Researchers discover new method to send data across fiber optic lines

Shawn Knight

Posts: 15,287   +192
Staff member

bell labs 400gbps

High-speed Internet – I’m talking data links between continents – is already blazing fast compared to what most people are used to but researchers have found a way to make it even faster by improving the overall reliability of the signal.

The problem with traditional fiber optic lines is that they have to transfer data over a very long distance. Doing so means you’ll need a lot of power to help push the data along but in the process, interference (or noise) is added into the equation.

A common method of handling the interference is by inverting the noise to its phase conjugate. The entire process is most easily understood in relation to how noise-cancelling headphones work. These headphones are able to listen to the surrounding environment and create an inverse signal to help prevent the noise from being heard by the wearer.

It seems easy enough but implementing it into a network scenario is tough considering the infrastructure of the Internet. Such methods have been proposed but they are ultimately impractical and too expensive to deploy.

A researcher by the name of Xiang Liu and his associates discovered a method to make noise filtering easy. By sending the original data as well as the phase conjugate at the same time, interference in one wave is mirrored by the other and both are canceled out. Ultimately what this means is faster connections and better reliability.

To demonstrate this, the team recently sent data nearly 8,000 miles at speeds of 400Gbps with few errors. Less errors means more data can be sent as the originals wouldn’t have to be resent as often.

Permalink to story.

 
Uhm.. this is using the exact same principle employed in balanced audio (XLR), humbucker pickups in electric guitars, etc..
 
I don't understand the concept of noise canceling a digital signal. When a digital signal is comprised of only 0's and 1's, how can you cancel out interference? It's not as if a third state in binary can be introduced to counter a stray signal.

Perhaps the answer lies within my confusion of phase conjugation.
Phase conjugation is a physical transformation of a wave field where the resulting field has a reversed propagation direction but keeps its amplitudes and phases.
 
The interpreted signal is digital (0's and 1's) but the real signal is analog. When noise is added a binary digit might be interpreted wrong so a dirty zero become a 1 an vice versa. Hope this helps :)
 
Back