Soda water. Because getting drunk isn’t on my agenda. But it fits the vibe and people think your having a vodka tonic or something

Even after you get your ideal setup with all your traffic transversing your network to a single host, you have bottle necked the whole network to the speed of that single host.

Usually in networks devices are able to talk to each other directly across switch fabrics and not interdesr with other traffic.

Say you have four devices A B C D each pair trying to send 1GiB/S of traffic to each other over a GbE network connected to the same switch. A,B gets 1 GbE and C,D gets 1 GbE. For a total concurrent speed of 2GbE.

In your model since all traffic has to hit the central wireguard node W first you can only get 1GbE speed concurrently

Sadly, and honestly, this.

Using an LLM with 4-year-old data is a better experience than digging through three pages of Google blog spam

The synology.com account, not the NAS account

is such a hassle it wouldn’t really pass in any company

Hate to tell you, this is now the norm. Right now, today, thousands of corporate travelers!

Company creates a travel laptop, perhaps even just a completely empty kiosk laptop. Corporate traveler downloads critical data to the laptop in an enclave (like a presentation). They have a two-factor token with them. If they need to get back to the corporate network for whatever reason, they use remote desktop software and no data is stored on the local device. They’re given policies telling them that if the computer is out of their possession, or view at any time, that the device is not to be used whatsoever afterwards. Contact security and let them deal with it.

When the traveler comes back to the mothership, laptop is checked into IT, it’s completely wiped.

Does remote desktop software suck? Yeah. It’s better than the alternative though

Well I’m ranting about this process, I have other complaints.

Synology.com - if you want to add a second factor to your account, requires a phone number to be the master factor, in case you lose your second factor. So if you’re worried about Sim jacking, or even just not having a consistent phone number for the lifetime of the deployment, it’s kind of a terrible practice. There’s no way to unlink all phone numbers from an account, you can only replace them with a new phone number.

Synology does actually support hardware USB keys, but only as a secondary factor behind SMS… Ai ya.

You can use a hardware security key, like a yubi key, or a software fido2 equivalent.

That way it satisfies the two factor requirement, without using a phone number.

For initial registration you can use an SMS service like SMS pool or the others, you pay a little money, you receive a real text message to a real phone number. You just don’t have access to that number in the future

Your voice, vocabulary choice, lighting conditions, power interference frequency, can all give away parts of your location and identity. You have to choose what level of paranoia is sufficient

The most anonymous, would be to have a v-tuber like model, respond and parrot LLM generated voice audio, from a script that’s been translated a few times. Or pay a voice actor from Fiverr to read your script.

Of course this whole time, using a VPN.

Fair enough. They did not explicitly say they removed it for complexity.

The facts are: they started with a protocol that had perfect forward secrecy, and they removed it, but not for philosophical reasons.

They were not opposed to perfect forward secrecy

In today’s ecosystem there are products that use onion networks and provide perfect for secrecy like simple x, and briar over tor…

You’re welcome to make any decision you like, if you want to use session go right ahead. I’m not going to stop you, and I’m happy you’re doing so. We’re all better for choice

It’s not a lie. I have read their post. And my interpretation reading between the lines is they dropped it because of complexity

Ipmi interface?

Wow. Where is all this hate coming from?

People like to experiment, and tinker, and try things in their home lab, that would scale up in a business. Just to prove they can do it. That’s innovation. We should celebrate it. Not quash people

Okay. Do you want to debug your situation?

What’s the operating system of the host? What’s the hardware in the host?

What’s the operating system in the client? What’s the hardware in the client?

What does the network look like between the two? Including every piece of cable, and switch?

Do you get sufficient experience if you’re just streaming a single monitor instead of multiple monitors?

Remember the original poster here, was talking about running their own self-hosted GPU VM. So they’re not paying anybody else for the privilege of using their hardware

I personally stream with moonlight on my own network. Have no issues it’s just like being on the computer from my perspective.

If it doesn’t work for you Fair enough, but it can work for other people, and I think the original posters idea makes sense. They should absolutely run a GPU VM cluster, and have fun with it and it would be totally usable

Fair enough. If you know it doesn’t work for your use case that’s fine.

As demonstrated elsewhere in this discussion, GPU HEVC encoding only requires 10ms of extra latency, then it can transit over fiber optic networking at very low latency.

Many GPUs have HEVC decoders on board., including cell phones. Most newer Intel and AMD CPUs actually have an HEVC decoder pipeline as well.

I don’t think anybody’s saying a self-hosted GPU VM is for everybody, but it does make sense for a lot of use cases. And that’s where I think our schism is coming from.

As far as the $2,000 transducer to fiber… it’s doing the same exact thing, just more specialized equipment maybe a little bit lower latency.

Can you define what acceptable latency would be?

local network ping (like corporate networks) 1-2ms

Encoding and decoding delay 10-15ms

So about ~20ms of latency

Real world example

Fiber isn’t some exotic never seen technology, its everywhere nowadays.

Moonlight literally does what you want, today! using hvec encoding straight in the gpu.

Try it out on your own network now.

Yes, for some definition of ‘low latency’.

Geforce now, shadow.tech, luna, all demonstrate this is done at scale every day.

Do your own VM hosting in your own datacenter and you can knock off 10-30ms of latency.

However you define low latency there is a way to iteratively approach it with different costs. As technology marches on, more and more use cases are going to be ‘good enough’ for virtualization.

Quite frankly, if you have a all optical network being 1m away or 30km away doesn’t matter.

Just so we are clear, local isn’t always the clear winner, there are limits on how much power, cooling, noise, storage, and size that people find acceptable for their work environment. So there is some tradeoff function every application takes into account of all local vs distributed.

If the picture has cocaine worth $1,000,000,000 … either cocaine has gotten REALLY expensive, or inflation has been HUGE…

Same same ‘Market Cap’

Nobody expects to clear at that price.

100% ^^^ This.

You could do everything with openstack, and it would be a great learning experience, but expect to dedicate about 30% of your life to running and managing openstack. When it just works, it’s great… when it doesn’t… ohh boy, its like a CRPG which will unlock your hardware after you finish the adventure.