How does SaaS performance / response time compare to an on premise install? I'm especially curious to hear from people who have moved from one to the other, to see if performance improved or degraded.
Can't say for on-premise. But for Saas response and uptime are inline with other webservices. Each on-premise architecture may produce different results, but you can control those. I'm sure your account rep can arrange for a sandbox test environment from the cloud if you want some hands on testing.
We have been very pleased with our On Demand service. For us the key issue is less about system performance and more about support. If we had to go through internal channels to get support it would be very difficult to get simple things done, such as system refreshes, etc. Our experience with On Demand support has been mostly very positive (sure, there are a few bumps here and there).
Agreed - Good point. SaaS support is reliable and responsive as of late (last year or so). Kudos to them.
This is quite difficult to answer. Depending on your hardware configuration for on-premise, you could have better or worse performance than SaaS. We have recently moved from a hosted solution to SaaS, in some areas we improved performance and in others we lost a little in performance. But when we migrated we upgraded to 14.3, converted reports to JasperSoft, and had to remove triggers and other components that were not supported by SaaS.
There are benefits and drawbacks to both solutions. I would not base my decision to move solely on performance. Remember a lot of CA PPM performance issues are not H/W related but configuration/customization related (Audit, Attribute value protection, heavy customization's, poorly designed process/scripts/queries, etc).
Personally I can argue for both on-premise and SaaS (I like both). To select what is appropriate in a particular case I would want to know what the customer objective is for CA PPM now, 2 years, and 5 years down the road, functions they will use in CA PPM, user base size, etc.