The Woes of Munich's Linux Migration

mikrorechner writes "The H Online has a writeup of the problems encountered by LiMux (Wikipedia entry), one of the most prominent Linux migration projects in the world, trying to introduce free software into the highly heterogenous IT infrastructure of the City of Munich. Quoting: 'Florian Schiessl, deputy head of Munich's LiMux project for migrating the city's public administration to Linux, has, for the first time, explained why migrating the city's computing landscape to open source software has taken longer than originally planned.'" Here is Shiessl's blog, in which he details some of the transition problems.

Comments

Because every project is late

[0racle]

Everyone always underestimates how long anything non-trivial is going to take. In this case it seems like not only were they trying to migrate to a new platform, but also trying to undo every past mistake, oversight and quickly implemented solutions that appeared on the surface to work just fine. That's going to take just a little while to get done.

Re:Because every project is late

[Locutus]

but it sounds like most of the problems were due to underestimating how many non-standard development tools and products were used and the trouble getting those over to GNU/Linux. Many of them required either the original vendor to port to an open standard or replacing the existing product with one which was based on open standards. The first option meant that most likely a Microsoft Partner Program member would have to be hired to provide the same product for the GNU/Linux clients. This might normally be an easier option except being a _Microsoft Partner_ often times means you are not allowed to work on other platforms. So the 2nd option is most likely their only choice and that is more expensive in that it would require all users to change the underlying software they currently use for the task.

All in all, this sounds like confirmation that Microsoft's strategy of proprietary API's and patented IP was successful in making it costly to leave their platform. It also shows that it is not impossible and in the long run, it will probably be shown that getting off the Microsoft treadmill might be expensive up front but over time, become very cost effective. Rip and Replace most often ends up resulting in a better, faster, cheaper solution when managed well.

LoB

Re:There is no free lunch

[Sir_Lewk]

Or you know.. buy an open system....

Bad title is bad.

[Statecraftsman]

I recommend to read the blog as it's more informative and it's also rather optimistic. Not just woes as the title would lead you to believe. Of course making the switch to free software takes work, but it's a great opportunity for constant improvement and as Mr. Shiessl points out, there is much digital waste to be cleaned up on exit from the proprietary.

No free lunch, but a range of benefits.

[khasim]

The advantage with FOSS is that Germany can hire German programmers to modify the software used by Munich's government (which is also German).

If they stuck with proprietary products, who would they be paying to improve it?

Re:No free lunch, but a range of benefits.

[Em Emalb]

The advantage with FOSS is that Germany can hire German programmers to modify the software used by Munich's government (which is also German).

If they stuck with proprietary products, who would they be paying to improve it?

This is an insightful post. However, I firmly believe if a US poster made this comment (about the US government) their comment would be labeled a troll comment.

Re:No free lunch, but a range of benefits.

[haeger]

Yes, because it's not like there's a large number of german companies that specialize in windows development and managing windows. That's only something that open source has.

Let's be a little honest about the benefits of OSS please. There are plenty, but saying that proprietary software is bad for the local economy is just misleading.

Re:No free lunch, but a range of benefits.

[h4rr4r]

No, but paying for windows is exporting a heck of a lot of wealth. Proprietary software made in the local country would have this advantage too.

Re:No free lunch, but a range of benefits.

[h4rr4r]

I wonder how much of the "cost savings" on software licenses is being consumed by developer hours recreating functionality.

Even if it consumes every last cent, it would still be a big win. That is money you spent in the local economy not exported, plus it means they are free from future payments for this tech.

Re:No free lunch, but a range of benefits.

[dave562]

So rather than benefiting everyone, they benefit the local economy. That seems sort of selfish, in a nationlistic, protectionist sort of way. (I'm just teasing, I tend to promote local solutions whenever possible). The real meat of my question is what the savings really are. They are spending a certain amount of money on "licensing fees". They are going to stop spending money on licensing fees and start spending it on "in house development". What I'm curious about is the real difference between the two.

I used to do IT work for a local city government here in California. Their building and planning department used a Windows based permit system. It was the same permit system used by government agencies all over the state, and the country. Permits aren't that complex and the functionality could have easily been recreated as a web app, or even a local app. Lets say that Munich uses a similar permit system. Where is the benefit of having their own system that is different from everyone else? The permit system I was familiar with exported to Word (.doc), Excel (.csv) and PDF. It supported OCR of the city's specific forms.

The only way I can see the Munich solution being a win is if they sell or lease their code and applications to anyone else who is willing to pay for them. If they don't, they're simply replacing one system with another. On top of it, they're replacing one system that is used by numerous organizations (and understood by that many people) with a system that only they understand (and is therefore that much harder to find / train people on). I'm not talking about Samba and OpenLDAP and what have you. I mean the day to day applications that support the government processes.

Re:No free lunch, but a range of benefits.

[Maxo-Texas]

The benefit is if they develop an open source permit system, then many people can use it for free.
Many people will contribute to the open source permit system.

If it's based on open standards, then other folks will be able to develop compatible permit systems in the future.

They won't have to buy a copy for version 1992, then version 1995, then version 1998, then version 1998se, then version 2000, then version 2003, then version 2005, then version 2006, and finally for version 2010.

With closed data and closed source- you pay and pay and pay. (and will continue to pay in the future).

And it they go belly up or stop supporting the product, then you are really screwed.

---

All of my personal software stack except dragon dictate is now opensource products that use open source data formats (and support most proprietary formats as well).

When the 2007 versions of office came out- they were damn hard to climb the learning curve (about 5-7 months to get back full productivity and some of my 2003 documents became unprintable-- which I solved by moving them to openoffice).

Munich had a real hairball. At the end of the move, their systems will be much cleaner. And they won't have to rebuy the same software 10 more times over the next 30 years (if the current track record holds).

Re:No free lunch, but a range of benefits.

[Bert64]

But when talking about long term, the benefits of not being locked in to proprietary products are huge.

Re:No free lunch, but a range of benefits.

[tftp]

They are going to stop spending money on licensing fees and start spending it on "in house development". What I'm curious about is the real difference between the two.

A licensing fee, especially one that is sent abroad, is not contributing to the education or employment of citizens of the country. If you hire local developers, they will become good at programming and will be able to design more software later. This is exactly the question of giving a man a fish or a fishing rod.

If you take this situation to the extreme, as an illustration, you can have a country that spends $100M yearly on licensing and still has not a single programmer who can write "Hello, World". This means that those $100M will have to be spent year after year.

Re:No free lunch, but a range of benefits.

[Bert64]

Buying proprietary software doesn't benefit everyone, it only benefits the single vendor of that software (to your own detriment often, as you get locked in)...
Buying locally shifts that benefit away from a single foreign entity, to one or more local entities which is beneficial for government who get their tax revenue from those same local entities.

However, by using open source they are contributing benefits to everyone... Any development they contribute back will benefit everyone, even any bug reports they make will ultimately benefit the community as a whole.

Re:It's like the old saying...

[Anonymous Coward]

Windows costs $$
You can either deploy it yourself or hire someone to do it for you.

OS X costs $$
You can either deploy it yourself or hire someone to do it for you.

Linux is free
You can either deploy it yourself or hire someone to do it for you.

Here's his problem...

[MikeRT]

Previously, around 1,000 staff had been maintaining the 15,000 PCs making up the Munich computing landscape in 21 independent IT centres. There was, according to Schießl, no common directory, no common user management, no common hardware or software management. There were more than 300 applications in use, many of which did the same job. On the desktop side, there were 21 different Windows systems with different update levels and security settings

You can't convert a bureaucracy like this anymore than you can build a political/military empire by invading a dozen good size countries and trying to integrate them all at once. Rome wasn't built in a day. They should have gone in first with the intention of standardizing things, straightening out all of the kinks and quirks each little fief had. All of the file servers here where possible, all OpenOffice there...

Re:Here's his problem...

[oatworm]

By the sounds of things, that's kind of what they ended up doing. Thing is, when you have a hodgepodge like this, you have to standardize on something, and that's going to affect and change whatever is around that isn't already adhering to that standard (i.e. most everything). The problem that they had, near as I can tell, is they decided on the solution before they determined what the problem was - they decided they'd standardize on their LiMux client, then started filling in the blanks. Granted, with a project this size and given the free-for-all nature of their original IT structure, there's going to be a fair amount of blank-filling anyway; if you try to document every single dependency before you go in on something this size, half of them are going to change and mutate before you get done with the documentation, which means you're going to have to update your documentation, which will then be out of date somewhere else, so you'll have to... you get the idea. Even so, it sounds like they got bit by the things they didn't know they didn't know, which happens quite frequently.

problem: poor standards compliance

[Onymous Coward]

More to the point: Moving away from a vendor-locked-in infrastructure is hard.

Any time you build on top of quirks and such that deviate from standardized protocols, upgrading will be hard.

Re:how much did this all cost?

[CityZen]

It's not really possible to asses that. The article really doesn't have much to say about Linux, so much as it was about all the crufty patchwork of multiple systems they were using before. There's a big cost associated with continuing to use the current kludges, though it is difficult to assign hard numbers to, since they come in the form of lost opportunities and inefficiency spread throughout the whole organization.

Moving to any modern, unified system, whether based around Microsoft software or OSS, is a tremendous task for a big organization like that. And without a parallel universe (that made the other choice) to compare to, you cannot really say which choice was better. You can only guess. Sure, you can try to make an educated guess by trying to figure out how much of the legacy applications will still work on the new system without changes, but until you try to actually do that work, you won't know how wrong you were. [99% compatible is worthless if you were depending upon the 1% of things that don't work.]

Re:There is no free lunch

[Sir_Lewk]

Aaaah, I see now. If once piece of software is rubbish, then surely any other pieces of software under the same license must also be rubbish!

With this in mind I think it is safe to say that we can write off proprietary software from seriously competeing in the real world, you would not believe how many stories about proprietary software messing up I can find...

What is that? That's not actually what you were claiming, you were just being offtopic? Oh, I see...

Re:how much did this all cost?

[Josef Meixner]

They hope to save in the future. As a lot of the costs are consolidating their terrible IT landscape it is not clear, what a migration to the latest MS offering would have costs, either. It is not as if it would have been free either, who knows how many of the macros would have broken down when run in a current version of Excel, who knows how many old programs might stop working on Vista (and be it due to a stupid installer). It would have been cheaper, at least probably because a lot would have still worked, but when they write that they found 21 different Windows setups with differing patch levels and security settings, I am not so sure if it really would have been cheaper.

What they probably hope is, that the next migration will be cheaper, the OSS they use won't cost them to upgrade, the costs of the upgrade in work to be done by their IT department are probably not very different when upgrading a Linux solution from a MS solution. But all the work to get their systems closer to a common base might actually make the next big roll out simpler and therefor cheaper.

Problems with linux installations

[bigredradio]

la la la la la la la ...I can't hear you ...la la la la la la la la

Perspective

[Locke2005]

How does this compare to the problems experienced by people migrating 15,000 clients running various Windows releases to Windows 7? Is migrating to Linux more or less costly than migrating to the latest release out of Redmond?

Re:There is no free lunch

[david_thornley]

You know, I have yet to find a closed-source OS that can run everything I want. In fact, there's no single OS that will run everything I want. For my personal preferences, Linux (along with Wine and similar programs) does a good job. For yours, I don't know.

Re:There is no free lunch

[Darinbob]

Reading the fine article, it seems the big problem was not propriety versus not proprietary, or Microsoft versus non-Microsoft. It was that they were trying to fix a very heterogenous and confusing mess with a homogenous consistent infrastructure. All stuff that should be upgraded and fixed, but doing that takes time and effort and often isn't worth the hassle. Ie, a proprietary system with non-open standards can't be ripped out and quickly replaced with something else, no matter how bad the proprietary system is. Or say some department has critical applications based on ActiveX, you can't toss that out if you don't have a compatible replacement.

Just like most IT infrastructures, there's a lot of duct tape architecture. It's easier to start from day 1 with a new infrastructure than to try and revamp an existing one.

Re:how much did this all cost?

[obarthelemy]

aren't you missing something, in between closed source and custom ? like .. open source ? which is what TFA is about ?

Re:There is no free lunch

[Bert64]

It is actually extremely rare for anyone to do a proper evaluation...
I know people who will evaluate multiple options based on their marketing literature and create a spreadsheet comparing feature checkboxes...
Some people won't even pay lip service to doing an evaluation, and will just choose something quite arbitrarily.
In the munich case, he chose open source and open standards for the significant long term benefits they will provide...

Give it a few years and noone will be able to argue against it, and the costs of migration and retraining often cited as reasons not to use open source will actually work the other way.

Re:There is no free lunch

[Bert64]

VBA was used because there were no other options when you're already locked into an MS stack...
Corel always made a much better suite than MS, and yet they were pushed out of the market by an inferior product... It's not about how good something is, its about how heavily marketed or pushed via other means it is.

Re:It's like the old saying...

[Lulfas]

The funny part is, from the article, they chose to go with Linux even though the estimates were almost 50% more than the Windows based solution.

Re:Why so prominent?

[Jah-Wren Ryel]

Why is the Linux migration project in Munich so prominent, as mentioned in TFS?

Because the guy who wrote it is German and lives in Munich.

There's nothing stopping you from writing up a submission about Banco do Brasil yourself. You seem to have access to a source with a whole bunch of good information, I'm sure a success story like the one you described would get coverage on slashdot too if someone made the effort to submit it.