[TriLUG] ROI moving gov to Linux?

[John Vaughters via TriLUG]

>Why anybody starts a project with Microsoft's SQL is beyond my capability to understand. >Expect a lot of push back from those who only know Microsoft. 

You are toooo kind. I'd say expect temper tantrums if you try to take away MSSQL.
I have offered to save tens of thousands of dollars by moving to MariaDB (MySQL) and I have been met with pure obstruction in some cases. Both Oracle and Microsoft have the worst licensing rules I have ever seen. I personally have no use for them ever. Oracle audited us and we did ok, but had to restructure a few licensing issues, nothing out of ordinary, but they hit us up a few years later because we were downloading their virtual software by many of our individual users that just wanted to try it out. But they tried to strong arm us into saying it must be being used for commercial purposes and that we owed them money. We basically told them to shove it and don't worry, because none of all the users that downloaded it thought much of their product anyway, so don't worry we won't be using it and we will be purging as much Oracle as possible in the coming years.
But the Microsoft users of MSSQL, which are mostly clueless about any other product and actually think MSSQL is bar none the greatest product ever and that anything open source must be worthless. "How could a FREE product be any good anyway?", they say. Not knocking MSSQL as it does seamlessly integrate to many products by default, but we have archiving systems that do nothing but collect data and have ZERO integration requirements. And the MS folks refuse to see that paying for expensive complicated licenses as a bad thing. Out of about 20 internal customers only 3-4 use linux with the free archiving. 
It is an uphill battle all the way. But I keep pushing anyway, because I may actually get to the top of that hill one day. The above comments are from the past decade or so and actually we have made some progress over time.
But good luck in your efforts with MS folks. 
One thing I liked about the Largo story is that they used Citrix for the MS products that were required. That sounds like an awesome solution. The main point is there is a size that matters to save money. You need people with the right skills to make it work. You need to have a few highly skilled, ie. higher paid, people to analyze the whole solution and come up with the path, or you risk complete failure. Consider some good consultants and some training for the existing folks if you ever get the opportunity. It's most likely going to require restructuring of the software and the way people work. Not an easy task. 
My personal experience of getting people to try Linux is generally mixed. If they mostly browse the internet, I have had good experience getting them to use Ubuntu. If they like Office products, LibreOffice is like poison to them. They would rather walk through the brier patch. Me personally would prefer LibreOffice, but I understand the transition sucks.

All in all it's a tough sell and you need a strong leader to just lay it out and tell everyone that it is happening like it or not. That could be your biggest hurdle. Otherwise the door for complaints is wide open and fall guys are being located.
Good Luck!!!
John Vaughters