While it may seem that moving to a containerized architecture be most logical thing to do, doing it right takes a lot of effort, investment on resources and patience before we can actually reap the benefits. About 2 and a half years ago we undertook an ambitious project in our Datacenter to pool all our resources together to optimise resource consumption, increase fault tolerance and scale faster. In this talk we are going to retrace our journey of setting up a container orchestration framework using Mesos and Marathon and migrating services from production infrastructure hosting almost a million customers to the new system with emphasis on challenges faced and learnings from the migration and the way forward.
- The problems of operating from a datacenter
- The light at the end of the tunnel (The objectives)
- Our Choice of Platform: Why Mesos and Marathon?
- The Grand Plan (Architecture of the new system)
- The much needed metrics
- The challenges and obstacles and the solutions to them
- The benefits we have reaped
- The learnings and the way forward
Sundar Raman Ganesh is an Operations Engineer at Endurance Intl. Group who has experience in managing Enterprise Mail systems and is primarily involved in design and implementation of scalable solutions using container orchestration platform Mesos/Marathon.
{{ gettext('Login to leave a comment') }}
{{ gettext('Post a comment…') }}{{ errorMsg }}
{{ gettext('No comments posted yet') }}