Engineers Set New World Record for Internet Speed (gizmodo.com) 43
"Imagine being able to download every single movie and TV show on Netflix in less than a second," writes Gizmodo:
Researchers at University College London have the ability to do that with a new world record they set for fastest internet — 178 terabits a second, or 178,000 Gbps. Lecturer and Royal Academy of Engineering Research Fellow Dr. Lidia Galdino and team collaborated with Xtera and KDDI Research on the project. According to UCL's announcement, that speed is "double the capacity of any system currently deployed in the world."
To get that insanely fast speed, UCL researchers used a greater range of wavelengths than what's typically used in fiber-optic cables and different amplifier technologies to boost the signal. Fiber-optic cables tend to absorb signals (well, the photons that are transmitted through the cable to make the signal) after a few miles because of the material the cables are made out of. Repeaters, which are like a wifi extender, are needed to re-transmit those signals so they can travel for a longer distance. So what the researchers managed to do is not only extend the signal, but also massively amplify it... 5G on the high-band or millimeter wave spectrum operates on 24 GHz and above and can transmit data up to at rate of 1 to 3 Gbps. But the internet speed Dr. Galdino and team achieved uses a 16.8THz bandwidth to get 178,000Gbps. Makes 5G seem rather slow when you put those numbers side by side.
This kind of system would be cheap to integrate with our existing internet infrastructure, too. According to UCL, upgrading amplifiers at certain intervals would be a fraction of what it would cost to install new optical fiber cables, roughly $21,100 every 25-62 miles (40-100 km) versus $594,000 every 0.62 miles (1 km), based on today's conversation rate of £1 to $1.32). This sounds like it could be a worthwhile solution to help shrink the digital divide, something that the current pandemic has further illustrated the seriousness of.
To get that insanely fast speed, UCL researchers used a greater range of wavelengths than what's typically used in fiber-optic cables and different amplifier technologies to boost the signal. Fiber-optic cables tend to absorb signals (well, the photons that are transmitted through the cable to make the signal) after a few miles because of the material the cables are made out of. Repeaters, which are like a wifi extender, are needed to re-transmit those signals so they can travel for a longer distance. So what the researchers managed to do is not only extend the signal, but also massively amplify it... 5G on the high-band or millimeter wave spectrum operates on 24 GHz and above and can transmit data up to at rate of 1 to 3 Gbps. But the internet speed Dr. Galdino and team achieved uses a 16.8THz bandwidth to get 178,000Gbps. Makes 5G seem rather slow when you put those numbers side by side.
This kind of system would be cheap to integrate with our existing internet infrastructure, too. According to UCL, upgrading amplifiers at certain intervals would be a fraction of what it would cost to install new optical fiber cables, roughly $21,100 every 25-62 miles (40-100 km) versus $594,000 every 0.62 miles (1 km), based on today's conversation rate of £1 to $1.32). This sounds like it could be a worthwhile solution to help shrink the digital divide, something that the current pandemic has further illustrated the seriousness of.
Clearly no Moore's law yet. (Score:2)
Only a few months ago a story here was run about a speed record being set in fibre at 44Tbps in Australia. Now in London we have 178Tbps, a 4x increase in speed. I'm glad we're not reduced to minor incremental improvements yet.
Re: (Score:2)
And with Spectrum Cable, this page still took two seconds to load.
Re: (Score:2)
"Internet Speed" - I don't think those words mean what the article writer thinks they mean.
They're talking about bandwidth, not speed. Also, it was in a lab, not a connection to the Internet.
So "Engineers Set New World Record for lab bandwidth".
Re: (Score:2)
And? Doesn't change my point. The previous 44Tbps was also lab record bandwidth.
I bet Comcast is salivating (Score:5, Insightful)
at making hundreds of millions of dollars in less than a second thanks to their shitty data caps.
Re: (Score:3)
Monopolies of scale.
What will conspiracy theorists be saying? (Score:3)
There are groups warning us that 5G gives us Covid-19 infections (or is it cancer: I find it hard to keep up). That is at a puny speed of 1-3 Gbps. At the insane speeds described in this research, is it going to turn everyone into werewolves?
Re: (Score:2)
is it going to turn everyone into werewolves?
Hopefully
Re: (Score:2)
Re: (Score:1)
Better analogy (Score:5, Insightful)
"Imagine being able to download every single movie and TV show on Netflix in less than a second," writes Gizmodo:
More like imagine hitting the Xfinity data cap in 50 milliseconds, then paying Comcast $4,450 in extra charge every subsequent second.
Nothing to do with digital divide (Score:2)
This is interesting science which may have practical application, but it has nothing to do with the so-called "digital divide." The digital divide is a social phenomenon caused by some people, or whole classes, neighborhoods, or regions of people, who simply can't afford a high speed connection, or don't have one available. Making fiber optics faster won't solve this problem. It will take a much larger buildout of regional and neighborhood infrastructure, coupled with some way of making access more affordab
Re: (Score:2)
Re: (Score:2)
It would be more interesting if they'd test it at the salt flats, maybe somebody would crash a rocket car into it or something, make some news.
The headline is as dumb as can be, it isn't the internet that is super-fast, it is just one private hop that is that is that fast. Their internet is still internet-speed. Their total bandwidth went up.
The first hurdle (Score:2)
If that is given, I suppose replacing emitters and receivers (including repeaters) is not such a big deal in many cases.
But this simply isn't a given, which is why we have such projects as SpaceX' Starlink. In my eyes this is the major factor in the digital divide. It's not that those people lack bandwidth in their fibre connections, it's that they don't have fibre connections.
Every movie and TV show? (Score:3)
Re: (Score:2)
Yeah, I was also figuring that NetFlix has more than 23TB of movie data.
(Pedantry: yes, it's actually 23 TB exactly - there's no mention of protocol overhead :) )
Re: (Score:2)
Re: (Score:2)
It is not 22.25T bytes. We don't know what the number is as we don't know the coding overhead, but that number is definitely the wrong one.
You're on a technology site, pedantry is never pedant enough. :p
Re: (Score:2)
It is not 22.25T bytes. We don't know what the number is as we don't know the coding overhead, but that number is definitely the wrong one.
You're on a technology site, pedantry is never pedant enough. :p
It is exactly 22.25 TBytes. We just don't know how much of that is payload.
(since you asked for more pedantry :) )
Re: (Score:2)
However, even if you could download it that fast I don't know what you're g
Re: (Score:2)
I was doing a thought experiment and looked up the information. Netflix as of a few months ago had about 36,000 hours of content. Their own website says 3 GB per hour (though they do have some content in UHD which I believe is 7 GB per hour). Assuming it's all HD then that's 108 TB. So, just about 5 seconds. SInce it's not all UHD I'll say you can easily do it in 10 seconds for sure. But, your're right definitely not in 1 second. However, even if you could download it that fast I don't know what you're going to store it on as it looks like DDR5 tops out at about 51.2 GBps and a good SSD only has a write speed around 520 MBps :).
You'd have to store it on a Beowulf cluster.
Re: (Score:2)
Didn't happen (Score:1)
Re: (Score:2)
You can't even handle that much porn.
Pull Fiber - Future Proof... (Score:2)
This is why I've said for about the last decade that you should always pull single-mode fiber cable over copper whenever possible, even if it costs a bit more if you want to future proof your install. The longer the run you are making, the more important this is.
I was working on a military aircraft upgrade and this was my recommendation. Stop pulling copper if at all possible, pull fiber. It's lighter, less susceptible to interference, and future proof. You only pull copper when you need power, any oth
Re: (Score:3)
Signal integrity isn't going to be the same of course, but it will work in a limited capacity.
For a fibre optics cable that's broken to be spliced together again, the equipment required is a lot more complicated. It's not that suited for a quick field fix. Getting a single mode fibre to work again even in a limited capacity requires
Re: (Score:2)
It's a good idea, until some some idiot comes out with Lightbeams 2.0 and your optical fiber isn't compatible anymore.
What the (Score:2)
"Imagine being able to download every single movie and TV show on Netflix in less than a second," writes Gizmodo
If you just downloaded the good stuff you could do it in less than an attosecond.
Re: (Score:2)
It would probably take a while just to read the whole list of *titles*. And since it would take a good fraction of a year [leicestermercury.co.uk] to watch them all, what's the point? If you could download all the videos into your *brain* in a month -- now you're talking.
Apples to Oranges. (Score:4, Informative)
Fiber service life (Score:1)
Shannon-Hartley theorem (Score:1)
Up to 178,000 Gbps* (Score:3)
*Speeds may be lower depending on location, day of the week, time of day, sun, wind, whether Aries is in ascendence, the phase of the moon, &/or the thickness of the interweb tubes.
Yours sincerely,
Your only neighbourhood ISP.
How mang football fields long is it? (Score:2)
Repeaters, which are like a wifi extender, are needed to re-transmit those signals so they can travel for a longer distance.
SMH. Is this how we should be explaining technical subjects to lay people?