Everything is a Transaction: Unifying Logical Concurrency Control and Physical Data Structure Maintenance in Database Management Systems
Everything is a Transaction: Unifying Logical Concurrency Control and Physical Data Structure Maintenance in Database Management Systems
复制标题
DOI：
--
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
Ling Zhang;Matthew Butrovich;Tianyu Li;Yash Nannapanei;Andrew Pavlo;J. Rollinson;Huanchen Zhang;Ambarish Balakumar;Daniel Biales;Ziqi Dong;Emmanuel Eppinger;Jordi Gonzàlez;Wan Shen Lim;Jianqiao Liu;Prashanth Menon;Soumil Mukherjee;Tanuj Nayak;Amadou Latyr Ngom;Jeff Niu;D. Patra;P. Raj;Stephanie Wang;Wuwen Wang;Yao-Tin Yu;William Zhang
中科院分区：
文献类型：
--
作者：
Ling Zhang;Matthew Butrovich;Tianyu Li;Yash Nannapanei;Andrew Pavlo;J. Rollinson;Huanchen Zhang;Ambarish Balakumar;Daniel Biales;Ziqi Dong;Emmanuel Eppinger;Jordi Gonzàlez;Wan Shen Lim;Jianqiao Liu;Prashanth Menon;Soumil Mukherjee;Tanuj Nayak;Amadou Latyr Ngom;Jeff Niu;D. Patra;P. Raj;Stephanie Wang;Wuwen Wang;Yao-Tin Yu;William Zhang
几乎每个数据库管理系统（DBMS）支持在过去十年中创建的交易，实现了多个并发控制（MVCC）。 - 结果，交易的逻辑语义与DBMS的基本实施者之间存在脱节。这些数据结构和非交易维护任务。在与系统的交易语义集成的MVCC DBM中。交易处理。我们表明，DAF可以支持垃圾收集和索引清洁，而无需损害性能，同时支持更高级别的实施目标，例如非阻滞模式的变化。
Almost every database management system (DBMS) supporting transactions created in the last decade implements multi-version concurrency control (MVCC). Still, these systems rely on physical data structures (e.g., B+trees, hash tables) that do not natively support multi-versioning. As a result, there is a disconnect between the logical semantics of transactions and the DBMS’s underlying implementation. System developers must invest engineering efforts in coordinating transactional access to these data structures and non-transactional maintenance tasks. This burden leads to challenges when reasoning about the system’s correctness and performance and inhibits its modularity. In this paper, we propose the Deferred Action Framework (DAF), a new system architecture for scheduling maintenance tasks in an MVCC DBMS integrated with the system’s transactional semantics. DAF allows the system to register arbitrary actions and then defer their processing until they are deemed safe by transactional processing. We show that DAF can support garbage collection and index cleaning without compromising performance while facilitating higher-level implementation goals, such as non-blocking schema changes.