The biggest elephants in the room are cloud providers, but I didn't find an easy alternative yet (hetzner, ovhcloud). ATM, the idea to the business is sold, that data resides somewhere near by in a datacenter, EU proximity. However, the EU businesses are realising that, well, whole region is at a mercy of one person.
Hetzner is great value, but their networking has a few issues:
1) Networking is mostly limited to 1Gbps. Even private networking. You can request a 10Gbps NIC, but it has to be housed in the correct data center and adds a $48 monthly fee.
2) Private networking is IPv4 only so dual-stack private networking isn't possible. Also each public IPv6 address is /64. Would be nice to get a /56 to setup dual-stack IPv6.
3) Can't specify a subnet to assign a server to when using hcloud API/Terraform. You have to specify the required IP on the subnet explicitly.
4) As I understand it, the private network traffic isn't truly secure between tenants, so needs to be encrypted between nodes anyway.
Still, I'm betting they'll fix these issues as their offering grows.
People are really quick to forget the fire that destroyed one of their data centers a few years ago and which did not get addressed in any way by OVH for months.
They also learned nothing from it, and are repeating the exact same mistakes.
I stopped hosting even my personal blog on OVH because of how garbage it is.
Yes they are such chaos internally. Even their support tells you different things every time. I kept having issues around my IRC bouncer on one of my servers (kimsufi, their budget brands). Some support people said yeah no issue as long as you don't do anything illegal. Others said I'd get insta-banned, and sometimes I did have issues and had to call them to get re-enabled.
Now, I have to admit I haven't been a customer of them for 10 years due to exactly this. But yes the fires exposed a lot of the same I left them for.
I left to go to DigitalOcean but it became too expensive and then I found Scaleway which I'm a happy customer of for years now.
Both companies are excellent, and I'd absolutely trust them with my business, but neither can replace something like AWS. The friends I have at companies who are actively using AWS are all relying on a fairly large number of AWS only services. Either they'd need to stand up their own replacements and host those services on VMs, or in some cases rewrite parts of their stack.
E.g. if you're using AWS Cognito then you're not going anywhere.
Exactly! You can get a bare minimum, like a virtual machine (EC2) or storage (S3), which probably enough for small and medium enterprises (SME). However, if we move beyond, I'm not sure as I don't have experience with them. Now, if I'm building a prototype, I want something quick and just a lack of Cognito is a deal breaker.
IMHO Aws is designed for totally embracing their philosophy and language. You don't understand two Aws Devs talking to each other. Even organizations are internally structured for Aws operations. This create something even stronger than a dependency.
> Making yourself a subsided of Amazon was never wise
True, but the AWS pricing doesn't make sense otherwise. If you're not using the managed services, then the value proposition is no longer there. Using those services is what allows you to build massive systems for relatively cheap, with much less staff. We had a project that was to be moved from on-prem to Azure (same deal), it went from thousands of Euros per month to fitting into the a free-tier, but only because we could use managed services. Spinning up the same VMs would cost more than hosting it ourselves.
I like them a lot but they only have EU DCs, if you are looking for Global (or at least Asia) you're out of luck for now. Perhaps this disconnect from US services might give them the impulse to spread out though! I'm really happy with them as a customer and I don't have needs beyond Europe anyway.
I've found Scaleway for AWS-style managed backend services fronted by Bunny (https://bunny.net/ - also EU-based & owned, but with global CDN DCs) works well! Bunny have nearly 30 DCs in Asia alone.
Problem with Hetzner is they don't have the self hosted DCs in pacific region yet. They have Singapore for their PaaS solution, but if you want those cheap second hand servers then have to be in EU
I think this is less of an issue than people actually think - if it gets to the point where this becomes a real problem, individual EU countries can force the datacenter owners like Google/MS to change ownership structure for these datacenters to EU-based subsidiaries or completely new companies if they want to continue to operate.
This doesn't matter as far as the concerns about US warrantless surveillance laws go because those laws also apply to subsidiaries of US companies. IIRC Microsoft tried to argue that its EU subsidiary could not comply with US requests and lost.
Usually everything is through subsidiaries. For tax and profit allocation purposes the way it works is that you set up subsidiaries in tax-friendly jurisdictions and then channel the profits to them through contracts between subsidiaries.
The general point is what does "moving away from US cloud services" mean, then?
Does it mean not using infrastructure actually located in the US? Or does it mean effectively boycotting US-owned companies that may be fully located, including infrastructure, in Europe?
Not one of these, but I tried on a small, Lithuanian, language. The catch is what the language has complicated grammar, but not as bad as Finnish, Estonian and Hungarian. I asked to summarise some text and it does the job, but the grammar is not perfect and in some cases, at a foreigner level. Plus, it invented some words with no meaning. E.g. `„Sveika gyvensena“ turi būti *atnemitinamas* viso kurso *vykišioje*.`
One way to test this is through a placebo test, where you shift the treatment, such as moving it to an earlier date, which I have seen used successfully in practice. Another approach is to test the sensitivity of each feature, which is often considered more of an art than a science. In practice, I haven't observed much success with this method.
You can apply it to estimate the impact of any business decision if you have data, so not only IT companies can benefit from it. However, the problem arises when the results don't align with the business's expectations. I have firsthand experience with projects being abandoned simply because the results didn't meet expectations.
"The data, which looked at more than 700,000 fully vaccinated people in Israel, found that only .04 percent of people contracted Covid-19, according to the head of Israel’s Ministry of Health, Dr. Sharon Alroy Preis at a press conference on Thursday.
The type of vaccine Israel has mostly used is the Pfizer/BioNTech vaccine with a small number of doses from Moderna. The two vaccines use the same technology, and in clinical trials were found to be about 95% effective.
"
https://edition.cnn.com/world/live-news/coronavirus-pandemic...
Can you share your datasource claiming that efficacy is only 50%?
You’ll have to actually read it, or find a tldr from someone who knows what they are doing.
The same calc that leads to 0.04% above would say overall covid mortality is 0.08% and is intentionally misleading - but no news media so far (in Israel or outside) has done any review. They are all parroting the misleading numbers.
Anything but primary data these days is suspect. (Pfizer’s own primary data, btw, says severe disease is 75% RRR with CI 0-100%. Did you see anyone reporting on that?)
Edit: to add: Israel is an a horrible bind right now. The vaccine is only 50% effective but gov has done very badly in handling the disease and promoted the vaccine as “our way out”, so the gov is attempting to hide that data. Unfortunately for them, Israel is too small and the culture not supportive of hiding this, even if at this point the mass media is cooperating with the government.
"We demonstrated an effectiveness of 51% of BNT162b2 vaccine against SARS-CoV-2 infection 13-24 days after immunization with the first dose. Immunization with the second dose should be continued to attain the anticipated protection."
I.e. if you take only the first dose of the two required then 13 days after taking it you already get an effectiveness of 50%. Therefore they recommend that people take the 2nd dose as prescribed. They don't mention effectiveness one week after the 2nd shot, which is what Israel considers to be immune (i.e. "green passport" status).
For severe disease, Pfizer had 4 in the control group, and 1 in the vaccine group. That’s 75% RRR, with a confidence interval that spans 0-100%.
Look at the graphs here and in Pfizer’s. Immunity seems to taper at 12-18 days and stay there. The 2nd vaccine does not seem to confer more immunity - just expected to last longer time. This article looks at more severe disease than Pfizer’s 95% endloint (though possibly not severe as their “severe disease” endpoint - I don’t remember what definition each uses, but just in Israel the “severe” definition has changed several times)
Please stop spreading lies. The paper is only about "The effectiveness of the first dose of BNT162 ..."
50% efficacy after the first dose is exactly the number that BioNTech reported in its Phase3 trial. So nothing unusual here. That is why the 2nd booster shot is needed to get to 95%.
No it is not. Have you actually read either Pfizer’s report or this paper? Looked at their data?
The graphs, both in Pfizer’s original report and in this paper, show that past about two weeks, there is almost no change. Please look at the Pfizer graphs and tell me where the great advance is from 50% to 90%
Also, if yo actually read the Pfizer report, you’ll notice the 95% is for mild disease - a totally uninteresting result. For severe disease, it is 75% with a CI spanning the whole 0-100% range.
I won't hold you to this, even though you are guaranteed to lose in this bet even if you are right - because the trial for 5-12 that would allow to vaccinate them is not expected to yield any data until later than that. Even if you are right, it will still be here in that age group.
Also, there's a nature paper from last week that shows incredible correlation between Sun UVB at 34% of equator and the date that a wave starts. Covid will be gone by June, just like last year, thanks entirely to that effect (and possibly arrive again in Sep/Oct, or not). I'm not interested in meterology, but it's possible that even if I'm right, by end of March covid will be very dim in correlation with UVB exposure.
The amount of data actually attributable to cause and effect is incredibly miniscule. I do data science for a living. The headlines are basically all wrong, most discussions are, and quite a bit of the primary data cannot be right.
added: p.s - Did you actually read the papers before calling me a lier (or after?) or just did it based on headlines. I'm not holding it against you either way, just wondering.
Presumably yes. The site of the guy who sued FB has a template you might be able to reuse[1], although it mentions the Irish implementation of the Data Protection directive, whereas Under Armour Europe B.V. is Dutch, so you should probably change that.
Robots and automation. I think in near future (put your number here) only cool jobs will be available for humans and shitty jobs will be done by robots. So, we need to find a way how to feed and entertain let say 90% of population.
