Federal IT Managers Leaning Toward Open Source

(Via Potomac Tech Wire) "An increasing number of government agencies are choosing open source architecture when upgrading their information technology systems, according to a new report from Potomac-based Larstan Business Reports. According to a survey of IT managers in the civilian and defense sectors, 63% of respondents either agreed or strongly agreed that architecture with "open" (or accessible) source code should be adopted by their organization. In addition, 60% agreed or strongly agreed that Unix and Microsoft environments can be "expensive to deploy and maintain and can limit overall flexibility."

Comments

[Rob Fay]

That information surprises me. As I mentioned in my blog post that Dave saw, back in 2003 the Department of Homeland Security named Microsoft as the Department’s “primary security provider.”

I am a federal contractor, and despite suggesting to my client that they consider an open source cms solution, for instance, they will not budge from MS. The agency already has deals with MS for its os, databases, support, etc. so it does not make much financial sense to make waves.

While it may save some money lumping in all these products, how much money is spent on security? Much like one is taught to diversity one's 401K, diversifying one's technology choices, while more expensive on the onset, might end up saving money in the long run. Anyone have any studies on this?

[Dave]

[Rob's talking about this post on his blog: http://robfay.com/2005/04/21/fair-access-security-and-innovation/]

I'm not aware of any studies conducted by unbias parties that demonstrate better outcomes for either heterogeneous or homogeneous technical environments. However, I think it stands to reason that relying on a single vendor for anything as variegated as a technology infrastructure is just a bad idea. In my mind, it's like the aristocrats of old Europe: everyone inter-bred until they looked great -- just as a pure-bred animal does -- but they were as dumb as a sack of hammers. I think blind allegiance to a single vendor or technical approach is as damaging over the long-term.

That said, there are some very compelling reasons to commit large swathes of one's operations to a particular technology. Apart from the numerous economies of scale (licenses, support skill costs, etc.), there's also the convenience factor: everything should (in theory) work well together without too much fiddling. I could argue for hours about the validity of these "benefits" over the long-term but they're very difficult to argue with on a short-term economic basis. And as most businesses and organizations operate on very short-term financial cycles, the lure of adherence to a single vendor is all the greater.