1 Introduction The LSM-tree has become a preferred solution for the storage structure of key-value(KV)NoSQL systems due to its efficient write performance[1,2].Its ability to efficiently process large-scale data write...1 Introduction The LSM-tree has become a preferred solution for the storage structure of key-value(KV)NoSQL systems due to its efficient write performance[1,2].Its ability to efficiently process large-scale data write operations makes it show high performance in handling TP(Transactional Processing)tasks.The increasing volume of real-time data processing and complex analytics tasks in current applications has made it necessary for databases to support HTAP(Hybrid Transactional/Analytical Processing)workloads.展开更多
大数据对存储系统的可扩展性、性能和成本等方面提出了更高的要求。瓦记录(Shingled Magnetic Recording,SMR)硬盘由于存储密度高、价格便宜,正逐步被广泛应用于大数据存储系统。但是,SMR硬盘的随机写性能较差,与快速的基于闪存的固态硬...大数据对存储系统的可扩展性、性能和成本等方面提出了更高的要求。瓦记录(Shingled Magnetic Recording,SMR)硬盘由于存储密度高、价格便宜,正逐步被广泛应用于大数据存储系统。但是,SMR硬盘的随机写性能较差,与快速的基于闪存的固态硬盘(Solid State Drive,SSD)一起构成混合存储时可以显著提升性能。同时,基于写优化的日志结构合并(Log-Structured Merge,LSM)树的键值存储已被广泛应用于许多NoSQL系统,如BigTable,Cassandra和HBase等。因此,如何基于新型的SSD-SMR混合存储构建出高性能的LSM树键值存储系统是一个具有很大研究价值的问题。首先建立基于SSD-SMR混合存储的LSM树键值系统的性能模型,然后针对SSD和SMR的硬件特征以及LSM树键值存储的软件特点,设计了一套面向SSD-SMR混合存储进行性能优化的LSM树键值存储系统,并基于LevelDB实现了该系统。在仅仅使用0.4%~2%空间的SSD的情况下,所提方法可以使SSD-SMR混合存储方案比普通磁盘方案的随机写性能提高20%,随机读性能提高5倍。展开更多
基金supported by the Fundamental Research Funds for the National Natural Science Foundation of China under Grant No.62472077.
文摘1 Introduction The LSM-tree has become a preferred solution for the storage structure of key-value(KV)NoSQL systems due to its efficient write performance[1,2].Its ability to efficiently process large-scale data write operations makes it show high performance in handling TP(Transactional Processing)tasks.The increasing volume of real-time data processing and complex analytics tasks in current applications has made it necessary for databases to support HTAP(Hybrid Transactional/Analytical Processing)workloads.
文摘大数据对存储系统的可扩展性、性能和成本等方面提出了更高的要求。瓦记录(Shingled Magnetic Recording,SMR)硬盘由于存储密度高、价格便宜,正逐步被广泛应用于大数据存储系统。但是,SMR硬盘的随机写性能较差,与快速的基于闪存的固态硬盘(Solid State Drive,SSD)一起构成混合存储时可以显著提升性能。同时,基于写优化的日志结构合并(Log-Structured Merge,LSM)树的键值存储已被广泛应用于许多NoSQL系统,如BigTable,Cassandra和HBase等。因此,如何基于新型的SSD-SMR混合存储构建出高性能的LSM树键值存储系统是一个具有很大研究价值的问题。首先建立基于SSD-SMR混合存储的LSM树键值系统的性能模型,然后针对SSD和SMR的硬件特征以及LSM树键值存储的软件特点,设计了一套面向SSD-SMR混合存储进行性能优化的LSM树键值存储系统,并基于LevelDB实现了该系统。在仅仅使用0.4%~2%空间的SSD的情况下,所提方法可以使SSD-SMR混合存储方案比普通磁盘方案的随机写性能提高20%,随机读性能提高5倍。
