LWN.net Logo

The real cost of open source (FCW.com)

This Federal Computer Week article covers the adoption of open source software by the U.S. Marshal Service. "For the past few years, the Marshal Service has been replacing SCO Group Unix with Linux in some back-office systems. Earlier this year, officials began implementing JBoss Web application servers, another open-source package, across the agency's 94 district offices. Traditional commercial alternatives would have cost $50,000 per processor in software licenses, and "that would have been cost-prohibitive," Campbell said. "JBoss is free upfront; we only have to pay for maintenance."" (Thanks to David A. Wheeler)
(Log in to post comments)

Scary Didio

Posted Nov 30, 2004 23:25 UTC (Tue) by bojan (subscriber, #14302) [Link]

Didio notes, quite correctly, that:

"Those [organizations] that are having the most success with Linux are the ones that are very self-reliant with in-house expertise."

The corollary to this statement is, of course, that one needs to become self-reliant in order to be successful with open source. Gee, that's sounds like a frightening thing, doesn't it? To be confident and knowledgeable enough to make things work the way you want them to work. Or worse, to accumulate in-house knowledge, therefore serving your "customers" better. Oh no, I think my head is going to burst of all this independent thinking...

Nah, I think it's much better if we all stick with the proven plan. Let someone else decide what we really need and charge us a fortune for it. Yep, a better thing to do, especially long term. Please forgive us Bill, for we have sinned (i.e. used our brains for a change ;-)

Once upon a time, there was an "IT budget"...

Posted Dec 1, 2004 5:18 UTC (Wed) by kirkengaard (subscriber, #15022) [Link]

Positive a troll though it is, I'll burst the bubble here. Consider that the good old days of in-house IT staff have been over for many corporations, especially smaller ones, for a few years now.
1) You have to pay people money for something tangential to your primary business objective; the more dedicated you want them to be to your systems integration plans, the more money you have to spend on them over time.
2) Business managers know, generally, two things: their business and their market. If their business isn't tech, they know what somebody else tells them, and they judge that advice the same way they judge anything else: by the reputation and track record of the messenger. "Very self-reliant with in-house expertise" doesn't just happen with most companies.
3) Everything, eventually, costs something, and they know that, too, sometimes because of painful illustrations. Business lives and dies on Return On Investment. Investing in staff, rather than 'outsourcing' the software work to someone like dear old Bill, or (increasingly) the guys in the Red Hats, can be very expensive, and has a generally nebulous return. Personnel is an area full of hits and misses, and picking geeks is a dark art for many small HR staffs (if you *have* an HR staff; see 1a, above). Packaged software systems may have a high price tag, but the investment meshes with the mindset. When the system you bought does what it's supposed to, and somebody else deals with the problems, the return on investment tends to look clearer from a managerial standpoint.

Who do we hear successfully adopting open source, by and large? Governments, government agencies, scientific groups, academic institutions, large corporations. What do they have? Resources to spare - of one sort or another - that apply to the problem. "In-house expertise" and "self-reliance". I don't mean to be the bogeyman here, but snipes about how all it takes is independent thought don't win the argument. You have successfully identified the 'corollary', but you miss what a big deal that self-reliance can be, and the obstacles to getting there. We need to clear those roadblocks, not just laugh at them. Think of small business as an offshoot of the "how do I get my grandmother to like using Linux" problem.

Once upon a time, there was an "IT budget"...

Posted Dec 1, 2004 5:56 UTC (Wed) by bojan (subscriber, #14302) [Link]

> We need to clear those roadblocks, not just laugh at them. Think of small business as an offshoot of the "how do I get my grandmother to like using Linux" problem.

I definitely agree with you - we need to clear the roadblocks. And they are being cleared daily by distro makers and upstream developers, sometimes even by "oursourcers".

My point was that with open source, businesses get to choose who they want to be and how they want to run their business. They can afford a customised system, because the base of that system is so cheap they can afford a bit of customisation on top. I have personally provided some small businesses (4 - 6 people) with fully customised tools that are literally running the core of their business. And they didn't go broke as a result of it. The reason I could do it is open source - my building blocks cost me (and them) nothing. So, they only had to pay for a bit of my time and that was it.

Better yet, the customised system is theirs, to be enhanced by someone else (if they so choose), because they received all the source licensed under the GPL. So the "self-reliant" bit doesn't necessarily mean that you have to have a hacker on staff. It means that you get your customised system designed to meet exactly what you need, but without the high cost and with its future guaranteed.

To make a (probably clumsy) analogy, if "bricks" are free, one can build a much more elaborate house for the same amount of money.

Once upon a time, there was an "IT budget"...

Posted Dec 2, 2004 1:28 UTC (Thu) by kirkengaard (subscriber, #15022) [Link]

> I definitely agree with you - we need to clear the roadblocks. And they are being cleared daily by distro makers and upstream developers, sometimes even by "oursourcers".

Heard and understood; I implied, but perhaps not clearly, that outsourcing works as well to Linux distributors who do the work as it does to MS, if not better, but you have to get them to see that, which is market share, our slow uphill battle. From DiDio's point of view, the little guys and the system customizers like you that serve them are off the radar. However, you haven't described how they are self-reliant now. They may be set for now, but unless you trained someone up to maintain the system you set up, they will, as you suggest, have to find you - or someone like you - again when they need. Self-reliant means what it says. A sound structure is not a necessary consequence of free bricks. A pile is a higher entropic state, and therefore a more probable consequence.

Once upon a time, there was an "IT budget"...

Posted Dec 2, 2004 2:51 UTC (Thu) by bojan (subscriber, #14302) [Link]

> However, you haven't described how they are self-reliant now. They may be set for now, but unless you trained someone up to maintain the system you set up, they will, as you suggest, have to find you - or someone like you - again when they need. Self-reliant means what it says.

I think you are still misunderstanding my point to some degree. They are self-reliant in the business sense - they know what they want (i.e. what differentiates them from the rest of the crowd), they know how to get it and they went for it. The reason a small business could do that is that they never got slammed with $50,000+ licence fees that would be associated with the base for thier customised software (just think various J2EE containers, for instance). So, they could afford a few grand of custom work instead and now they are self-reliant - they have their own solution nobody else does.

It is not possible to run a business - any business - in complete isolation. You always use other people to do things for you. Be it an electrician, a plummer, a copy-centre or something else. You want to focus on what you do and then you want to deliver it your way. That's what makes you as a business interesting. If you can do that, you are self-reliant - you don't have someone from the outside determining your business internals by forcing you into using pre-done, same-for-everyone software.

With the tailored solution, your internal processes are directly reflected in your software. With a canned solution, you only get what was already there and you have to submit to its ways. Not to mention that, given we're talking about proprietary software, you have no way of getting this thing from someone else when the time to upgrade comes (and it will).

With open source, you can ditch the original guy (me), find other people that understand the same thing (even employ some if you grew in the meantime), give them the source and have another set of cusomisations done to it. When you spend big on licences you can't do that any more. And, you don't have the source for your building blocks, so you can't do anything with that either.

Don't assume that "accumulation of in-house knowledge" means something related to programming or software directly. It has to do with what you want to do as a business. It just so happens that open source lets you be self-reliant on your own processes and methods to turn your business-think into customised software cheaply, either by "outsourcing" or by employing.

Once upon a time, there was an "IT budget"...

Posted Dec 1, 2004 6:43 UTC (Wed) by dkite (guest, #4577) [Link]

The shift may be occuring now by skilled in-house staff. It's a natural.
We need a quick solution to a local problem, no budget; bang together a
linux server of sorts on the cheap. It satisfies, so with next issue
Linux or some FOSS solution is considered. Eventually the C** notices,
wants to institutionalize or formalize the thing, buys Redhat or SUSE
support etc.

That market has fed much of the growth up till now.

Next is the more conservative, less resourced. A friend works for a
community college, has a Novell network with Win clients. A couple of
months ago he surprised me by bringing up Linux, lots of good things to
say, and was looking forward to getting some boxes going with Novell
linux. He is run off his feet, usually has to implement other's ideas,
doesn't have time or freedom to experiment. But Linux is starting to
encroach.

This illustrates that the expertise doesn't have to be in-house.

Derek


Once upon a time, there was an "IT budget"...

Posted Dec 1, 2004 13:26 UTC (Wed) by cpm (subscriber, #3554) [Link]

Nice summary; I wanna take a couple of issues with it though

"Business lives and dies on Return On Investment."
While this is certainly explictly true, it can be said also;
that Business thrives or fails on adaptation and innovation.

Richard Branson has said, and I quote "The time to get into a business is when it's abysmally run by someone else".

On the one hand, nearly any service, any product, any level of "help" in the form of business, life, legal, health and such consulting
can be purchased from folks doing it the so-called proper way. The ugly truth is,
most of them suck at it. Bill Gates was once heard to say that "Any kid working
in a garage with a good idea can put us out of business."
This is certainly true, at least in spirit, hence Microsoft's
vicious anticompetetive business practices. In view of all of these things, and many others, yes.
If your goal is to maintain the status quo, lay everyone off, outsource everything, take all measures
to prop up your stock prices, try not to get caught by the justice department doing so,
and so on. Then yes. Doing things "by-the-book" as taught to all MBAs, is the right thing.

If, however, your goal is to build and grow a business that is competitive, that innovates towards that end, that is healthy, that can stand scrutiny, then you must do things differently, in order to "get ahead".

Once upon a time, there was an "IT budget"...

Posted Dec 1, 2004 18:26 UTC (Wed) by ordonnateur (guest, #6652) [Link]

The time has come to add up the downside of 'outsourcing'. True it takes time. a bit of luck. and skill to build an in-house capability. The benefit is that old-fashioned thing: loyalty and dedication. Build a tech team you can trust and get advice that really does have the interest of the organisation and its purpose at heart.

Once upon a time, there was an "IT budget"...

Posted Dec 1, 2004 21:04 UTC (Wed) by emkey (guest, #144) [Link]

And higher management that is willing to listen to them. Thats the tricky bit. Far to often the technical side of a business will be making recommendations that are key to the long term success of a business and getting nowhere, even in situations where those recommendations make good business sense. This is due in large part to many managers distrust of things they don't understand.

Of course it is also due in part to technology peoples sometimes poor track record in regards to delivering what they promise.

Once upon a time, there was an "IT budget"...

Posted Dec 2, 2004 1:34 UTC (Thu) by mbg (subscriber, #4940) [Link]

While it may require expertise to hire good IT staff, I'd suggest that it also requires expertise to purchase and manage packaged or outsourced IT, i.e. not get ripped off.

While we're at it, why not contract a consultant to help with your IT staff selection?

The real cost of open source (FCW.com)

Posted Dec 1, 2004 6:33 UTC (Wed) by ninjaz (guest, #2083) [Link]

Unfortunately, this article passes off the standard FUD as "valid concerns":
"With open source, who's going to support the hundreds of thousands of users?" asked Quazi Zaman, platform technology specialist manager for Microsoft's federal division, based in Washington, D.C. "With commercial software, end users have direct vendor support, third-party systems integrators and help desks. Then there's the training piece. How am I going to reduce enterprise costs if I have to get thousands of people up-to-date in using open source?"

Hidden costs

Those are valid concerns that are certain to grab the attention of IT managers squeezed between tight budgets and growing end-user expectations.

This ignores the fact that if open source is brought in as a standard in a company (or division, etc) environment, vendor support is available from SuSE or Red Hat as it would from Microsoft. It also propagates the false dichotomy of Commercial Software vs. Open Source software (eg., Red Hat and SuSE are both) Regarding helpdesks, hopefully a company deploying Open Source would hire or cultivate internal helpdesk expertise for their environment. Third party integrators? Hasn't anyone heard of IBM Global Services? ;)

Granted, any time you switch software, you need to adjust to supporting it (whether it be a new version of Windows, Office, SuSE, etc) The fact that you'd have to migrate (during a migration...) is a valid concern, but what was stated in the previous paragraph was clearly false and has been for some years.

The real cost of open source (FCW.com)

Posted Dec 1, 2004 16:05 UTC (Wed) by vmole (subscriber, #111) [Link]

"With commercial software, end users have direct vendor support...

Now, that is a complete load of crap. I used to work for a company that had some fairly large Microsoft developer licenses, along with the usual Windows, Office, etc. licenses. None of them allowed end users to contact MS: it all had to be channeled through a few designated individuals. Most commercial software support works this way - and for good reasons.

The real cost of open source (FCW.com)

Posted Dec 1, 2004 16:19 UTC (Wed) by iabervon (subscriber, #722) [Link]

It seems to me that those are valid concerns, in that you ought to worry about them and only be satisfied after doing research to determine that they aren't actual problems. It's not that they're dumb questions (although the claim that end users of commercial software have any help from vendors in laughable), it's that there are good answers ("IBM, Red Hat, or Novell" and "most people won't notice much of a difference"). I particularly like how, later in the article, actual end users (rather than marketting types) give those answers.

The real cost of open source (FCW.com)

Posted Dec 1, 2004 9:30 UTC (Wed) by Seegras (guest, #20463) [Link]

What I don't understand: Why do they always ask Microsoft? Microsoft most probably does not know anything about cost-savings that can be achieved by using non-Microsoft products; and Microsoft of course has an interest to portray all non-Microsoft-solutions as inferior in one way or the other. So you'd really be stupid to ask "Quazi Zaman, platform technology specialist manager for Microsoft's federal division" about Linux and expect anything but Linux-bashing marketing-drivel.

The real cost of open source (FCW.com)

Posted Dec 1, 2004 15:12 UTC (Wed) by hppnq (subscriber, #14462) [Link]

Old news!

But still very amusing. ;-)

Copyright © 2004, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds