Recently places like @SIDN (Dutch national operator of .NL) have been claiming that nobody in Europe can deliver their computer needs, and that they are therefore forced to outsource operations to American cloud providers. Meanwhile our own IT industry denies this. Here I delve into what's going on, and how Europe is being Cloud Naïve instead of Cloud Native.
Managed k8s is not necessary.
A good managed load balancer is. My current list of "need" for a cloud provider, with which i regularly revisit Europeans one (because i still believe that technically using US one is gdpr illegal)
on demand managed servers/vm with local nvme ssd storage
load balancers equivalent to aws alb
S3 like
managed postgres with automatic cluster management (see RDS)
@bert_hubert@SIDN afaict, only Scaleway in europe is going that direction. And from friend that tried and are migrating away now, the database and load balancer offering are not even meeting the lowest bar.
@bert_hubert@SIDN Kubernetes is nowhere offered reliable and solid (also not in American cloud providers). Kubernetes is a poorly written and designed platform full of issues - they even say this on their on website (https://research.nccgroup.com/2023/04/17/public-report-kubernetes-1-24-security-audit/). It has only niche cases and for the majority of the use cases there are better, more lean and more cost-efficient (in terms of efforts and infrastructure) solutions. Kubernetes is also not cloud native (it does not even know AZs or regions!).
@jornfranke@SIDN nevertheless people I trust tell me managed kuberenetes works for their usecase. It is in any case not great to offer an unnecessarily poor kubernetes experience.
@bert_hubert@SIDN Many people claim this - who would admit a fault that they use "fancy" technology that is not useful? I would ask for their KPIs. They will have higher personnel cost, more complexity, less security, less achieved, a lot of Kubernetes clusters are mostly idle and waste energy, a zoo of Kubernetes cluster as Kubernetes is not good in isolating things etc., it is the wrong abstraction for most problems.
@jornfranke@bert_hubert@SIDN Bert, did you analyze the market incentives there? Suppose OVH or Hetzner come up with their own version of cloud storage, would it sell? Probably only if there is dome kind of standard, as you say. But could such a standard emerge without Amazon and Microsoft? And if it did, how long before "embrace and extend"?
@jornfranke@huitema@SIDN I wrote the piece very carefully and in consultation with users of such services. I note in the article the many limitations of these services.
@bert_hubert@jornfranke@huitema@SIDN even more. The ovh service has been known to never reach one nine for years. It is a shitshow, and I do not think this is a powerful enough statement. Any managed service offered by ovh is laughably bad with a UX even more atrocious than AWS.
OVH managed services are a clown show not worth even considering.
@huitema so this is indeed a thing, hence my suggestion that companies at least flesh out a minimal and reliable set of cloud native services. This would allow them to at least get government tenders and appreciable revenues. From that point on the weakness of 'big cloud' is currently their complexity and surprise billing, which is also something new entrants to the market could address. But it will be very hard work.
@huitema meanwhile the incentives for European providers up to now have been rather simple it appears, cost cutting and delivering rock bottom pricing on simple server services. I can't blame them, there is money to be made there. But then don't complain if you don't get business for more advanced services which they've neglected.
@bert_hubert The silicon valley school of system design emphasizes "build a moat" in order to secure a monopoly. Typically relying on network effects and economies of scale. For the cloud service, what is the moat? It cannot just be individual services like S3, because cheaper copies are doable. Security? Identity? Customer support? It is very hard to compete without understanding that.
@huitema that part is relatively simple as far as I can see - customers aren't even migrating between GCP/AWS/Azure. It is a total lock-in once you are up and running. Switching costs are huge since people assemble services they barely understand let alone can migrate. Also customers quickly lose the competence required to do such things once locked in. It is a dream scenario.
The problem is that changing provider gives you basically nothing. None of them massively undercut price, because the price are not as extravagantly high as it looks.
There are no move between them because why the heck would you do that?
Add comment