UK Scientists Build 10Gbps Receiver to Boost Fibre Optic Broadband Lines

Scientists working out of the Optical Networks Group at University College London have developed a cheaper and simpler receiver for optical “full fibre” (FTTH/P) broadband networks, which makes it possible to dedicate a wavelength of light to each subscriber and “avoid bandwidth sharing.”

Admittedly it’s already possible to do what is being described above by using highly sensitive hardware known as coherent receivers, but these are costly and the UCL team say they’re “only financially viable in core networks that link countries and cities” (i.e. not much use for the local access network between networks and subscribers).

Modern networks, even the latest fibre optic ones, can be complicated beasts under the hood (see the example diagram below) and by adopting the benefits of coherent receivers into something that comes in a smaller, simpler and cheaper package then this could be improved (the new kit requires just a quarter of the detectors used in conventional receivers).

Apparently this simplification was achieved by adopting a coding (alamouti) technique to fibre access networks that was originally designed to prevent signal fading in wireless communications. This approach has the additional cost-saving benefit of using the same optical fibre for both upstream and downstream data.

The coding scheme removes the need for any optical polarisation tracking unit in the receiver, which leads to a “significant reduction in optical complexity” compared to the conventional polarisation and phase-diverse coherent receiver architecture, allowing the implementation of the proposed simplified coherent receiver using the minimum number of optical components possible. In plain English..

Dr Sezer Erkılınç, Lead Researcher, said:

“This simple receiver offers users a dedicated wavelength, so user speeds stay constant no matter how many users are online at once. It can co-exist with the current network infrastructure, potentially quadrupling the number of users that can be supported and doubling the network’s transmission distance/coverage.”

Professor Polina Bayvel, Co-Author, said:

“BT Openreach recently announced that fibre access is a key focus and must improve. With high-capacity broadband a priority for the UK government, we will be working to reduce the electrical power requirements of this technique to make this commercially viable in the nearest future. We believe that it has real potential to provide high-speed broadband connectivity to every home, which will support the growing digitally enabled economy in the years to come.”

The receiver has already been tested on a Dark Fibre network installed between Telehouse (east London), UCL (central London) and Powergate (west London). The team successfully sent data over 37.6km and 108km to eight users who were able to download/upload at a speed of at least 10Gbps (Gigabits per second).

Essentially this would create a more efficient network, although it’s unclear how easy it would be to adopt into today’s already complicated networks and heavily regulated environment. On top of that we also take issue with a few of the claims that the team make.

Dr Erkılınç notes that average UK speeds are currently 36Mbps and he claims that, by 2025, speeds of “over 100 times faster will be required to meet increased demands for bandwidth-hungry applications such as ultra-high definition video, online gaming, and the Internet of Things” (i.e. 3.6Gbps). Firstly, 36Mbps reflects Ofcom’s speedtest based study and so doesn’t accurately reflect the availability of faster networks.

On top of that we’d be very surprised if everybody needed 3.6Gbps by 2025 for online gaming (latency is more of an issue here, not raw speed, and existing connections do fine) or UHD video streaming (a 4K stream needs about 20-30Mbps and future 8K + HDR streams may push to around 100Mbps after compression). As for the so-called Internet of Things, most IoT kit tends to be fairly low bandwidth.

The study also appears to suggest that performance drops at peak-times (busy periods) would no longer be a problem with their approach, although such issues are not merely the fault of the complex local access infrastructure. Capacity, especially from international connections, will always have to be shared to some degree for residential users and the sake of affordability.

Those of a more technical mind can find all the details here.