performance and properties of codes that toleratemultiple failures. This project's mission is explore, design,evaluate and disseminate erasure codes for multiple failures thatsignificantly improve the performance and failure coverage of currentcodes. The research component is centered around unifying disparatetechniques from classical coding theory, disk array technology, andasymptotic graph theory.This work will impact all applications where multiple storage nodesare used in tandem, and the risk of one or more nodes failing issignificant. New, efficient codes will be developed and evaluatedwith respect to the best currently known codes. Additionally, theimplementations of these codes will be disseminated to the communityof systems programmers that need them, and tutorial material will bedeveloped so that the codes may have wide impact. This is a seriousissue, as most current codes are presented in a theoretcially formalway that puts them beyond the reach of the programmers who need toimplement them. Finally, erasure codes will be used as the basis formotivational programs on math and science at K-12 schools in EastTennessee.

存储系统在大小、复杂性和受欢迎程度上呈爆炸式增长，随着它们的增长，它们自身也面临着各种故障。因此，支持这些系统的软件必须能够容忍越来越多的故障。擦除编码是一种技术，它允许用额外的编码信息来丰富存储节点，以便在一个或多个节点发生故障时，可以从幸存者那里恢复原始数据。虽然纠删码已经存在了几十年，但人们对容忍多重故障的码的结构、性能和特性的理解还存在着相当大的差距。该项目的任务是探索、设计、评估和传播针对多个故障的擦除码，从而显著提高CurrentCodes的性能和故障覆盖率。这项研究集中在统一不同于经典编码理论、磁盘阵列技术和渐近图论的技术，这项工作将影响所有同时使用多个存储节点且一个或多个节点发生故障的风险很大的应用程序。将开发新的高效编码，并参照目前已知的最佳编码对其进行评估。此外，这些代码的实施将分发给需要它们的系统程序员社区，并将编写教程材料，以便这些代码可以产生广泛的影响。这是一个严重的问题，因为大多数当前的代码都是以理论上正式的方式呈现的，这使得需要实现它们的程序员无法实现它们。最后，擦除码将被用作田纳西州东部K-12学校数学和科学激励计划的基础。