Kazumara

It's like people still don't know about Schrems II or the Cloud Act.

Or they somehow seriously think that the EU-US Data Privacy Framework resolves the issues that killed the EU–US Privacy Shield?

In my org email went to shit after they outsourced it and lost the institutional knowledge. Now we suddenly have random things happen, like a second layer of quarantine appearing, and nobody can explain it. Any support request is copy pasted forward and backward to the outsourcing provider. If the outsourcing provider's response makes no sense it's forwarded to you internally none the less, and without comment.

My colleagues tell me that back in the nineties we were running an X.400 email gateway in this very company before it was clear that Internet email would be the one to win the protocol wars. We were at the forefront of email developments then.

And we're still a god damn tech company. We're a registry (not registrar), network provider, security services provider, cloud provider, etc. But email is now apparently too hard for us, it's a sad state of affairs.

Yes, then you only pay tariff for the powder, for the big jar, and the transpacific shipping of the big jar. That's what I meant by "as far as it was packaged in China".

I feel like it's a CEO's job to care about all aspects of the company he is supposed to lead.

At the border you pay tariffs on all the costs up to that point, because they are all considered to be part of the value of the shipment as it crosses the border. So the price of good, plus the price of packaging (as far as it was packaged in China), plus the price of the freight shipping are tariffed together, which makes the result of the calculation a little worse, but fundamentally you're right.

Wow interesting! Thanks for checking the label!

Yes, and that's before muxing!

I need only 75 GHz of spectrum to send 400 Gbit/s through our country. We've currently got a link running between Zürich and Lugano (two amplifier sites in between, before and after the alps), and I've got 4400 GHz of usable spectrum with our currently deployed system, so if we needed it and spent like two million dollars we could deploy 23.2 Tbit/s within months, using just our normal commercial stuff, on a single fiber pair.

Using G.657.A2 fiber you could get away with 7.5 mm bend radius, or 15 mm diameter, for the innermost layer of the spool. That's around 5/8 inch for freedom units.

But then again if you went that tight you'd need like 56'000 windings for 10 km. That sounds like a fuckton, and like we can't ignore the outer diameter being larger.

Approaching it from the other side: The fiber diameter with coating but without any mantle is 0.25 mm. If you want to put 10 km on a 100 mm long spool you could put in 400 layers lengthwise, and each layer would have to be a spiral of 25 m (of course you'd spool it outside in, not layer by layer, but should be mathematically similar enough). Using this spiral calculator and some random changing of the values it looks like an outer diameter of 91 mm (3 & 5/8 inch), and inner of 15 mm and a thinkness of 0.25 mm would work for a 25 m spiral.

Or if we go for 125 mm drum length, so 500 layers, with 20 m each we get 82 mm (3 & 1/4 inch) outer diameter.

Or if we go for 150 mm drum length, so 600 layers, with 16.7 m each we get 75 mm (3 inch) outer diameter.

So yeah I think your estimate was pretty spot on, if the 10 km length is the right assumption.

Today's wires aren't actually wires, they are optical fibers. It must be G.652 or G.657 from telecom use, since that's commercially available en masse. I think most likely would be G.657.A2 because that can be bent tighter. Here's an example data sheet from a random google search. I wrote it in a different comment already, but the core has 9 micrometer, the cladding 125 micrometer and the coating 250 micrometer diameter. For telecom applications you'd add at least a mantle, or more likely use a cable with many fibers in little pastic tubes wrapped around a metal core for stability, 12 x 12 is fairly standard. Here of course it's just a single fiber without mantle being spooled off.

Bidirectional transceivers (so both directions on a single fiber) can do 100 Gbit/s ethernet too. No way you'd do that for drones of course, but just to show how far you can get with a single fiber.

bend the fiber around a pencil and only experience degraded signal

Interesting. Are you using G.657.A2 then?

Those news are already not so new any more. We've had reports of those two months ago.

Since fiber optic wire guided missiles exist it's not that much of a leap to think it should work with drones too, so long as the weight works out.

Fiber is really really thin. 9 micrometer core diameter and 125 micrometer cladding diameter (incl core) and 250 micrometer coating diameter (incl core, cladding). The 10 km spools we use in our lab for network equipment testing are boxes of only like 20x20x10cm, and those aren't optimized to be extra small with bend insensitive fiber. I can totally believe the 1.2-1.4 kg for 10 km in the article.

Edit: leak -> leap