@inproceedings{9217113dff1c40d6a86b6ea42814f231,
title = "Laser: Load-adaptive group commit in lock-free transaction logging",
abstract = "Log manager is a key component of DBMS and is considered as the most prominent bottleneck in the modern in-memory OLTP system. In this paper, by addressing two existing performance hurdles in the current procedure, we propose a high-performance transaction logging engine Laser and integrate it into OceanBase, an in-memory OLTP system. First, we present a lock-free transaction logging framework to eliminate the lock contention. Then we make theoretical analysis and propose a judicious grouping strategy to determine an optimized group time for different workloads. Experiment results show that it improves 1.4X–2.4X throughput and reduces more than 60\% latency compared with current methods.",
keywords = "Group commit, Load-adaptive, Lock-free, Logging",
author = "Huan Zhou and Huiqi Hu and Tao Zhu and Weining Qian and Aoying Zhou and Yukun He",
note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2017.; 1st Asia-Pacific Web and Web-Age Information Management Joint Conference on Web and Big Data, APWeb-WAIM 2017 ; Conference date: 07-07-2017 Through 09-07-2017",
year = "2017",
doi = "10.1007/978-3-319-63579-8\_25",
language = "英语",
isbn = "9783319635781",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "320--328",
editor = "Cyrus Shahabi and Xiang Lian and Jensen, \{Christian S.\} and Xiaochun Yang and Lei Chen",
booktitle = "Web and Big Data - 1st International Joint Conference, APWeb-WAIM 2017, Proceedings",
address = "德国",
}