arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 1809.03261cs.DB
收藏

基于Skiplist的LSM树

The Skiplist-Based LSM Tree

Aron Szanto

摘要:

日志结构化合并(Log-Structured Merge,LSM)树提供了一种分层的数据存储和检索范例,这对于写优化的数据系统是有吸引力的。在内存中保持一个有效的缓冲区,并推迟更新超过其初始写入时间,该结构提供了快速的操作热数据。由于结构的每一层在逻辑上都是相互独立的,因此该结构也有利于机会优化和粒度优化。在本文中,我们介绍了基于跳跃列表的LSM树(sLSM),一种新的系统中,LSM的内存缓冲区是由一个序列的跳跃列表。我们开发的理论和实验结果表明,sLSM固有的调整参数的广度允许它广泛的灵活性,在各种各样的工作负载的优异性能。

英文摘要:

Log-Structured Merge (LSM) Trees provide a tiered data storage and retrieval paradigm that is attractive for write-optimized data systems. Maintaining an efficient buffer in memory and deferring updates past their initial write-time, the structure provides quick operations over hot data. Because each layer of the structure is logically separate from the others, the structure is also conducive to opportunistic and granular optimization. In this paper, we introduce the Skiplist-Based LSM Tree (sLSM), a novel system in which the memory buffer of the LSM is composed of a sequence of skiplists. We develop theoretical and experimental results that demonstrate that the breadth of tuning parameters inherent to the sLSM allows it broad flexibility for excellent performance across a wide variety of workloads.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑