This seems like a questions which starts with an assumption ‘IaaS is the best option for moving user applications to the cloud’ and then seeks justification to the statement.
It’s best to first gather data about current state of “user application” that is a candidate for moving to cloud based IaaS (as against in-house IaaS which more often than not impacts resume of tech engineers, their efficiencies and KPIs).
- What is the current “interactive response time” experienced by end-users?
- What is the connectivity between application’s “home data center” and all it’s users?
- What kind of hardware is used for running the applications? Is it new or aging?
- What is the CPU, disk space and memory utilization trends on this application?
- How does CPU, disk space and memory utilization looks like during peak demand?
- What is the software stack needed to run current state of application?
- What is the frequency of use?
- is the application mission critical / business critical?
- What kind of infrastrcture is in place for backup and disaster recovery for the application?
- What is the connectivity of application to storage devices? SAN, NAS, DAS, etc.
- Collect statistics about IOPS, throughput and latency on storage devices during normal and peak performance.
Once you have these technical and operational metrics available at your disposal, ensure that IaaS of your choice can deliver acceptable performance on all the parameters that are critical for you to maintain at current level to ensure user experience does not degrade.
Now add to this the cost of maintaining this application in-house or a period of future 3 to 5 years (including people skills, software upgrade and licensing costs)
Now compare that with the cost for the similar period for your selected IaaS provider. Add to it the cost of migration and new skills acquistion.
Now factor is benefits of moving to cloud, CapEx v/s OpEx, new technology, no need to upgrade hardware, ability to attract skilled workforce <Cloud Strategy Checklist for Decision Makers - This blog might be a good strategy level read for you).
Once you tabulate your data and compare it, the answer to your question will be staring at you, and it would be obvious.
Jigar