Intel's Silicon Photonics Link boasts 50Gb/s transfers

Matthew DeCarlo

Matthew DeCarlo

Posts: 5,271   +104
Staff

Intel has announced a new silicon-based optical data connection technology that is capable of transferring data at a rate of 50Gb/s and could reach speeds well beyond that. By using a combination of transmitter/receiver chips, lasers and fiber cables, Intel says Silicon Photonics Link is fast enough to transfer a high definition movie in one second.

A transmitter chip composed of four lasers beams light into an optical modulator which encodes data onto them at 12.5Gb/s. Those four beams are then combined and sent over a single optical fiber for a total data rate of 50Gb/s. When it reaches the opposite end of the link, a receiver chip essentially reverses the process, separating the beams and converting data back to electrical signals. By adding more lasers per chip and scaling the modulator, speeds of 1Tb/s are possible.

Intel believes its integrated hybrid silicon optics solution is poised to replace today's copper wiring, which suffers from signal degradation and is said to be reaching its limits. Silicon Photonics Link is still in its very early stages, existing solely as a concept and products based on the technology won't arrive for at least three to five years.

The chipmaker says this project is separate from Light Peak, which is "an effort to bring a multi-protocol 10Gbps optical connection to Intel client platforms for nearer-term applications." Whereas Silicon Photonics research "aims to use silicon integration to bring dramatic cost reductions, reach tera-scale data rates, and bring optical communications to an even broader set of high-volume applications."

Permalink to story.

https://www.techspot.com/news/39790-intels-silicon-photonics-link-boasts-50gb-s-transfers.html

 
7 meg connections are the standard for broadband home access
I mean this sound all cool, but buisness's already get direct fiber links into their offices and those haven't come close to reaching there limit. I don't think anyone really needs this new technology.
 
Man... everybody is going to need this new technology! You won't have a choise either... they will kill old technologies and make you buy new ones...
I once read about an IBM processor that uses light beams instead electrical pulses...that will come out in 2013... I wish I live to see this new technology!
 
@ Yad & paynetrain: I'll never understand sentiments like these, especially coming from people who purportedly like technology enough to register and comment on a technology article, from a technology site.

Do you realize that this technology is ultimately designed to replace copper interconnects in processors, not just beat the pants off USB or Firewire? Copper interconnects are reaching their limit in terms of how closely they can be packed together. Crosstalk, voltage bleed, and heat generation will eventually cause chip manufactures to hit a wall in terms of clock speeds and feature size shrinking.

Lightpeak is simply an early step towards chips that operate entirely with optical interconnects. No heat generation, no crosstalk, no voltage bleed; just pure information exchange at the speed of light. How could you not love that??

And even if they never achieve this, I would gladly take an interface for hard drives and flash memory that ran at 50Gb/s.

Overkill? Too fast?? Never. Give people the bandwidth, and they will find creative ways to use it....

Though most of it will probably still be used for porn.
 
I agree with fizziks, what is wrong with you guys? There will be a day in the near future where 50 gb/s connections are considered slow. If you think this is "overkill" then you have no right being on this site.
 
This is actually pretty cool, I can see the improvements on certain things, imaging this being used with online video rentals such as netflix or blockbuster, rather then waiting hours for that HD movie you can download it and watch it as soon as you pay, how sweet is that. People that call it overkill are retarded and don't know ****.
 
fizziks, I think they need to try and transfer a really large file on some USB stick just to understand the 'pain of slowwwww transfer rates' we are stuck at the moment, data transfers amongst various components is just as much an issue as well.

The fact is, with time as new technology becomes available, software start taking advantage of it until the limits are reached and the cycle of evolution restarts again. I doubt people making such comments have actually used 'internet at 14.4Kps modem speeds' ......... ? I wouldn't bother going further back with even slower connection rates for now.

The simple truth is, there is a time when a certain type of technology is needed to help meet people's requirements/expectations with regard to productivity or pleasure or whatever the objective is, so for 50GB/s it may not be today but the day will come when something like that will be norm of the day.
 
All i was saying is that the technology is great but can't be properly used for our time.
I mean our current cpus,motherboards,ram and most importantly SSD's won't keep up with all that speed.
Hopefully by the time this tech is applied we have a lot of progress in all computer components.
 
I dont understand what you mean yad? Should we not bother trying to improve something just because it is not a bottle neck at this moment in time? This is how progression happens.
 
I see where Yad is coming from, this at the moment is a useless tech because all other components are the bottleneck and things like the Quick-Path Interconnect (X58 chipsets) are not really used to there full potential at the moment anyway, so if you put this kind if tech onto an x58 chipset with the latest CPU's etc you would see no difference at all.

Maybe in 20 years or less though I think these will come into use as CPU's develop and get past the current performance barrier.
 
I guess one thing which is missed in the whole argument, they technology is not yet ready for mainstream, hence it won't show up for next 3-5 years any way.

20 years it bit too far off, having seen pace of changes in the tech industry in last couple of decades, I wouldn't bet against it start to show up in products by mid of next decade.
 
The whole point is that they have found something they think will be beneficial, and now it's off to the design races to make it something that could be utilized in actual applications. As with every advance, there is an idea, a breakthrough, a "see what we can do" phase, and then a long and arduous process of trying to take that cool new tech and making it actually usable. This particular interface and data transfer technique may never see the light of day in a product we use, it all depends on if it can be adapted properly. But, in the process of working on that, there may be more discoveries made that will benefit us down the road. It's progress, and they are (as always) working NOW on what we may be using later. Just imagine if electronics pioneers back in the day had been shortsighted and taken a "that's overkill" or "we'll never need that kind of speed" attitude. We'd still have computers that take up an entire room just to do simple math problems. Instead, they are always looking down the road, and pushing for faster/stronger/better.
 
I agree with Archean and Vrmithrax,
Technological advances have always been utilized in some way, either in and of themselves, or as stepping stones for future innovation.
Almost all advances in computer technology over the last 2 or 3 decades have been utilized in as short as 4 or 5 years, approximately the time it takes for a new generation of computers to come around.
Just look at USB 3.0, Intel is not even making it compatible with current tech, but third parties are already embracing it, with both chip sets and hardware to utilize USB 3.0.

Any innovation will be assimilated into the market very quickly, maybe even more quickly than any of use have dared say... Time will tell, but we probably don't have very long to wait.
 
guys, this is NOT related to internet or home networking, this is SILICON photonics... this is something that is used on the motherboard and inside the CPU, etc.

for example, currently video cards use PCI express 16x links... those are 1gbps/s per link...
While this silicon photonics is 12.5gbps per link.
 
This work on fiber up to 50 meters. That's about 164 feet.

So this isn't for Internet backbone access, but looks like a batter fit for data center's and storage networks.
 
Any technology integration, including PC-s, is built with lots of compromises in mind, where speed of data exchange is one of the first to consider. Introduction of new ways of speeding up data exchange removes the bottlenecks from the older designs, making it possible to create new platforms based on efficiency that is not bound by speed; it is a doorway to a better architecture.
 
thats nothing compared to this https://www.techspot.com/vb/topic19010.html
 
You must log in or register to reply here.

Similar threads

A
Starlink satellites exchange petabytes of data per day through space laser links
Replies
6
Views
258
Endymio
Endymio
A
Amazon's internet satellites will communicate using lasers in space
Replies
6
Views
309
saladbarsmash
S
midian182
China claims to have launched world's first 1.2 terabit-per-second internet backbone
Replies
6
Views
312
havok585
havok585

Latest posts

Back