纠删码存储系统中数据修复性能优化研究进展与展望

Research Progress and Prospect on Performance Optimization of Data Recovery for Erasure-Coded Storage Systems

关键词:  分布式存储系统|纠删码存储系统|数据修复操作|数据可靠性|性能优化

Key words:  distributed storage systems|erasure-coded storage systems|data recovery operation|data reliability|performance optimization

摘要:

纠删码被广泛应用于分布式存储系统以保存在线应用的用户数据。当部分存储节点发生故障时，纠删码存储系统需使用新的存储节点替换原有失效节点，并恢复失效的用户数据。由于需要执行数据编码、传输和读写等操作，纠删码存储系统通常需要消耗较长的时间执行数据修复操作，存储的用户数据将长期处于不可靠状态。为了保障存储数据的可靠性，研究学者提出了多种数据修复性能优化方案以减少数据修复时间。本文介绍了数据修复性能优化问题，分析了各个应用场景下主要的性能瓶颈和性能优化难点，总结了提升数据修复性能的主要技术方案和研究工作，并对数据修复性能优化研究领域的未来发展方向进行展望，为纠删码存储系统设计人员准确选择适合特定应用场景的数据修复性能优化方案提供思路。

Abstract:

Erasure codes are widely used in distributed storage systems to store user data for online applications.When some storage nodes fail,the erasure-coded storage systems need to replace the original failed nodes with new storage nodes and recover the failed user data.Since the erasure-coded storage systems need to perform the data encoding,transmission and read-write operations,erasure code storage systems usually take a long time to perform data repair operations,and the stored user data will be unreliable for a long time.In order to ensure the reliability of stored data,researchers have proposed a variety of data recovery performance optimization schemes to reduce data repair time.The problem of data recovery performance optimization were introduced,the main performance bottlenecks and performance optimization difficulties in various application scenarios were analyzed,the main technical implementation schemes and research work to improve the data recovery performance were summarized,the future development direction of the data recovery performance optimization research field was prospected,which provided ideas for the designers of erasure-coded storage system to accurately select the data recovery performance optimization scheme suitable for specific application scenarios.