"The Fourier transform is a well-known mathematical trick that can in essence extract the different colours from an input beam, based solely on the times that the different parts of the beam arrive.
The team does this optically - rather than mathematically, which at these data rates would be impossible - by splitting the incoming beam into different paths that arrive at different times, recombining them on a detector."
At the speed of light, the latency impact of an optical implementation of the Fourier transform operating within the confines of an equipment rack should be negligible. In this case it sounds like they intend to implement the optical computation directly on a silicon chip.
>The trick is to use what is known as a "fast Fourier transform" to unpick more than 300 separate colours of light in a laser beam, each encoded with its own string of information.
No, the "trick" is decades of hard work and research in materials, electronics, photonics, signal processing and more.
Journalists like to believe you can sum up any scientific result in one sentence understandable by mortals. It turns out it's not the case.
It takes a very unfavourable reading to come to that conclusion, IMO.
The particular discovery that is the core of this, is to do this transform. That doesn't imply that (a) doing that is simple (b) getting to the point where you can do the trick doesn't build on decades worth of science.
And, just out of curiosity, what's the alternative to simplification? "It's fast. There's no way you'd be able to grasp why. kthxbye."
I find this an appropriate, consumer-oriented summary right at the start of the article - you cannot expect the BBC News to dive too deeply into the matter and especially not during the first 2 paragraphs.
Also, consider they did add that touch you were missing at the end:
"Think of all the tremendous progress in silicon photonics," he said. "Nobody could have imagined 10 years ago that nowadays it would be so common to integrate relatively complicated optical circuits on to a silicon chip."
And since the scientist they interviewed was German, you should also factor in certain translation issues like a German scientist explaining Fourier transformation and splitting of light beams to a British news reporter.
The intrinsic limit comes from the frequency range for which the fiber is transparent (or rather, the attenuation is low "enough"). And this in turn depends on the composition of the glass the fiber is made from.
The wavelength range currently standardized for telecommunication is 1260nm-1675nm [1], which corresponds to 59 THz as far as I can tell. I don't know which wavelength range Vint Cerf had in mind.
Can someone "compare and contrast" this with existing CWDM and DWDM technologies, which also use different "colors" (think different frequencies, really) traveling over the same fiber?
The team does this optically - rather than mathematically, which at these data rates would be impossible - by splitting the incoming beam into different paths that arrive at different times, recombining them on a detector."
This is cool, and makes sense.