Public or private, which is cheaper? A ever-present, hotly-debated topic but one that often lacks substantive data. In this session we'll leave the politics aside and focus on the facts!
A major data processing company was recently trying to determine what cloud model would meet their needs most cost effectively: public cloud, or private hosted OpenStack. After much searching, they discovered that there really isn't much data available on the topic. They asked us to look at their workload and make a recommendation.
What we found surprised us. In this sure-to-be provocative session, we'll dive into the following topics:
- Unit cost economics
- Reservations: paying for capacity vs paying for performance
- Urban sprawl / orphans
- The "feature black hole"
- Financial modeling
We'll present our findings - which will surely prompt a healthy debate with session attendees!
Unit Cost Economics:
- Capacity reservation versus performance
- Hardware utilization
- Thin provisioning
- The law of large numbers
Urban Sprawl
- A day in the life of an application
- Orphans
- Sprawl... by design?
Feature Dependency
- AKA the black hole
- AKA cloud vendor lock-in
- AKA getting stuff done faster!
Financial Modeling
- Building a model
- Day 1 cost analysis
- Years 1-5 trend analysis
Other Factors
- Performance
- Security
- Customization
- Ease of use
