Percona Live: Data Performance Conference 2016 Logo

April 18-21, 2016

Santa Clara, California

Developing for extreme performance - learn from MySQL Cluster

Developing for extreme performance - learn from MySQL Cluster

Experience level: 
50 minutes conference
Tools and Techniques
High Availability


While MySQL Cluster wouldn't be MySQL Cluster without presenting a new level of extreme performance this session is focused on the how-to. We previously demonstrated 1 billion writes per minute, and exceeded 200 million queries per second (QPS), all with open source software running on commodity servers. MySQL Cluster is a high-available in-memory auto-sharded distributed data store. It has native NoSQL node.js and java connectors as well as SQL access through the MySQL Server. In this technical session we'll take a deep dive on how to scale to those extreme levels of real time performance and what you can learn from that for your own solution. Discover how network, CPU, interrupts handling affect performance of a distributed system and how sometimes delays and bottlenecks are not that bad after all.


Bernd Ocklin's picture

Bernd Ocklin

Director MySQL Cluster Development, Oracle


Bernd Ocklin, responsible for running the MySQL Cluster team at Oracle, recently celebrated his first 10 years with the MySQL organization. MySQL Cluster is the high performance and availability core of the largest mobile phone networks and for many popular gaming back ends. Bernd first worked as a solution architect for high available database and clustering solutions and then joined the MySQL Cluster development at Sun/Oracle. To his credit, he has designed and programmed machine learning and data analysis solutions at his own company in addition to writing a database core prior to using MySQL databases. The outcome of such work has proven extremely useful in big scale data processing, i.e., for fraud detection at the world’s largest derivatives exchange.

Share this talk