Well, I can see where this is going, and IMHO it's not the solution that fixes all - it never will be.
Imagine what happens when developers on their own system start kicking each other out of the development RDP sessions because of their continued behavior that caused them to be thrown out of the actual QA platform in the first place. Add to this the fact that in many regions, (unproductive) developer time is very - I say, very - expensive.
The logic goes a long way: each QVS system (PROD and QA) suddenly starts a fight for resources with QDS. So you split them off. In the end you have 5 or more systems where 2 with a decent amount of cores, RAM and common sense should be sufficient.
It think operating a QlikView Enterprise solution should be a balanced.act: 1. proper configuration, 2. continued platform monitoring and management and 3. discipline. Without the latter in the form of a set of sound rules, every architecture will lead to a complete lock-up sooner or later. And as you said before in another comparable discussion on the same subject, Janaki: you can't keep adding RAM, CPU's or entire systems to the setup forever. At least not without decent scaling & sizing arguments.
One final tip: try to eliminate everything that is unpredictable. The cost of adding a QlikView document or developer should be known before they arrive on the scene. It will make it easier for you to plan (and budget) the growth of your QlikView environment.
Some background information for this post can be found here:
http://community.qlik.com/thread/148298
http://community.qlik.com/thread/148299
.png "Former Employee"){kind=icon}
