Percona Live: Data Performance Conference 2016 Logo

April 18-21, 2016

Santa Clara, California

Why would I run MySQL or MariaDB on POWER anyway?

Why would I run MySQL or MariaDB on POWER anyway?

 19 April 5:15 PM - 05:40 PM @ Ballroom C
Experience level: 
Beginner
Duration: 
25 minutes conference
Tracks:
New and Trending Topics
Scalability/Performance
Topics:
MySQL
High Availability

Description

This session presents the technical reasons why you might want to consider the POWER architecture for MySQL/MariaDB production workloads. MySQL on POWER8 was the first hardware/software combination to run 1 Million SQL Queries per second, but that was an artificial workload (PK lookups) and not indicative of any real world performance so this session tackles features of the POWER architecture, the POWER8 processor and OpenPower machines that can make a difference in production environments. We'll look at: - raw compute power - many cores, many threads, a couple of sockets - memory bandwidth - Reliability and Serviceability: elegance in the event of hardware failure - Open Sourceness of an OpenPower system: open source OS (Linux) and open source firmware (OPAL) - virtualisation overhead and density - does it hurt performance to run in a VM? - can you squeeze lots of VMs running MySQL onto a single machine?

Speakers

Stewart Smith's picture

Stewart Smith

Linux Kernel Developer, IBM

Biography:

Stewart currently works for IBM in the Linux Technology Center on KVM on POWER and OPAL (the OpenPower Abstraction Layer) firmware, giving him a job that is even harder to explain to non-Linux geek people than ever before. Previously he worked for Percona as Director of Server Development where he oversaw development of many of Percona’s software products. He comes from many years experience in databases and free and open source software development including MySQL, MySQL Cluster and being one of the founding Drizzle developers. He’s often found hacking on free software, taking photos, running, brewing beer and cycling (yes, all at the same time).

Share this talk


Slides