How we reduced the launch of a 60 node production cluster on the cloud from 3 days to 2 hours.
Submitted by Sri Harsha (@sriharshachillara) on Friday, 18 April 2014
Section: Crisp talk Technical level: Intermediate Status: Confirmed
To discuss the approach and the processes followed around system topology, configuration management, database versioning and automated deployment to optimize the setup time of a new cluster.
For a SAAS based B2B company, as the business grows into multiple geographies, there is a requirement of deploying production clusters into data centers across the globe. This poses a challenge of uniformly deploying a large application stack, with several dependent components, the major issues being managing system topology, system configurations, database versioning and infrastructure setup. These problems are common to all cloud providers like AWS, Google Compute Engine, Rackspace etc. The processes, practices and concepts that we followed in tackling these will be applicable to any SAAS provider.
The session covers our learnings and experience tackling these challenges.
I'm a developer at Capillary Technologies.