Hetzner partnering with Ampere to bring ARM systems to their cloud | Hacker News Reader

Hetzner partnering with Ampere to bring ARM systems to their cloud

hetzner.com

71 pointsrglullis30 comments

I am over the moon with announcement. Being allowed with to play with a dual socket 160 core Amprere machine at the an ARM HPC User group meeting was a super interesting experience.

I as an existing Hetzner customer just called up their hotline and it is not known whether these machines will also be made available with the Hetzner Cloud or just as dedicated machines. A truly exiting time to be alive.

If a european company offered a reliable and reasonably priced hetzner-based "cloud" with managed k8s, managed databases, object storage, networking and maybe some FaaS system I would use that in a heartbeat.

I'd try to build it if I had the right technical/sysadmin background. Seems like a realistic task for someone with a small, scrappy team who does.

It seems reasonable that Hetzner will eventually offer all of that, but they're taking their sweet time with it. Meanwhile, Scaleway offers many of these things (obviously not Hetzner-based).

I feel like Hetzner is great at what it does: reliable, reputable, affordable server hosting. I imagine it to be a company full of top notch, old school sysadmins. Not sure they have the developer focus (?) to build an appealing AWS/Scaleway style service.

https://m.youtube.com/watch?v=5eo8nz_niiM

Fun to watch their DC :)

Interesting they are not actually doing Server Rack and Server units in U but actual good old fashion desktop.

I thought about building such a thing and have the right background. What do you think would be a good service to start with? I thought about k8s and then go on from there.

Yeah I guess k8s would be the best start, if you want to achieve aws-style features without a ton of development resources. I know barely enough to use k8s as an application developer though, so maybe not. Adding MiniO and a simplified, pre-configured Knative experience would be my next step, because everyone (including me) seems to love Google Cloud Run.

For networking not sure if you can/should integrate Hetzner's native networking and load balancers or should do some kind of overlay network like fly.io.

For managed databases and other services it would be great for you if first and third party "in your cloud" solutions like Aiven, Elastic Cloud, Snowflake etc would support hetzner but I've never seen it. Fly.io seems like a nice model of offering semi-managed databases.

With those things in place, and hetzner's few regions, some costumers might need Cloudflare in front of their services, but that might not be a bad thing. Is Hetzner in "Bandwidth Alliance"?

That’s exactly what Scaleway offers.

Scaleway ?

This seems like an interesting development and maybe could lead to even more affordable cloud offerings by Hetzner sometime down the line (of course the AI/ML spin of the article might imply something else, which is fine too).

As it stands, right now they seem like one of the most affordable hosts out there. In comparison, Contabo's UI is a bit ancient and apparently they overprovision resources, most other hosts like DigitalOcean/Vultr/Scaleway are all more expensive.

I guess Scaleway had Stardust instances which were really affordable but everything with higher specs was a bit less affordable, even though their managed services are pretty nice.

If it were more on topic, i'd ask what people here are using for when they want VPSes for a side project or something along those lines, though as it stands i'm all for seeing more ARM offerings, as long as running certain software on them isn't too much of a problem.

I really liked Scaleway's ARM offering as long as it lasted. The old native one (no virtualization). It was rock-solid in my experience.

You kind of have to wonder about why it didn't work out and become more popular to keep the product around.

Then again, i feel the same way about Intel Atom and most other low-power offerings - it puzzles me that they never got big.

Having low cost/high density options for hosting stuff online would be pretty great. After all, lots of people already run Raspberry Pi clusters at home.

I'm very open to corrections, but from what I recall of it, it was extremely custom to them, in terms of hardware, and seemed experimental. I suspect the logistics of managing a lot of custom equipment that wasn't making big money (as Amazon does with Graviton) wouldn't have been super appealing long term.

I had the same impression. For a turnover of 50 EUR per year and instance it was hard to see (for me as a complete outsider) how it could be an interesting business. Amazon charges more (as usual...), I know even less about their hardware.

The advantage Amazon has is that their hardware isn't really their USP whereas those Arm devices were for Scaleway at the time. While AWS does make a song and dance about Graviton (and quite rightly, IMO) AWS can really develop and shuffle in whatever tech they want behind the scenes as long as it runs customer workloads.

Wouldn't RISC be better suited for web server workloads than CISC anyways? If that is correct then how come it hasn't dominated that sector?

Unlike in the eighties and nineties, CPU front-end (instruction decode) is hardly an issue in modern times. Oversimplified, front-end translates instructions into something the CPU really understands. This part of the CPU used to be significant amount of die area, but not so anymore.

CPU back-end is what matters, and the answer there is that it doesn't make practically any difference.

For general purpose computing, RISC vs CISC is practically a dead topic now.

Of course you would not want a complicated CISC front-end in microcontroller. But that's another topic.

Ask the ARM server vendors that routinely canceled their projects.

Also, RISC vs CISC is for people stuck in the 80s. For example RISC-V's special sauce isn't some random "its better suited for web server workloads" theory with zero references or reasoning provided by a random guy on HN. It's widespread adoption. It's the fact that it was intentionally kept simple to make it easy to implement to the point that it probably is missing lots of functionality that we take for granted in modern servers.

Remember all the fanfare around Cloudflare blog posts for a qualcomm ARM server SoC that was never released? Yeah the performance was surprising but the amount of hardware specific code the Cloudflare guys had to write was surprising too.

> Also, RISC vs CISC is for people stuck in the 80s.

Sometimes I joke on this topic to the akin of Indo-European languages vs Sino-Tibetan languages, or in layman's term but a little offensive [1], English vs Chinese. This analogy is quite intriguing to me, since English is relatively easier than Chinese, and Chinese has more entropy than English, like RISC is much simpler than CISC, while CISC does more in one instruction than RISC in general. Of course, the true holy grail would be VLIW, which is like Esperanto, barely anyone knows and uses.

But as I can speak both languages and recognize the difference between both ISA designs I don't really think it mattered, but there will always be some stubborn people who bitch about it. This kind of behavior sometimes lead me to think about Linguistic relativity [2] and even extend the whole argue into the highly political topic of Linguistic determinism [3].

[1]: there are couple more of those Indo-European and Sino-Tibetan languages. Deutsch is an obvious candidate.

[2]: https://en.wikipedia.org/wiki/Linguistic_relativity. aka "what you speak could have affected what you thought"

[3]: https://en.wikipedia.org/wiki/Linguistic_determinism. aka "what you speak affects what you thought"

> Remember all the fanfare around Cloudflare blog posts for a qualcomm ARM server SoC that was never released?

Announcing stuff that is never released seems to be a common pattern for Cloudflare. They announced their AWS S3 compatible competitor "R2" over 18 months ago [0] without a release, and became very silent about it. At this point it is just vaporware.

[0]: https://www.cloudflare.com/en-gb/press-releases/2021/cloudfl...

As was stated in the earnings call for Q1: "We're on track for R2 to progress to open beta in Q2 and then be generally available in the second half of 2022."

So, very soon now.

Great to see some progress. But the fact that it is only announced in an earnings call - where you kind of have to answer such questions - makes me sceptical. There is probably a reason why there is no mention of R2 anywhere on the Cloudflare site or other channels targeted towards developers?

> Ask the ARM server vendors that routinely canceled their projects.

Pre-Graviton/M1, you could argue there was a chicken and egg problem on the datacentre and higher-end compute side of ARM wherein smaller vendors couldn't achieve the volume, and so couldn't achieve the economies of scale to price an ARM chip competitively and drive revenues to reinvest, so fewer binaries were built keeping it niche, so it continued to lag in compatibility and compute and its power advantage wasn't enough by itself. Once big guns like AWS joined the fray, ARM's performance was improved as well as its natural advantage in power consumption and with big budgets they have the staying power to wait for network and ecosystem effects to win in the datacentre. Ironically, that probably creates a better space now for smaller players to tease out other customization benefits for other verticals on a stronger core datacentre presence and larger customer base. But before that happened, none of them had the resources of an Intel and AMD to compete for the datacentre by themselves, even if the x86 architecture and legacy created a ceiling everyone could see and ARM was a natural suitor.

> Remember all the fanfare around Cloudflare blog posts for a qualcomm ARM server SoC that was never released?

Like all Cloudflare fanfare posts, it was to get a discount from Intel.

I mean, that was sort of my question. If the performance is decent then why would you not write hardware specific code, especially when operating at cloud scale like Cloudflare. I think I just lack a good understanding of the complexity at hand.

Interesting! I'be been trying to get my hands on an Ampere server in Oracle cloud without any success... I hope I will be luckier with Hetzner.

Oracle did not even accept my credit card (which works fine in AWS and Scaleway). So I decided it's probably best to follow general wisdom: Don't do any business with Oracle.

They rejected mine too. Not sure why. I was wondering if that happened to others as well and also didn't look too far because oracle.

Try a different region. Never had issues getting ampere in uk oracle

Reply on news.ycombinator.com