April 20, 2022
SQL
《Grammar-like Functional Rules - for Representing Query Optimization Alternatives》 SIGMOD 1988
概述优化器如果要做到可扩展,则要求执行计划的变换策略需要作为数据而不是代码存在。本质上优化器就是一个专 ...
Read more
April 15, 2022
Programming
《BOLT: A Practical Binary Optimizer for Data Centers and Beyond》 CGO 2019
《Ispike: A Post-link Optimizer for the IntelR ItaniumR Architectur》 CGO ...
Read more
April 10, 2022
SQL
《Extensible/Rule Based Query Rewrite Optimization in Starburst》SIGMOD 1992
这篇文章是介绍Starburst的基于规则的优化器实现框架。
背景很多业务的SQL语句或类SQL语句,尤其是中间件软件自动生成的SQ ...
Read more
March 20, 2022
Database
VLDB 18’ Content-Aware Lock Scheduling for Transactional Database
相关的证明在 Contention-aware lock scheduling for transactionaldatabases.Technical Re ...
Read more
March 11, 2022
Database
Aurora在SIGMOD18’的论文《Amazon Aurora: On Avoiding Distributed Consensus for I/Os, Commits, and Membership Changes》中描述了Aurora关于共识方案的选择和Quorum方案的详 ...
Read more
February 19, 2022
Database
《Constructing and Analyzing the LSM Compaction Design Space》 VLDB’21
四个原语Compaction是LSM-Tree引擎中最重要的环节之一,一方面它主导了LSM-Tree的形状进而影响了LSM-Tree的包括性能、空间占用等 ...
Read more
February 18, 2022
Database
《REMIX: Efficient Range Query for LSM-trees》FAST’20
LSM-Tree引擎的scan操作是一个引擎固有的短板,其根源在于相比基于B+Tree的引擎如InnoDB的数据全局有序,LSM-Tree的数据有序性差,它由多个sort run组成,因此 ...
Read more
February 12, 2022
Database
《Evolution of Development Priorities in Key-value Stores ServingLarge-scale Applications: The RocksDB Experience》FAST’20
这篇文章是讲述Facebook大规模部署Rocks ...
Read more
February 12, 2022
Database
这是SIGMOD’21 上的论文《Rethink the Scan in MVCC Databases》,论文针对像HTAP等场景下可能出现大量versioned data中执行scan慢的问题提出了一种比较有效的方案。
Scan性能问题首先,论文看到了这种场景下scan慢的一个重要原因:
...
Read more
October 20, 2021
DistributedSystem
Scaling Replicated State Machines with Compartmentalization,VLDB 21’
本文是针对Multi-Paxos协议的实现方案瓶颈做的扩展性方案,主要的方法论是compartmentalization,这个词本意是分离,在这里作为一种 ...
Read more