Microsoft acquires Lumenisity!
Lumenisity is a spin out company from Southampton University’s ORC (Optoelectronics Research Centre), set up in 2017 to commercialise a special type of hollow core fibre known as NANF (Nested Anti-resonant Nodeless Fibre) that had been developed at the ORC.
As someone who started their career in fibre optics nearly 40 years ago, working for a spin off company from Southampton University, I have been following the Lumenisity story with interest for a couple of years now. The purchase of Lumenisity by Microsoft is a fascinating new development.
On OTT’s Certified Optical Network Engineer (CONE) course, we have woven the use of hollow core fibre into the case study, considering the application of low latency trading in financial markets. When we first introduced this topic about 10 years ago, it was treated with harsh scepticism from the more pragmatic characters in the case study, but the latest next gen version of the CONE course, (now available in the UK, Ireland, USA, South Africa, & Australia), has softened the approach. It notes the significant developments that have been made by Lumenisity with commercial deployments of CoreSmart® cable containing NANF in real networks such as EUNetworks in London. CONE course delegates can find further details on the CONE online course resources Field Guide: Special fibres.
What’s the plan?
It is interesting to speculate on Microsoft’s plans for their new acquisition.
- Will they restrict the use of this fibre to deployments within their own networks? Thus providing a source of competitive advantage over the networks of other hyperscalers.
- Will they continue to supply fibre to the very lucrative financial services market? This would provide a tangible return on their investment.
- With the financial backing of Microsoft, will it be possible to scale up the manufacture and deployment of NANF cable? This might make the benefits of hollow core fibre more widely available than is currently the case.
However, there are many technical challenges to scaling things up, all the way from making this very complex fibre, cabling it, installing it, splicing it, testing it (OTDRs don’t work quite the same when there is no glass to scatter the light!). Whilst Lumenisity have addressed many of these issues, for the modest installations that have been carried out so far, doing things at (hyper) scale is a different ball game.
So, what is it about Lumenisity and their Coresmart® cable with NANF that has caught the attention of Microsoft?
Well, to understand the full story, of course it would be best to go on the Certified Optical Network Engineer (CONE) course. But in the meantime, here are some of the key technical benefits of NANF and how they might help Microsoft with their network.
What are the benefits of NANF fibre?
The big thing about hollow core fibre is that the light travels through air, rather than glass, so the light travels 50% faster (300,000km/s rather than 200,000km/s). This results in lower latency for a fixed length link, or allows a longer link length with the same latency. So for the clusters of data centres in a region where the distance between data centres may be latency limited, as is the case for Microsoft, the maximum distance between data centre may be increased by nearly 50% (some aspects of the latency within equipment are fixed (attend CONE course for details!). This can allow a lot more flexibility in where you can site the data centres, potentially allowing the use of much cheaper locations. Given the costs of data centres and the number of them that Microsoft will be building, this might pay for the acquisition of Lumenisity in itself!
However, these longer distances can only be achieved if the loss of cabled hollow core fibre comes down to similar levels to conventional silica fibre. This is where the Lumenisity NANF technology has shown significant potential compared with other hollow core fibre technologies. Currently, cabled fibre performance is below 1dB/km and a 40km field trial has already been conducted by Comcast. But NANF performance as low as 0.22dB/km has been reported at conferences. So there is the promise of longer distances.
Actually, low latency isn’t the only interesting characteristic of this fibre, it can also tolerate very high power levels, with little or no non-linear impairment that limits conventional silica core fibres (non-linearity is another topic that we discuss on the CONE course). Without these non-linearities, we do not need to have high levels of chromatic dispersion to combat four wave mixing in closely spaced, high powered DWDM systems. The NANF is engineered to have low dispersion of a few ps/nm/km over the whole useable wavelength range. Low dispersion reduces the requirement for optical dispersion compensation on direct detect systems, and low dispersion also eases the task of the electronic dispersion compensation (EDC) aspect of the digital signal processing (DSP) in coherent systems.
NANF has been demonstrated to be capable of transmitting standard telecoms traffic including Ciena 400G DWDM traffic over 1000km in an amplified re-circulating loop. So it is not necessary to start from scratch with different transmission technology to use this fibre.
Of course the biggest downside of NANF to date has been the production cost. Whilst it is possible to manufacture 1000km or more of conventional fibre from a single preform, the lengths of NANF that can currently be made from a single preform are less than 10km. This is why the massive cost associated with hollow core fibre has only been justified, so far, for the lucrative high frequency trading market, where the latency reduction has allowed some to make faster trades than their competitors, even if it only by a few nanoseconds.
So how much will change with Lumenisity under the Microsoft umbrella? I’ll continue to follow developments closely and update our training courses appropriately! Perhaps we’ll see you or your colleagues on a course soon….



Recent Comments