? dCompaction:基于延迟的日志结构合并树的合并方法
Journal of Computer Science and Technology
Quick Search in JCST
 Advanced Search 
      Home | PrePrint | SiteMap | Contact Us | Help
 
Indexed by   SCIE, EI ...
Bimonthly    Since 1986
Journal of Computer Science and Technology 2017, Vol. 32 Issue (1) :41-54    DOI: 10.1007/s11390-017-1704-4
Computer Architecture and Systems << Previous Articles | Next Articles >>
dCompaction:基于延迟的日志结构合并树的合并方法
Feng-Feng Pan1,2(潘锋烽), Student Member, CCF, ACM, IEEE, Yin-Liang Yue3(岳银亮), Member, CCF, ACM, IEEE, and Jin Xiong1,2(熊劲), Senior Member, CCF, Member, ACM, IEEE
1 State Key Laboratory of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences Beijing 100190, China;
2 University of Chinese Academy of Sciences, Beijing 100049, China;
3 Institute of Information Engineering, Chinese Academy of Sciences, Beijing 100093, China
dCompaction: Speeding up Compaction of the LSM-Tree via Delayed Compaction
Feng-Feng Pan1,2(潘锋烽), Student Member, CCF, ACM, IEEE, Yin-Liang Yue3(岳银亮), Member, CCF, ACM, IEEE, and Jin Xiong1,2(熊劲), Senior Member, CCF, Member, ACM, IEEE
1 State Key Laboratory of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences Beijing 100190, China;
2 University of Chinese Academy of Sciences, Beijing 100049, China;
3 Institute of Information Engineering, Chinese Academy of Sciences, Beijing 100093, China

摘要
参考文献
相关文章
Download: [PDF 962KB]  
摘要 键值存储系统在当今互联网应用中发挥着巨大的作用。写优化的数据结构,例如日志结构合并树以及它的变种,被广泛应用于各类键值存储系统中,如Bigtable,RocksDB等。传统的日志结构合并树通过两层有序数据集或者多层有序数据集对索引变更进行延迟及批量处理,并通过类似于归并排序的方式高效地将磁盘上的多个有序数据集进行合并,而在多层合并过程中会由于大量键值对重复读写引发写放大问题,导致性能下降。本文针对合并过程中的写放大问题,提出了一种新的合并方法—dCompaction,其核心思想是通过延迟调度部分合并操作方式,来减少合并过程中的键值对重复读写的问题,从而提升了系统的吞吐率。本文以RocksDB为原型在其上实现了dCompaction策略,并利用YCSB进行大量的测试,其结果表明与RocksDB相比,dCompaction在保持读性能不变的情况下,写性能提升40%左右。
关键词键值存储系统   日志结构合并树   写放大   延迟合并     
Abstract: Key-value (KV) stores have become a backbone of large-scale applications in today's data centers. Writeoptimized data structures like the Log-Structured Merge-tree (LSM-tree) and their variants are widely used in KV storage systems like BigTable and RocksDB. Conventional LSM-tree organizes KV items into multiple, successively larger components, and uses compaction to push KV items from one smaller component to another adjacent larger component until the KV items reach the largest component. Unfortunately, current compaction scheme incurs significant write amplification due to repeated KV item reads and writes, and then results in poor throughput. We propose a new compaction scheme, delayed compaction (dCompaction) that decreases write amplification. dCompaction postpones some compactions and gathers them into the following compaction. In this way, it avoids KV item reads and writes during compaction, and consequently improves the throughput of LSM-tree based KV stores. We implement dCompaction on RocksDB, and conduct extensive experiments. Validation using YCSB framework shows that compared with RocksDB, dCompaction has about 40% write performance improvements and also comparable read performance.
Keywordskey-value store   Log-Structured Merge-tree (LSM-tree)   write amplification   delayed compaction     
Received 2016-08-01;
本文基金:

This work is supported by the National Key Research and Development Program of China under Grant No. 2016YFB1000202 and the National Natural Science Foundation of China under Grant Nos. 61303056 and 61379042.

About author: Feng-Feng Pan received his B.S. degree in computer science and technology from Central South University, Changsha, in 2010. He is now a Ph.D. candidate in the Institute of Computing Technology, Chinese Academy of Sciences, Beijing. His research interests include big data storage and management, storage systems, and big data analysis.
引用本文:   
Feng-Feng Pan, Yin-Liang Yue, Jin Xiong.dCompaction:基于延迟的日志结构合并树的合并方法[J]  Journal of Computer Science and Technology , 2017,V32(1): 41-54
Feng-Feng Pan, Yin-Liang Yue, Jin Xiong.dCompaction: Speeding up Compaction of the LSM-Tree via Delayed Compaction[J]  Journal of Computer Science and Technology, 2017,V32(1): 41-54
链接本文:  
http://jcst.ict.ac.cn:8080/jcst/CN/10.1007/s11390-017-1704-4
Copyright 2010 by Journal of Computer Science and Technology