InnoDB is well-known and popular, but naturally there's room for improvement. There are many hard-coded constants – what happens if we turn them into command-line parameters? InnoDB is also incredibly complex, with a lot going on under the covers. What algorithms can we change to make InnoDB faster and more scalable, especially on modern hardware?

Writing a custom engine can seem intimidating and creating a generic engine can be. This presentation will show how to refine the requirements to a manageable subset. We will discuss the problem addressed, the approach chosen, storage engine design and API required to implement the design. We will talk about lessons learned implementing the storage engine to improve flexibility and performance.

This session will introduce DTrace, OpenSolaris's Dynamic Trace Facility, which can be used to give unprecedented access to the inner workings of MySQL. We will explore specific MySQL DTrace probes and the DTrace pid provider, which allows the gathering of information about any function within a process.

In this session we present OpenLDAP/back-ndb which gives an LDAP interface to MySQL Cluster. The LDAP interface on Cluster offers some nice ways of scaling out, as well fantastic performance, and we discuss how this is done. Moreover, it is possible to simultaneously access the LDAP data using e.g SQL! A short demo will be given.

Enterprise email is broken. From managing metadata to the inefficiency of multiple message storage, IT administrators have long bemoaned the task of implementing a scalable, affordable, reliable messaging architecture. Zimbra and MySQL’s combined expertise has altered the way administrators control their email servers through the development of a more effective back end for messaging architecture.

What is libdrizzle? A complete client library rewrite for the Drizzle project. Why is this interesting for MySQL users? It supports the MySQL protocol and provides a non-blocking I/O interface. This session will demonstrate the power of connection pooling and how to run parallel queries to decrease overall processing time. PHP examples will be used, but other languages will be very similar.

Maatkit is a set of tools that helps MySQL users and administrators fill otherwise unmet needs. The toolkit is absolutely packed with functionality. This session will teach you how to get your daily work done faster, safer and smarter.

Come learn the fundamentals of how to leverage Gearman, the open-source, distributed job queuing system. Gearman can help in building scalable MySQL-backed applications or use through a user-defined function. Gearman’s generic design allows it to be used as a building block for a wide variety of applications, from farming out image processing to building your own Map/Reduce implementation.

Everybody knows that indexes can boost SELECT performance, but how do you verify that you are using index correctly? In this session, I'll explain detailed mechanism of traditional MySQL B+Tree indexes, optimizer plans, and how to boost performance. Not only speeding up SELECT, but also speeding INSERT are covered.

The memcached Functions for MySQL are User Defined Functions (UDFs) that give you the power to interact with both MySQL and memcached in one place, giving you the ability to have read-through or write-through caching using a number of tricks, all without having to have caching logic in the application. Learn from Patrick Galbraith how to take advantage of these great new UDFs!

Habbo, the world's leading virtual world for teens, has been running on MySQL ever since year 2000 and serves millions of users each month. It's global operation and continued product development depend on timely analysis of masses of data, a challenge we solved using Infobright's Brighthouse DWH engine. I'll tell you why and how we did that.

Craigslist serves millions of classifieds to millions of users around the world on a daily basis and MySQL is our database of record for all this data. But how do you efficiently design a search system to handle this volume?

In blazing speed we will cover the most important features on MYSQL Cluster 7.0: online add node, multi-threading of the data node and a bunch of other features that pushes the limits of MySQL Cluster even further. A demo will be made showing how to add a node.

MySQL deeply embedded version, libmysqld, is coming back in MySQL Version 5.1. This talk gives an overview of libmysqld, the advantages and the limitations, as well as example applications and sample code. The talk is centered around C applications, so knowledge of C is assumed to appreciate this talk fully.

Learn why MySQL was chosen as core database in mission critical system for fund managers of financial firms. In this mission critical financial system, in which database manages more than 1TB and handles over 200 concurrent transactions per second, all operations of front office team are managed including positions, executions and capital management.

MySQL is a great DBMS with a vast feature set. It is lightweight and easy to use "out of the box". What it lacks is a good High-Availability (HA) solution though. There are different approaches to create a HA environment - Depending on software setup and application preferences. This talk will cover everything from pros and cons of different solutions to real world experiences.

An updated overview of MySQL High Availability deployment options with DRBD. Covers DRBD for low-latency workloads (leveraging Dolphin Express interconnect technology), four-way replicated storage redundancy, end-to-end replication integrity checking, and more.

In the daily work of MySQL Support Engineers, alongside genuine problems, there are some that are plain mistakes. These mistakes can become forum questions, or bug reports, or finally Support issues. Even the most advanced users make terrible mistakes that become interesting support case stories. This presentation is a collection of those.

Amazon EC2 is a new platform on which to run MySQL. This session evaluates the performance possible on the variety of hardware configurations available with EC2.

Binlogs are essential for the MySQL Replication. MySQL Proxy can unleash the full power of binlogs: filtering, rewriting, splitting and merging.