Whenever I see those statistics, I always find them very weird. Many of our products are made in China, but those are counted towards China's global emission, not Europe's. It also applies to the need of power generation for delocalized industry or importing food. A country handling a lot of industry will have a very high per-capita emission, but the people living there may actually have very low individual emissions.

They aren't, most emission statistics are counting imported emissions.

China's figure aren't that great on their own, with exportation deducted.

EDGAR assigns emissions geographically to the country where the emitting activity occurs. (so in this case no)

Each pod has weight. These are written to your cgroup (cpu.weight in CGroups V2). CFS is scheduling based on these weights.

There's a part about it: https://github.com/inevolin/k8s-cpu-limits-analyzed#3-withou...

Not in the domain at all but wasn't this the purpose of STIR/SHAKEN but the biggest blocker was that large telcos hadn't deployed it? I think Android and iOS dialer app show you whether there was Caller ID Verification done or not so you may be able to drop anything unverified.

Worth a read about Signal's Sealed Sender and how it doesn't preserve its properties with receive and read receipts: https://news.ycombinator.com/item?id=46022279

We had the opposite experience. We contacted them through their form this year and they were happy to discuss with us even when we communicated from the beginning we wouldn't be customer in the short to medium term and even for single rack systems.

How did that compare to the big firms? HPE, Dell, and so on.

The rack were not insanely expensive, expandable so you don't have to build fully stacked racks from the beginning. Purely based on a hardware, compares to other blade systems but pricier.

The interesting part was the software, management interface, Terraform provider and how everything just fit together. Having storage, network and compute all in a single managed rack package brings a lot of value and brings down TCO. Really appreciated the security group like approach to network policies.

Unfortunately I haven't worked with HPE or Dell recently so I'm not sure what they currently offer.

You don't go to oxide for lowest cost.

I think the idea is that, in fact, you do.

Not one-time cost, but rather TCO.

You get a fully auditable hardware and software BOM, and you get an air-gapped, API-driven, on-prem cloud.

You still have a fairly large up front cost ($600k base in 2023ish numbers, probably over $1M now?), and an ongoing support subscription.

I think it’s fair to assume the idea is for it to eventually be more cost effective than other solutions as they scale up.

I've never been under the impression that is the case.

They are filling a niche no one else is for a small pool of fish with virtually endless pockets.

That really doesn't sound like a good business though.

I can't even fathom this statement tbh, it doesn't make any sense. They are in the most profitable and protected niche there can be in the US.

Historically, selling to a small niche of price-insensitive customers (e.g. mainframes, defense) breeds complacency and eventual failure. You can see VMware go down that death spiral over its history where every release was more expensive with less innovation.

> Historically, selling to a small niche of price-insensitive customers (e.g. mainframes, defense) breeds complacency and eventual failure.

VMware was a publicly traded company got more than a 20 year run where they dominated virtualization and captured over 80% of virtualized servers. That a publicly traded company sold out to another publicly traded company over and over after two decades where they got scraped for change in the couch cushions doesn't surprise me at all.

> (e.g. mainframes, defense) And yet mainframes are still in use, still getting purchased, and Defense is by far the largest segment seeing movement in the US right now. TONS of dollars, much of it going into AI, drones, etc. Much of which needs digital and infrastructure development and deployment for R&D and operations.

Right now is probably an amazing time for them sales / interest wise. A LOT of money going into software and hardware attestation right now.

I'm not entirely sure why you're getting downvoted. I wasn't able to find a source for the data but it is probably new vehicle registrations. Europe here seems to include Norway, Sweden, Finland, Iceland, Switzerland.

Sweden, Finland and Iceland would be part of any definition of Europe? They're all EU countries, and (irrelevant to that) are now all NATO members too.

Switzerland is not but its meager population can hardly significantly skew the sales stats

Iceland is not in the EU

Why wouldn't it?

Typically European Union vs Europe.

they’ll get included again if they order a few humanoid robots that will be shipping by 2078

Anywhere from ¥20 ($3) to ¥100 ($15) in large cities. Depending on the city, type of ticket and whether it's China-exclusive or worldwide release. Looking at prices, ¥40 doesn't seem like a bad deal.

> The script is hardcoded to read the stress binary at /usr/bin/stress since it doesn’t exist, the program errors out. To fix this, edit stress_helper.py.

Script is good, but the directory is missing from PATH. The edited scripts hardcodes the path to the user's home. Earlier in the article, the binary was copied to ~/.local/bin/stress so the given code suggestion should not work either.

Not familiar enough about GCP internal, but AWS prices are way higher than what's quoted in the article. You'll get billed for bandwidth, should you want any logs of your control plane you'll pay them at a premium so high* it will dwarf the cluster price and machine price combined. And for ~$70/month, you get the cluster API, not the EC2 compute required for worker nodes.

* Because vended logs to s3 are not supported (https://docs.aws.amazon.com/AmazonCloudWatch/latest/logs/AWS..., https://github.com/aws/containers-roadmap/issues/1141)

It really depends on the market you are targeting and your threat model. If your threat model allows future decryption of the data by a passive listener, then you don't need to rush for PQC. If you are worried about your communications being archived for future decryption, then you need to deploy PQC now even if QCs aren't developed for decades (or ever).

The assumption that you care about this is baked into ANSSI certifications, otherwise you would usually not bother certifying your product. They warned in 2022 that they would do this (See Phase 2: https://messervices.cyber.gouv.fr/guides/en-anssi-views-post...) and will allow PQC-only algorithm no earlier than 2030.