Floschi’s Blog

or: Why does Munich’s switch take some time?

First: This column is my personal opinion and not necessarily the one of the Municipality of Munich.

Short answer:

There is not just a technical switch from proprietary to open standards and free software, but also a general improvement and reorganization of Munich’s IT. A reorganization to centralized IT services for our linux client. Nowadays we’re doing much more than planned in 2003, to gain an efficient and sustainable IT structure, based on open standards and free software. That’s a long-term strategic objective, not just related to free software.

Long answer:

Munich’s IT diversity

Munich’s IT history is very heterogenous. Munich’s IT as faced by LiMux in 2003 consisted of 21 independent IT units, every single one responsible for its IT operation. Different grown – and locally quite optimized – processes, tools and specific trained staff. 51 IT operating locations (small and big datacenters), about 1.000 IT staff for 33.000 employees. The technical diversity is a small mirror of the world’s different IT solutions. No common directory, no common user, system or hardware management. Different tools for software distribution and system management. More than 300 apps, many of them redundant, e.g. using Dreamweaver, Frontpage, Fusion etc. for HTML-editing. 21 different Windows clients, different patch levels, different security concepts. This was Munich’s IT situation when LiMux started.

Realizing “digital waste exit challenges”

We did a miscalculation in the beginning, yes. We were naive. LiMux tried to offer a single linux client which fits into every different environment inside the IT units. Theoretically possible, but this would have meant to strengthen diversity and to ignore the opportunity to learn from the past. Why should every IT employee reinvent the wheel again and again, let’s cooperate!

Our adjusted goal is to do technical stuff once and benefit 21 times in the future. A linux client maintained and supported by one unit, also providing common tools for user and system management. A client fitting into a standardized IT backend infrastructure, especially directory and file service.

That’s not a matter of free vs. proprietary software, that’s a matter of efficient and economic IT organization.

We realized that in 2007, when the first departments began to use our linux client as planned before. But in other units the migration stagnated. In many cases, their technical backend structure hampered or even declined cooperation. You know that there a bugs in a big company’s proprietary DHCP implementation answering requests very strange? And other proprietary tools which could not coexist with any other software management solution, because they are not configurable?

The lack of open standards for interoperability and the domination of lock-in interfaces was awful. Really awful. You don’t realize these efforts by just depending on one single vendor, being happy with his software and tools. You know Simon Phipps “the biggest enemy of freedom are happy slaves“? I remember this statement very often when thinking about the past. A lot of digital waste we once produced we now had to get rid of. Gartner calls this “exit costs”, I think about “digital waste exit challenges”.

Change of strategy: pilots and broader IT reorganization

So we changed our strategy from 2008 onwards. At first a pilot phase EVERYWHERE, to gain information about the degree of diversity. Every unit had to establish at least 50 desktops or 10 percent of their PCs with the LiMux Basisclient. Together with its common infrastructure, allowing but not neccessarily enforcing standardization. The second step for us should be to learn, to learn, to learn and optimize. Then to finish the rollout knowing the best practises and lessons learned.

In parallel the City Council decided on a extensive reorganization project for the whole IT in process and hierarchy matter. Not directly related to LiMux, but with many points of contact, especially whenever we’re talking about technical solutions. This is ongoing and helpful for LiMux, looking to IT improvements in a broader view than just the technics.

Yes, we can! Successfull pilot phase finished in 2009

Since the end of 2009 we have proven that our LiMux Basisclient (linux client) is able to be fully integrated into this heterogenous environments. We successfully finished the pilot rollouts. In total 3,000 linux clients throughout the city; an enourmous number of linux clients. Remember, the goal was to establish pilot projects of 10% (1,500) of our desktops. We run 3,000, twice as much. First step done.

By the successful switch to the open standard “Open Document Format” (using OpenOffice.org on all desktops) we cut off the plenty lock-ins from one (business) app to one single office suite, only available for – you got it! – one single operating system. Now we are free to chose whatever we want! I blogged about this in my 2009 review.

Nowadays we are doing the planning for optimization. We know we can do the migration, so we want to be more efficient than in the proprietary world before. Let’s learn and build a better IT. Then the next years will be the rollout years in Munich.

LiMux has a long-term agenda

Yes, LiMux has a long-term agenda. We could have switched to linux clients in just a few months, giving the order to all 21 IT units to set up a linux client until end of 2008. No further specifications, no standardization and no consolidation. I’m pretty sure they would have done this excellent and then I would have published great news in 2007 or 2008 “LiMux done, Munich completely on free software”. But if we would have done this we would have ignored this big opportunity for Munich’s IT as a whole. Quality over time! Not related to free software, but neccessary for cleaning up our IT.

We never ever will be happy slaves again

I won’t excuse me for being clever and adjusting the way to achieve better goals. Digital sustainability is a long-term effort and not only a matter of Linux vs Windows. It’s not a matter for or against Microsoft. There are many vendors trying to lock you in. We learned it and do our homework. We never ever will be happy slaves again. You, too?

I hope clarifying our approach helps to understand that LiMux is more than just a technical question. We are doing our homework and invest in the future of IT openness. Successful, as proven by our ODF switch and the many linux clients we’re using every day.

Cheers,

Florian

First: This column is my personal opinion and not necessarily the one of the Municipality of Munich.

OK, let’s start my FAQs with this simple question:

Why did the Municipality of Munich chose to use free software?

Short answer:

This was a political decision for independence, freedom of choice and a free market. The decision based on facts.

Long answer:

1. EOL for MS WinNT4

Microsoft company announced the end of life/support for Microsoft Windows NT4, which was widely used on the administration’s desktops. So the Municipality, like every other public authority, had to examine the next step.

2. Examining alternatives is a legal requirement

The City Council asked the administration the show alternatives. This was a demand of the german law. It’s unlawful to chose one solution or provider without examining alternatives. The administration decided to take this question serious and prepared a preliminary study together with an external consulting company.

This study examined three main factors of five different alternatives (ranging from proprietary software for operating system and office suite to free software for both, including some mixtures of both). The study answered the questions about

• Is it technical possible to use the alternative?
• How much does it cost to introduce the alternative?
• What are the strategic advantages and disadvantages of the alternative?

3. Let the politicians decide

Finally, the result was a draw of two examined alternatives. A draw in total, that means not in every detail aspect. The perfect situation to balance pros and cons, to discuss them and to focus on the things the coucil wants to stress. Great.

One of the two equal alternatives was based completely on proprietary software. This alternative had lower overall costs in the short term, but chosing this would have meant to tighten the dependencies on this proprietary software vendor. So the lower short term costs faced high long term costs. And from the strategic point of view, other products of the same vendor should have been introduced to ensure security and operation of the desktops.

The other equal alternative was based on free software. Of course, changing the architecture meant higher costs to invest in the beginning. But it was clear, that this was the only possible solution if we want to escape from the vendor lock-in.

The administration presented this draw to the members of the City Council in 2003.

4. Political decision for independence, freedom of choice and a free market

On May 26th in 2003, the City Council decided by majority of Social Democrats (SPD) and Greens for the long term strategy of the administration’s desktops. The main objectives to chose this alternative were

• Increase the independence from software and hardware vendors, consultants, external timelines
  This means to get the opportunity of choices.
• Reduce the hurdles for small and medium sized companies to take part in the software market
  More competition means higher quality for lower prices for all of us. Monopolies hamper this.
• Reduce the IT costs in the long term by deciding on our own.
  Of course an investment has to be made to get rid of the vendor lock-in, the so called exit costs (Gartner). But after the change, we all benefit from the opportunities of choices.

There were two decisions. The first one knows everyone:

• Free software and open standards for the desktop operating system and the office suite.

The second isn’t known as much as the first one, but also a very important one:

• Business (legacy) applications should be procured platform open in the future.

This long term goal means learning from the past and reducing the vendor lock-in again by applications which doesn’t take care of the operating system and office suite the user uses. Communicate with open standards and don’t force the user to a specific operating system or office suite.

5. Successful verification of the decision

The City Council decided and asked the administration to examine this alternative more in detail, before spending any money. This was the detailed conception phase in 2003 and 2004. Together with another IT and consulting company the administration re-examined the chosen free software way. And the result was quite the same. So the Council decided in 2004 to start the migration and authorized the money for the two times independently estimated costs.

My personal conclusion

This shows clearly that Munich’s decision is based on facts. Facts that have been revised by two independent consulting companies, by the administration, by the office for administrative audit, by the audit court of Bavaria.

So it’s not a kamikaze mission by some crazy free software enthusiasts, as blamed to me many times (e.g. by limuxwatch, imho the Microsoft paid anti lobbyist blog). But of course, it was a political decision which had to be made in 2003 and I admire the members of the 2003 City Council for this courageous decision against all lobbying attacks. It was the right decision during a time, when many other public administrations didn’t event think about alternatives, which was completely illegal, but common these old days.

I will continue this “LiMux question” column everytime when there’s some time left. Please feel free to send me your question.

Florian

Hi everybody!

There are so many questions asked to me during my talks and most of them are the same topics. So I decided to start a column here to answer them in a unsteady way, when there’s some time left to me.

At the moment I plan to answer:

• Why Munich chose free software
• Why it takes such a long time

Please feel free to ask me questions you’re interested in. I won’t guarante to anwer them quickly, but I won’t ignore constructive-minded questions.

All will be released in the category “LiMux questions“.

Yours,

Florian