Successfully scaling an application requires that you understand the businesses tolerances for data availability, performance, and cost before trying to tackle the challenges of backup, load balancing, and high availability. All too often, development teams are asked to build highly available systems, without justification or an understanding of the business needs for the data. These challenges are presented to Dyn’s DNS and Email Delivery engineering teams on a daily basis, and by asking these questions ahead of development, we’re able to successfully managed multiple terabyte-scale MySQL database instances across our global distributed datacenter footprint — with strict durability and consistency requirements. This talk explains how Dyn sets tolerances for downtime, the recovery point objectives, and recovery time objectives for these mission-critical MySQL servers, and how examining these factors early in the design process help our teams to deliver value to our business and to our customers every day.
Presented by Tom Daly at Percona Live Washington DC 2012 – January 11th, 2012.
