BBC News reports on a fascinating paper published today in the journal Nature Photonics which describes a technique for transferring data along a fibre-optic cable using a single laser for the light source. A record-breaking transfer rate of 26 terabits per second has been achieved by a team of physicists, with Karlsruhe-based doctoral student David Hillerkuss as the lead author. This, as the BBC helpfully points out, would allow the entire Library of Congress to be sent down an optical fibre in 10 seconds flat. What it means for freetard filesharers on a global scale is anyone’s guess.
The breakthrough is reported by the BBC as being due to the use of 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.”. This statement is potentially misleading, as existing forms of optical communication rely on fast Fourier transforms (FFTs) of one form or another. What is novel here is the high efficiency of the FFT developed by Hillerkuss and his colleagues.
And what exactly is a Fourier transform? It is a mathematical or physical operation that breaks down a time-varying signal into its constituent frequencies (i.e., its spectrum), and vice versa. The basis of Fourier theory is that any signal may be expressed mathematically as the sum of an infinite series of simple waves. In even more abstract terms, Fourier transforms are a means of mapping time and frequency spaces, and in the real world they are ubiquitous.
Telephony relies on Fourier transforms, as do broadcasting, sound recording, image processing and chemical spectroscopy. These particular examples generally incorporate Fourier transforms in digital electronic and computational form, where a numerical algorithm is used to approximate the exact mathematical transformation between the time and frequency domains. That said, Fourier transforms can also be physical, real-world processes, and a classic example is the simple lens, which carries out its work at the speed of light, with nary a mathematical formula in sight.
Data bits are transmitted along fibre-optic cable as discrete pulses of light. A single pulse will have a particular colour, or frequency, and one can increase the data transfer rate by simultaneously transmitting pulses with different colours. This is known as frequency division multiplexing, and traditionally it has involved the use of multiple laser beams to generate the different colour pulses. Fourier transforms then work on the different colours of the input beam, all of which arrive at the light detector at different times.
Any number of lasers can be used in frequency division multiplexing, but cost and equipment bulk are limiting factors. Hillerkuss and his colleagues have managed to get around such economic and engineering restrictions by using very short pulses, within which the light is divided into a number of paths, with discrete colours in each. This is known as a ‘frequency comb’.
The latest advance has taken the established frequency comb concept, and improved it with a fast Fourier transform the details of which are way beyond the scope of this simplified and hopefully comprehensible introduction to Fourier optics. What can be said about the new FFT is that it includes band-pass filters in optical circuits which allow through light with defined frequency ranges. By using such band-pass filters, the computational load in the receiver is significantly decreased, thereby simplifying the process and increasing its overall efficiency.
Improving the efficiency of computer codes and digital signal processing electronics can only be taken so far. What Hillerkuss and his colleagues have done is use optics rather than electronics to stretch out the light pulse time delays. This improves the efficiency of the fast Fourier transform, and allows the transmission of optical data with higher rates than has so far been achieved with single-laser systems. It’s quite a big deal. Really.
Hillerkuss et al., “26 Tbit s-1 line-rate super-channel transmission utilizing all-optical fast Fourier transform processing”, Nature Photonics (2011)