A day with the boys from GLLUG

Thu May 4 09:39:10 UTC 2006

Shane, a thoughtful and interesting post - thank you.

I work at a Fortune 500 company which is a Microsoft shop and although it's difficult to generalize - each company/institution has its way of doing things - perhaps I can offer some advice concerning bringing Free Open Source Software into a traditional corporate environment.

In the year I spent fighting for the right to test and deploy my company's first official FOSS solution (others were running GNU/Linux secretly in closets; I myself prototyped an intranet site on GNU/Linux in 1997), I would say the single biggest obstacle was the fear that there would be "no throat to choke" - no accountability. I successfully countered this argument first of all by choosing among the largest, most active FOSS projects, contrasting these with "orphans" - proprietary solutions my company had chosen in the past which have since disappeared. I asserted that whether we chose proprietary or Free software, we would need an integrator in any case for installation, training, deployment, frontline maintenance, etc. So I carefully chose my integrator, who was able to send serious IT people (with suits and ties :) to meetings and propose reassuring SLA contracts. The technical director at the integrator was co-author of a reference work on the subject; I bought the book and passed it around at every meeting. I also mentioned the widespread IT industry support for FOSS projects outside of Microsoft, since the collaborative development model - made possible by the Internet - has proven itself as a better way to produce reliable, secure software. I insisted that how no matter how many times we tried to choke the throats of our company's Microsoft account managers, they remained incapable of improving the security of their internet browser or improving multilanguage support of their SharePoint product; they could only apologize and ask us to wait for the next version, whenever that might be (!).

I contrasted the difference between software built to standards (W3C, ISO, Dublin Core, LDAP) and designed for interoperability, and proprietary solutions designed to lock in revenue streams whether in the customer's interest or not.

I showed how FOSS can be very easily prototyped, in most cases by simply downloading binary executables and running them in a sandbox. This contrasts with most proprietary solutions, where an integrator designs a pilot project as a step in commitment (human resources, financial, legal) to a bigger project. In this vein I discussed the advantages of avoiding gigantism in our IT projects, and deploying focused projects which reduce risk, minimise business process disruption, and shorten project timelines.

After mentioning that there are 90,000 registered projects on Sourceforge (this was in 2004), I showed a short list of the FOSS titans: GNU/Linux, Apache, MySQL, PHP, etc. I cited examples of very successful companies relying on FOSS, such as Google with Python.

In my company's case, I chose an application which runs on Windows servers as well as GNU/Linux; I explained that this solution could be deployed immediately on today's Windows servers, compatible with existing corporate standards while minimising IT staff impact (very few GNU/Linux experts at my company); at the same time, opening a path to better performance, reliability, and security in the future the day we run the application over GNU/Linux. I added that although all 40,000 PCs and 90% of servers (the rest being proprietary Unices) at my company run Windows, I would bet that the new generation of IT personnel at our company were all running GNU/Linux at home and surfing the web with Firefox.

I avoided the Free versus Open Source discussion, since in my context this was picking nits; I spent many hours explaining that large successful Free software projects are not three guys in a garage, but extremely competent engineers many of which are employed by familiar IT companies. I did however successfully negotiate with our legal department to invert our standard intellectual property clause; my company explicitly renounced any intellectual property claims to software developed for and financed by our FOSS projects, so that it may be made available to the community. I explained that this contribution could be considered our end of a deal in which we got great software for free, created by others who have used and improved it before us. I felt that this aspect was very important; although I explained that the GNU GPL did not obligate us to publish source code since there was no distribution (we are final end-users), I thought it necessary to set that precedent for a FOSS project.

Finally, I demonstrated that for my project, my chosen solution would be at least seven times cheaper than the proposed Microsoft solution, the major differences being in licensing costs and in consulting/development necessary to adapt the MS solution to our needs.

Although I had the support of my business unit managers who were principally interested in the lower cost and faster deployment aspects, I had to endure some very tense meetings at which I was accused by IT executives of recklessness, incompetence, and worse. Several colleagues contacted me privately to say that they hoped I would succeed, although they themselves couldn't stick their necks out. In the end, the corporate IT director was called upon to decide and spent half a day hearing the arguments. He approved the solution as a new corporate standard, saying: "we must be pragmatic, not dogmatic" and "we have gone too far in our dependence on one IT supplier". Prudently, he directed that such projects be limited at first to noncritical databases and informational intranets. Today, FOSS projects are flourishing at my company; the first GNU/Linux servers have appeared in production as replacements for proprietary Unix servers, after tests showed that our in-house Unix tools could be easily recompiled and run over GNU/Linux. And several IT managers are reporting success with virtualization solutions, running Windows server images over GNU/Linux.

One final note: my company (a manufacturer not in IT) has absolutely nothing to gain by publicly discussing our IT infrastructure; you will not find a news article on Google that talks about this. But Microsoft is fully aware that its top-tier customers are testing the waters, and that the industry has changed. The major consulting firms in daily contact with large corporations are also aware of the change. I think FOSS is a clear choice for any company or institution which has to watch its IT budget, and a viable choice for richer corporations who want to improve security and deploy applications more quickly. I believe that corporations can be persuaded to accept the advantages of keeping code Free, in the same way that sustainable development has become a top-of-mind - an issue unheard of ten or fifteen years ago.

Sean DALY.

> Message du 03/05/06 23:54
> De : "Shane M. Coughlan" <shane at shaneland.co.uk>
> A : discussion at fsfeurope.org
> Copie à : 
> Objet : A day with the boys from GLLUG
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
> 
> Recently I was down in London speaking to the Great London Linux User
> Group.  I just posted a blog entry about my adventure here:
> https://www.fsfe.org/en/fellows/shane/communicating_freely/a_day_with_the_boys_from_gllug
> 
> I had a great time, and we touched on some pretty topical things.  If
> you are interested in deployment of GNU/Linux and other Free systems in
> business environments, or in the danger that as Free Software gets more
> corporate we'll cooperate less, you might find something to mull over :)
> 
> Shane
> 
> - --
> Shane Martin Coughlan
> e: shane at shaneland.co.uk
> m: +447773180107
> w: www.shaneland.co.uk
> - ---
> Projects:
> http://mobility.opendawn.com	http://gem.opendawn.com
> http://enigmail.mozdev.org	http://www.winpt.org
> - ---
> Organisations:
> http://www.fsfeurope.org	http://www.fsf.org
> http://www.labour.org.uk	http://www.opensourceacademy.gov.uk
> - ---
> OpenPGP: http://www.shaneland.co.uk/personalpages/shane/files/publickey.asc
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.4-svn4127: (MingW32)
> 
> iQCVAwUBRFkmX9wG3M95JPpzAQgOzAP/YhYXLi9rPllh/Sjjmh7IeqcBOq/xqTUK
> cUhVRr09025zG7I6DRXHuPtor0SMzU2tOz9X8w2R1y4plJntJJ3zm2hMz0CNMM19
> GinPKV8avB7736/OBtARk7Fn24KXYpm2Gtod7Q15N2hROA7Sksr6/TbksXgQQAJV
> IVXsq4g32vI=
> =13KE
> -----END PGP SIGNATURE-----
> 
> _______________________________________________
> Discussion mailing list
> Discussion at fsfeurope.org
> https://mail.fsfeurope.org/mailman/listinfo/discussion
> 
>