OMG just do, EU needs some balls. Upcloud, Scaleway, Hetzner, OVH are all production ready . EU business leaders are so afraid of not using the biggest provider they are blind to how you IT can absolutely run 100% on compute from these providers and opensource
It's very funny to me that all of our municipal EU customers requires that we host in EU. But they all run Microsoft online email, Entra and so on themselves.
I haven't tried to leave Gmail, but I've run email servers for 15 years and have never had trouble sending email to Gmail (get plenty of spam from them though). I often read people saying Gmail rejects mail, but I suspect such people are running mail servers on cheap VPS providers, who have bad IP reputation.
IP reputation lists predate Gmail by many years (eg spamhaus). I know Gmail used to use them, and may still. It would obviously make sense to keep their own stats as well.
I've successfully exited Google Workspace for email, calendar and contacts about two years ago. Went with Fastmail. It's been rock solid.
Unfortunately, Drive and Voice are keeping me subscribed to their service despite me wanting VERY BADLY to exit it. Neither have good alternatives that check all the boxes.
Sure but let's start with not having all citizens' sensitive data that EU state orgs store in american cloud, and then we can talk about creating an EU based alternative to google/apple in mobile.
The people who need to allocate software resources aren't "the government" writ broadly, they're specific people who operate specific organizations within it. You run a team within your local ministry of housing, and you identify a new analysis you want to run monthly that will take a day on a 12-core Spark cluster. Do you:
* clear out a supply closet, buy 12 new machines to put in it, and idle them for 97% of the year
* ask 12 employees to please deliver their computers to the supply closet on the last Friday of each month
* email Emmanuel Macron asking him to find another team who already cleared out their supply closet with spare cycles
Or do you just use the cloud and not have to solve an operational challenge every time you have a new idea? (The actual pre-cloud status quo was that most people simply couldn't implement any idea that would require more than 1 personal computer.)
They built their own internal clouds, to varying degrees of jankiness depending on how important large scale compute was to their business. It was very important for Spotify, so they had a bunch of people working on it, to the degree that it was hurting focus on their core business goals.
In less tech-forward organizations, you usually just had "the server", and it was rude or even forbidden to just run programs on it on your own initiative. You had to coordinate which programs you wanted to run when. And you'd better make sure your program doesn't have a memory leak or something, because then nobody will be able to check their emails until the sysadmin fixes it.
Government agencies can pool their resources to operate a shared data center. If you take just a tax office, they are operating at scale where they have sufficient demand to operate 2-3 data centers per country in half of Europe. As for the skills, a for-profit SOE or a non-profit can deal with this as a regulated primary contractor. IT is a special case where consolidation and moving ops in-house actually makes sense at that scale. US gov does not do that likely for political reasons.
There are plenty of EU domiciled Managed Service Providers who do have the skills though.
Having your government infrastructure run in country and managed by your citizens seems like a good idea just in general. It helps to develop local skills and the people living in country have a better feel for the needs of the local people.
I am an American but this just seems like a good idea even if the current geopolitical situation was better.
There are European cloud providers though. American hyper scalers are not the only option. Lidl, Hetzner, OVH, ..etc.
I can spin up a dedicated server within 24-48hours or a VM within minutes on OVH. Also there have been plenty of white papers written about how much more expensive AWS is when compared to Hetzner or OVH.
The big cloud providers are quite expensive and come with a lot of geopolitical risk/baggage. European governments have safer alternatives within their own borders.
Edit: it’s Lidl who launched a cloud service not Aldi
Before the concept of cloud was a thing, every company and a computer room, smaller ones were literal closets, bigger ones were large data centers, with everything in between.
Are you saying deploying Debian/BSD on some servers in the basement of a government building is too complicated and more expensive than paying Microsoft/AWS?
Governments aren't scale-ups/unicorns to need the scalability and global availability of cloud, they're ossified known quantity entities with predictable userbases and traffic across a very specific geographical region. On-prem is perfect for that.
> In the fire, 384 battery packs were burnt, which took down 96 government systems. Whilst this is obviously still a huge loss, 95 of these had backups - but the G-drive system (government drive), used primarily by the Ministry of Personnel Management, did not.
> [...] reports estimate that 8 years worth of data was lost, and around 17% of central government officials are impacted
Being against on-prem just because South Korean government implemented on-prem poorly with no backup best practices and lost data one time, would be like if homo sapiens stopped using fire because a guy burned down his straw hut one time.
Yes but you're saying this guy needs to build his own house and trusting him to obey the fire and safety codes, when plenty of professionals exist that specialize in following those.
No it's essentially just that a bunch of hype people sold everyone on the idea "the cloud is the future" and so even government types think they have to do it to modernize even if it costs more, is less secure, and less reliable than just paying your own IT guy to do it.
On-prem is not expensive or complicated, people just make dumb choices. Any IT engineer with two years of experience can run a small on-prem data cluster.
Infrastructure is one thing. How many companies that change to a purely EU infrastructure would be routing some of their customer data through 3rd party data processors in the US? I've had to map out 3rd party data flows and it is a daunting task and eye opening even when using applications that map out the TCP and UDP flows.
Do people then null route all the CIDR blocks in the US and all US based CDN's? Has anyone been daring enough to try this yet? I predict DNS to be the first single point of success even assuming one uses a EU specific TLD for all domains.
It needs the cooperation of member states to implement something like this. The EU cannot really force any large state to do anything - IIRC France broke EU rules on national debt or similar and nothing could be done.
haha for sure some one has made a little aggregator for this and saving tokens. I bet you gotta dig for a while though before you find a company exposing Opust 4.6 to customers and not flash 2.5 lite
If you have an LLM on the untrusted customer side the wrost it can do is expose the instructions it had on how to help the customer get stuff done. For instance phone AI that is outside of tursted zone asks the user for Customer number, DOB and some security pin then it does the API call to login. But this logged in thread of LLM+Customer still only has accessto that customers data but can be very useful.
You can jailbreak and ask this kind of client side LLM to disregard prior instructions and give you a recipie for brownies. But thats not a security risk for the rest of your data.
reply