A good start would be switching all government systems from the Microsoft stack to a home-grown alternative. But efforts to switch to Linux are regularly stopped.
So we’re actually pretty progressive in terms of pushing open source while also supporting local business.
We operate more than 300 IT systems though, I’d say that around 75% of these run on windows only, many without suitable alternatives. So even on the technical side, we can’t swap our stack because our employees wouldn’t be able to do their jobs. We’re working to lower this, and stuff like web-apps and the rise of android/iOS devises has helped but it’ll probably take 25 years to happen, and here’s the thing, Azure is actually the most EU friendly, secure, stable and cost-efficient cloud platform, so a lot of those non-Microsoft software actually still run on Microsoft.
Then there is the employee training. I’m not sure if you’re aware, but the primary cost of running a public organization is your employees. They are the most important resource and replacing them is really expensive. So is retraining them, and it’s also something we already struggle with in terms of IT.
Most times when we switch a system, the technical implementation will go excellent. The organizational implementation will go horrible, however, because learning how to use an IT system is hard. It’s even harder to learn how to use it efficiently and often local management will be reluctant to invest enough time or focus on what they view as IT changes, because they are already understaffed in their primary function.
Windows, android and IOS are something most people have worked with before. So is the office package, and again, where is the non-Microsoft alternative to the office365 enterprise stack? Anyway, switching people who can barely “turn on the internet” (talking of course about the browser, but that’s what you’ll literally hear every day in our support center) to open source alternatives is a tremendous effort that nobody, nobody, outside of tech wants to do.
Then there is IT, we’ve run Microsoft for decades. Our staff is trained for it, they’re certified in it and despite what you may think, they’re actually cutting edge in terms of skill. At least in my country they are, we’ve yet to find an ADFS consultant that knows more about ADFS than our guys, in fact, we’ve had to send most of them home because they weren’t able to help us. Where do we find Linux replacements for those people? And how do we pay the 25 million it would cost us to replace the ones that don’t want to work with the Linux stack?
Lastly there is support. Microsoft may have a spotty reputation in popular culture, but their support has always been top of the class to the public sector. If we file a primary incident report to Microsoft, Seattle will be on the phone with us until it’s fixed. What Linux house will offer this service? Even if we found one, I don’t think we’d really want to throw away decades of partnership with an untested entity just because.
I wish we could run European based open source software, I really do, but the truth is that it’s complicated, highly improbable and the business case isn’t even there. Because it would be really expensive, and do you really want to sacrifice welfare to run Linux? Would your political leadership?
But then I see cases like the Munich project where they have been using Linux for several years[1], saving 11 million Euros in the process, until Microsoft moves their HQ to Munich and suddenly the council decides to switch back to Windows.
In Germany at least there is tremendous potential for cost savings if it were possible to develop one FOSS set of software for municipalities and share it across the country instead of paying millions in licensing costs.
I work primarily with digitization and efficiency, so I’m cursed in that I’m always thinking about cost efficiency. This is actually the primary drive behind my personal open source idealism. Since open source is customizable, we can cut away the clutter and focus on user experience and results.
But because of my background I can’t help but think about the man-hours they must have spent on the organizational implementation, and how many hundreds of millions of euro that must have cost them.
Maybe the German public sector is less digitized than we are, making the change lighter. Don’t get me wrong, we could certainly save money on licenses too. It’s just that the cost of doing it far exceeds the benefit, but I’d love to learn from Munich.
The cost of change is why I personally think hard switches (I did notice them calling it gradual on the wiki page, but I’d argue that it’s not if you swap OS before every other system is ready) are a mistake in enterprise. We’re gradually changing our plumbing to be more open, but we’re doing it so that nobody notices. Our MSSQL cluster is a good example, it’s currently necessary because of legacy software. In one or two license cycles I fully expect it to be replaced by open source however. If things go as planned then no-one on the business side is going to notice anything but the freed up funds.
> so a lot of those non-Microsoft software actually still run on Microsoft
As a FLOSS advocate myself, this matters less than you seem to imply. Microsoft is publicly traded, they'll pivot to goat herding if that's the way to maximize profit. So avoiding it is (imo) less important than avoiding their lock-in products. If Azure has little lock-in, might as well use it, icky as it may feel to me personally. It's not inherently anti-FLOSS, because Microsoft is not inherently anything.
> If we file a primary incident report to Microsoft, Seattle will be on the phone with us until it’s fixed. What Linux house will offer this service?
I think both Redhat and Canonical would love that business.
I'm not looking to argue as you seem to know what you're saying, but on the financial side, how many millions are spent on Windows, Office and other licenses every year? I'd say it's worth doing the math.
I also worry because of the political climate. It’s extremely unlikely, that the relationship between the EU and the US will really go bad, but we have contingency plans for things you wouldn’t believe, so naturally we have one for Azure.
I don't think it makes sense for a particular European country to develop its own distro, and certainly I don't think a state or county should be developing its own distro either, as it happened in Munich. Munich having to support its own distro is pretty crazy. At the EU level, administrations would get all the support and funding they need, pretty much.
I do think it's probably better to have a EU-developed distro than using Fedora or Ubuntu, though. Ubuntu at least already comes with proprietary software and licenses, so I'm not sure it would be that much better and trustworthy than Windows is in terms of national security. Let's imagine a war between some EU countries and the US in the future - would Ubuntu be free from manipulation? I'm not talking about hacking, which is to be expected, just good old hands greasing or compelling of the company to create malicious code.
Having an EU-wide distro that all administrations use and rely on would also have some other "unintended" but positive consequences, such as states not being as interested in quitting the EU, since all of their institutions rely on the EU distro. It wouldn't be the deciding factor in such cases, but I'm sure it would play a major role.
Now, do I think Munich gave up its distro because it was "too difficult and expensive"? No, absolutely not. The deal-breaker, or should I say "deal-maker" was the fact that Microsoft promised to build a whole headquarters there! And then suddenly Munich switched to Windows.
I strongly believe that without that coming into play, Munich would've just pushed through with the Linux distro, since they'd already been using it for the past 10 years, or worst case scenario they would've switched to Ubuntu and Libreoffice.
Microsoft used this as a PR technique "see, Linux doesn't work! Look at these guys, they've already switched back to us!" But it only happened because they made a major move to convince them to switch back to Windows, but 99% of the public doesn't understand or know that, so they believe Microsoft's PR, which seems to have worked very well, because it sounds truthful, even if Linux being hard to use and whatnot wasn't the deciding factor.
Microsoft did something very similar when it began patent trolling Android OEMs. They knew most companies would scoff at them and their patent trolling attempt, so they first got some highly strategic wins. First they got HTC to agree to pay the royalty, and then Samsung, the largest OEM.
Now you, and 99% of the people who don't know the whole story would say "but that sounds about right, looks like those companies thought Microsoft's patents were real so they agreed to the deal, no?"
Wrong. What Microsoft did was give these companies something very compelling in return - something where the net winner wasn't Microsoft in the deal, but those companies. If I remember right, HTC got free Windows Phone licenses for a while, and Samsung got a similar deal, and also free or very cheap laptop Windows licenses. There may have been other benefits for those companies, too.
So basically Microsoft didn't care it had to pay these companies to accept their patenting deal (however little sense that may make at first glance), because they knew that later they can push the story that these companies agreed to their patent licensing deals, because their patents are real and they are afraid of getting sued by Microsoft. Plus, they'd get to go to other companies and say "Look, Samsung and HTC, the most important players in the Android market agreed with our licensing deal, and you won't? Are you sure you want to go down that path?!"
I’m also a realist, and have you met the EU? We’ve been working on architectural standards for organizational data for almost 20 years, and it’s nowhere near finished. :p
If we can’t figure out a standard for how to define a department or employee object in a data model, then how on earth will we ever maintain a universal Linux distribution?
I think your strategy for implementation is flawed though. You really don’t want to tear the house down and build a new one when people are living in it.
I realize I’m a bit of a conservative in that sense, but people’s lives depend on our critical systems.
As far as Microsoft goes, they are a company. I think they are a great business partner in an ocean of bad ones. It goes against my personal political beliefs to use proprietary software, but the public sector isn’t primary a tech sector and we don’t operate on personal ideology. We use the best and most efficient tools, at the cheapest cost, to support our primary function.
https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32...
Each of these directives is implemented as a law in member states. The Dutch version is here: http://wetten.overheid.nl/BWBR0036795/
If they failed to adhere to the law you should do so, but I’d urge you not to request access in the future unless it’s really important. Because making a request of access is one of the most expensive things you can do.
We get a few requests each year, every department does, and every time we have to dig through a number of different systems and thousands of documents to prepare stuff that’s relevant. Then we have to read through every single document to make sure there isn’t any sensitive information and manually redact it if there is, while filing a report on why we redacted to internal review.
We do this, while expected to perform our normal functions by the way. I don’t know about the Dutch, but we take every request very serious, but for me personally it’s a little disheartening to know that we could hire 10-15 extra teachers/nurses/care takers if we didn’t have to deal with requests of access done solely by private citizens who feel like they “fight the good fight”.
I mean, who are you fighting? Your own society? The press is usually ahead of you too, but they actually have the resources to process the information they request.