Linked by Thom Holwerda on Mon 27th Nov 2006 10:55 UTC, submitted by Jean Claude
Linux French députés' offices will be equipped with a Linux operating system and open source productivity software. There will be 1154 French parliamentary workstations running on an open source OS, with, Firefox and an open source email client.
Thread beginning with comment 185788
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[2]: Nice...
by elsewhere on Mon 27th Nov 2006 19:53 UTC in reply to "RE: Nice..."
Member since:

Munich, one of the largest yet announced their ~45 million dollar 14,000 PC migration back in ~2002 and with any luck they should have about 500 machines on Linux by the end of this year.

Personally, I think the fact that migrating more than 400 machines over to Linux still makes the news is pretty depressing.

Let's keep it in perspective though. The actual deployment only began in September and is scheduled to roll out through 2009.

The decision in Munich wasn't a kneejerk reaction based in anti-MS sentiments, and it wasn't all about eliminating license fees. The objective was to remove long-term dependencies on proprietary software, and the city went through a proper consultative and analysis stage before making any decisions. This included studies of the software used, the viability of open source replacements, useability, resource requirements etc. IBM and Novell/Suse were both involved in this stage, providing expert guidance on the integration of linux and OSS applications in enterprise network environments.

This was followed by a testing phase. In addition, decisions were made on software and they opted to roll their own distro. A service was designed to facilitate imaging, management and application deployment to simplify the rollout and ongoing maintenance.

In short, Munich did their due diligence. Organizations with large infrastructures cannot make whimsical decisions concerning platform migrations etc., these things need to be studied, considered and re-studied to the nth-degree the determine impact across all areas of the business (or government in this case). Large companies will go through a similar process just determining how to approach upgrading from NT to XP, or XP to Vista.

The process in Munich was also sidelined for a period during the whole software patent issue, until the government was confident that their use of linux and other OSS applications would not be impacted.

Linux deployments need to be looked at realistically, perhaps smaller organizations can shorten that period of time and start migrating but larger organizations don't have that luxury. There is a considerable amount of analysis, preparation, training and testing that needs to be done.

Munich should be a blueprint for other organizations, particularly at the public level, in managing their own studies. In fact, much of the groundwork that was done for Munich can benefit other organizations or institutions considering a similar migration.

I've never believed linux will be a wholesale replacement for Windows in enterprises, but I've always maintained that when viewed with an objective set of standards, linux is surprisingly adaptable and suitable for a variety of deployments. I'd rather see strong successes that proceed at a glacial pace than fast "look at me!" type headline-grabbing deployments that ultimately fail due to lack of understanding and preparation.

The fact that Munich as made it this far and has begun actual deployment is a significant success, in my view anyways.

Reply Parent Score: 4

RE[3]: Nice...
by NotParker on Mon 27th Nov 2006 19:57 in reply to "RE[2]: Nice..."
NotParker Member since:

Munich also chose to spend 12 million more on Linux than on Windows.

That sends a clear lesson to places like Birmingham (

Linux is way more expensive.

SUSE was headquartered in Munich at the time. Thats the only reason Linux won. Local politics.

"Documents obtained by USA TODAY show Microsoft subsequently lowered its pricing to $31.9 million and then to $23.7 million — an overall 35% price cut. The discounts were for naught.

On May 28, the city council approved a more expensive proposal — $35.7 million — from German Linux distributor SuSE and IBM, a big Linux backer."

Edited 2006-11-27 20:01

Reply Parent Score: -1

RE[4]: Nice...
by Almindor on Mon 27th Nov 2006 21:19 in reply to "RE[3]: Nice..."
Almindor Member since:

You forgot the fact that from long term they went with FAR FAR cheaper deal.

Reply Parent Score: 1

RE[4]: Nice...
by Johann Chua on Sat 2nd Dec 2006 05:30 in reply to "RE[3]: Nice..."
Johann Chua Member since:

I'm sure you enjoy your checks from Bill Gates and/or Steve Ballmer very much.

Reply Parent Score: 1

RE[3]: Nice...
by NotParker on Mon 27th Nov 2006 20:06 in reply to "RE[2]: Nice..."
NotParker Member since:

The double tragedy of Munich is not only the extra 12,000,000 its that only 80% of the desktops can be moved to Linux. 20% will stay with Windows XP to run applications like AutoCAD.

That will cause extra costs to run two OS infrastrutures.

Munich is great lesson for business. Linux is expensive.

Reply Parent Score: 1

RE[4]: Nice...
by elsewhere on Tue 28th Nov 2006 04:39 in reply to "RE[3]: Nice..."
elsewhere Member since:

Munich is great lesson for business.

I agree. Let's take a look.

only 80% of the desktops can be moved to Linux. 20% will stay with Windows XP to run applications like AutoCAD.

Hmmm. An organization with 14,000 desktops decides to do a full scale analysis of their infrastructure, business requirements, organizational objectives and future needs assessment. They discover that out of those 14,000 desktops, only 2,800 of them actually *require* a rich-client OS like Windows for mission-specific applications. Interesting.

That will cause extra costs to run two OS infrastrutures.

Maybe. But then based on the fact that Windows is only required on 20% of the desktops, this implies that Windows and a Windows-based software stack are possibly being used unnecessarily on the remaining 11,200 desktops.

So then the smart thing to do is look at the licensing costs of the software for those 11,200 desktops compared to alternatives, whether free in cost or not. In addition, we need to compare the expense of the necessary Microsoft server infrastructure and management utilities incremental to those additional clients.

The LiMux distribution they are using for those 11,200 clients is free. As is the core software running on it. In addition, these clients will connect to linux-based servers that do not require client licenses at additional cost. They will be centrally managed by an OSS-based application that does not require incremental client fees. None of the software involved in the core stack will require annuity software maintenance fees. So that's something to start with.

It's possible that even with the non-existent license costs of their chosen client and the necessary support infrastructure to support 11,200 desktops alongside the 2,800 platforms they are locked into, it could still be cheaper to simply purchase the necessary licenses, CLA's, servers etc. Or it might not be.

Of course license cost is only a small part of the equation. As are annuity fees for maintenance contracts etc. I guess it boils down to whether linux can be deployed and managed at a cost equivalent to or less than a Microsoft solution.

I suspect that maybe it can, depending upon requirement. After all, I've been using it daily for a good couple of years now. No games, no high-end multimedia editing, I don't need that for work. Just good old fashioned document processing, email, web, connecting to MS shares on our corporate network, citrix, that sort of thing. Believe it or not, I installed it and it just works. I don't even have to recompile my kernel every week or write my own software patches or execute a cryptic array of CLI commands every time I plug a thumb drive in. I know, shocking.

So it's entirely possible that, looked at *objectively* (sorry, I'm not intentionally trying to use terms that are over your head, here's a link that should help explain that term:, a properly managed and deployed linux desktop could be useable for Joe Average whether in the civil service or simply doing admin work at your average enterprise. I know that in my own company, our IT department doesn't force employees to install their own OS or software, doesn't expect them to edit the registry or download applications or access online forums for support in order to get their desktops working. I can only assume other organizations are similar, and that IT departments would not expect users to install or configure linux for themselves. Maybe I'm crazy.

Of course, neither my opinion or yours really matters. Opinions have no place without some degree of empirical data to back them up. I would certainly hope that before the City of Munich would commit to this, they would perform studies and tests coupled with a proper cost-benefit analysis weighed against their objectives and requirements before making any decisions, the same way any smart well-run organization would operate.

Oh, wait a minute, I almost forgot. They did that, didn't they?

Linux is expensive.

So is Microsoft. So are Porsches. So is universal health care.

Concepts of cost and expense are relative. You keep harping on the fact that Munich should have just rolled over and accepted the discounted licenses to save that money.

But they stated right from the beginning that their objectives were not short-term cost reduction, it was long-term. Long-term cost-reduction often entails short-term expense, and vice versa.

So the easy decision would have been to cave in, accept Microsoft's deal, and let it become somebody else's problem down the road when licenses were again up for renewal, or Microsoft wasn't feeling as generous, decided to change their support plans etc. But again, short term budget savings were not the objective here.

There were considerable costs involved. Much of it entails consulting, migration and training. In fact, training is half of the budget and they're already saying based on the initial testing and rollouts that training will not be as big a concern or obstacle as initially projected, based on the test systems and initial deployment.

Your argument about it being more expensive is simply in comparison to Microsoft's license fees. But the costs Munich are carrying amounts effectively to a one-time investment. Once the deployment is completed and the training and support infrastructure is in place, the costs for ongoing maintenance and expansion will be considerably reduced. THAT was the whole objective. It will take some time, and it will take some money, but they have planned for that and their projected returns center around that. That's how business decisions are made.

The same cannot necessarily be said for a Microsoft-centric strategy since they will simply expect to be paid on demand, even more so the further entrenched they get in the network. Hell, this whole exercise could have been unworkable if Windows was maybe required on 30% or 40% of the desktops instead of only 20%. Fortunately for them Windows is terribly underutilized and not really necessary in the vast majority of their desktops.

Change is never the easiest thing in the world, but that doesn't justify resisting it at all cost.

Munich is great lesson for business.

So let's close where we began. I agree with you, I think that Munich will be a great lesson for businesses. It will provide a template and a methodology for organizations considering a similar move. Doesn't mean that it's the right move for everyone, I've never said that or believed it myself, but their success will certainly lower the bar a bit more for others that are sure to follow. Good stuff.

Reply Parent Score: 5

RE[4]: Nice...
by Johann Chua on Sat 2nd Dec 2006 05:31 in reply to "RE[3]: Nice..."
Johann Chua Member since:

Windows is even more expensive when you factor in downtime due to viruses and malware.

Reply Parent Score: 1