1TB in 1 second? How is that even possible?

Ghosty:

1TB in 1 second? How is that even possible?

I was about to ask this too. Even if this is theoretical performance, I don't get how it's possible. Let's say hypothetically these researchers got their hands on a dual-socket AMD Genoa system, with 24x channels of DDR5. Not even that would reach 1TB/s (it would be a little bit under that). This is not accounting for the obvious issue of how a network interface can tap into all that bandwidth. You could argue "yeah but what if it were highly compressed?" well, the receiving end of the data needs to decompress it in order for this to count, and you can't decompress 1TB of data in less than a second; that's just not possible today. Now.... it's possible they got 1Tb/s, as that is "only" 125GB/s. That would be achievable, and impressive.

schmidtbag:

I was about to ask this too. Even if this is theoretical performance, I don't get how it's possible. Let's say hypothetically these researchers got their hands on a dual-socket AMD Genoa system, with 24x channels of DDR5. Not even that would reach 1TB/s (it would be a little bit under that). This is not accounting for the obvious issue of how a network interface can tap into all that bandwidth. You could argue "yeah but what if it were highly compressed?" well, the receiving end of the data needs to decompress it in order for this to count, and you can't decompress 1TB of data in less than a second; that's just not possible today. Now.... it's possible they got 1Tb/s, as that is "only" 125GB/s. That would be achievable, and impressive.

1TB a second would make more sense.

Ghosty:

1TB a second would make more sense.

Remember, this is wireless. I'm pretty sure nobody has done 1Tb wirelessly.

schmidtbag:

Remember, this is wireless. I'm pretty sure nobody has done 1Tb wirelessly.

The receiver is very close, less than 1KM away, but even so. It would take longer than 1 second to pass between two points. Unless the laws of gravity and physics have been broken.

or rather ..... how to reach your bandwidth cap limit in less than 1 second on your mobile data !

Ghosty:

The receiver is very close, less than 1KM away, but even so. It would take longer than 1 second to pass between two points. Unless the laws of gravity and physics have been broken.

Gravity on radiowaves plays no role ....at all . ok you can tell that the particles can be affected by gravity in a massive scale ... like not even escaping a blackholes gravitational pull , but for all intensive purposes and practical here on earth gravity effects is not even worth calculating ! Now laws of physics no laws are broken here see to up your data rate all you have to do increase the frequency the higher the frequency the less the signal can travel and be readable and less penetration of obstacles , and this is not all ! adding multiple frequencies for random example 14 khz and 14.1 khz signal multiplexed as far your can separate em again effectively you literally doubled your bandwidth ! I imagine this is very short distance and involves ... A LOT of channels to do that ! The speed of transfer is the speed of light since light is also a particle wave ! So trabsfer this kind of data in 1 second is not out of the possibility the harder part would be to have fast enough saving space storage to fill it with that data as it comes :P

Venix:

or rather ..... how to reach your bandwidth cap limit in less than 1 second on your mobile data ! Gravity on radiowaves plays no role ....at all . ok you can tell that the particles can be affected by gravity in a massive scale ... like not even escaping a blackholes gravitational pull , but for all intensive purposes and practical here on earth gravity effects is not even worth calculating ! Now laws of physics no laws are broken here see to up your data rate all you have to do increase the frequency the higher the frequency the less the signal can travel and be readable and less penetration of obstacles , and this is not all ! adding multiple frequencies for random example 14 khz and 14.1 khz signal multiplexed as far your can separate em again effectively you literally doubled your bandwidth ! I imagine this is very short distance and involves ... A LOT of channels to do that ! The speed of transfer is the speed of light since light is also a particle wave ! So trabsfer this kind of data in 1 second is not out of the possibility the harder part would be to have fast enough saving space storage to fill it with that data as it comes 😛

Kind of like a multi spectum array of fequences all at the same time? I suppose the only problem then would be the data being corrupted, given the high rate it is sent and received?

Ghosty:

Kind of like a multi spectum array of fequences all at the same time? I suppose the only problem then would be the data being corrupted, given the high rate it is sent and received?

well yeah the more tightly you pack frequencies the harder it is to separate em , pretty much every signal today is multiplexed passing multiple frequencies threw one carrier (air at this instance ) , the ancient by todays Standard ADSL passing multiple channels as well just on different carrier (copper cables ) . Jam pack enough frequencies and theoritically you can pass infinite data since frequencies are infinite as well your limit is your ability to reliably read the signal ! I can't imagine what kind of gear you would need to even attempt to read 1 teraherz signal XD

I thought they already did that. The corona virus

Ghosty:

1TB in 1 second? How is that even possible?

Packet interweaving across multiple fiber links.

Espionage724:

Easy, I can just hand someone next to me a 1TB HDD 😛 Practically, I think this would be nice for wireless PCVR. 5G with 1ms latency is probably already pretty good, but 6G with 0.1ms latency and 1TB bandwidth sounds even better. Anything that allows for full-graphics fidelity without compression and low latency for VR sounds nice!

It all depends how it works over longer distances without degradation. Other factors like different traffic speeds and a more populated network come in to effect then too.

schmidtbag:

I was about to ask this too. Even if this is theoretical performance, I don't get how it's possible. Let's say hypothetically these researchers got their hands on a dual-socket AMD Genoa system, with 24x channels of DDR5. Not even that would reach 1TB/s (it would be a little bit under that). This is not accounting for the obvious issue of how a network interface can tap into all that bandwidth. You could argue "yeah but what if it were highly compressed?" well, the receiving end of the data needs to decompress it in order for this to count, and you can't decompress 1TB of data in less than a second; that's just not possible today. Now.... it's possible they got 1Tb/s, as that is "only" 125GB/s. That would be achievable, and impressive.

It was proclaimed to be 1TB per second. What they don't say is if it was concurrent transmissions or what. Very light on details.

Ghosty:

It all depends how it works over longer distances without degradation. Other factors like different traffic speeds and a more populated network come in to effect then too.

It will not work over longer distances , I assume they use even higher frequency than 5g . Means lower distance and lower object penetration , considering 5g and wifi 6 have issues penetrating a sheet of paper ... G6 should be even worse ... For this thing to work you will have to be outdoors and have an unobstructed line of sight with the antenna . How I can see it working indoors is having an antenna on your roof and tiny low power repeaters on your ceiling ... Few milivolts of signal shall be enough , so when it comes to radiation less than your ceiling lamp light ! Anyway that said when I passed the class in university about telecommunications g4 was not yet released and what I remember is a bit dusty ... So except if there is a breakthrough I am not aware off I should be spot on or close enough at least.

Astyanax:

Packet interweaving across multiple fiber links.

ummm, not. this is wireless 6g not ethernet

this is just semi-random noise. you have to take this with a grain of salt and an eye on geopolitics. i'm sure there's been an advancement of some sort given the national resources put into 5/6G. and if you don't think national resources have been employed i have a bridge across the Seine to sell you. but really this is just propaganda from a script at the foreign ministry. the market for these claims are the many countries that have opted out or are thinking of opting out from Chinese made IT infrastructure. there is something there, just not what they say.

Don't want to be shitty here but so far 5G is dramatically less capable than advertised. To get those speeds you need mmWave and that can be blocked by even something as innocuous as a tree standing between you and the access point. Without it you're basically stuck at a little over 4G specs. From the 6G claims it's clear that they're still working with the same technical limitations as mmWave so what exactly is this supposed to achieve. Definitely an impressive proof of skill but I'm pretty sure we are expecting and hoping for more. I really hope that when 6G arrives it delivers something that is actually usable for the general public, not something that is purpose built for specific circumstances. Not everyone lives in grasslands surrounded by 5G towers. Some governments actually expected their citizens to avoid going outside for the past few years, something about stagecoach robbers running around wearing fancy scarfs over their mouths.