Building a Scalable Orchestration Engine to Manage 30k Nodes
Submitted by aaditya sood (@aadis) on Saturday, 30 January 2016
Section: Full talk Technical level: Intermediate Status: Shortlisted
To share the lessons learnt and challenges faced when solving large-scale IT orchestration problems. Hopefully provide best patterns and lessons so people following in our footsteps don’t make the same mistakes.
Sharing the lessons of the 5 year journey involved in building large-scale orchestration software in India and the challenges involved. This talk will cover the evolution from a monolithic model to asynchronous micro-services over 3 generations of our orchestration engine, what real-life problems we solved and the challenges we encountered.
This orchestration engine has been in production for over 5 years across multiple BFSI and other enterprises in Asia-Pac. The National Stock Exchange of India (NSE) uses this engine for IT orchestration, Disaster Recovery etc, with over $25B in daily transactions passing through the applications managed by this system.
This talk will cover:
- The Orchestration Problem - Three Generations of Orchestration in Detail - Lessons Learnt
Co-founder at Calm.io. Previously was the architect at medical imaging startup Erudent. Earlier worked at Yahoo Image Search & Oracle ERP. Has been known to use his .emacs file in self-defense.