Concurrency Control in RDBMS
In this tutorial, we will delve into how modern relational database systems handle concurrency. A guided tour of the important features in different RDBMS is provided first, after which we should have sufficient ammunition to discuss the inevitable trade-offs that come with nearly every design aspect. As the word "trade-off" already indicates, there is no single perfect way to do concurrency control - so this won't be a "mine is better than yours" event. Naturally, MySQL will be covered, but also most other well known free/open source and proprietary RDBMS. Buzzwords: ACID: Atomicity, Consistency, Isolation, Durability, Transactions, COMMIT, ROLLBACK, multi-versioning, locking, dirty and repeatable reads, phantoms, serialisation, lock escalation, clustered indexes, next-key locking. This session is aimed at people who are interested in learning (more) about how RDBMS do their magic. It may not be suitable for newbies, but you really don't have to be a hard-core hacker to dig this stuff. In fact, when using RDBMS it is a great asset to have sufficient understanding of this topic.
Arjen Lentz (36) is originally from Amsterdam, The Netherlands. Together with his wife, baby daughter and black cat he resides in Kenmore (Brisbane), Australia. Arjen initially started with programming, but has since branched out to blends initial geekness with more human interaction - like writing and training. He has been working for MySQL AB since 2001. He is now MySQL AB's Community Relations Manager, a job which involves cool stuff like jet-lag, hanging out with MySQL users and drinking assorted beverages on company time. He is also the program chair for the MySQL Users Conference.