Multicore Fiber: Just a Cool Design for Now

Avatar photoPublished on
Blog 1535305249 213

Research on MultiCore Fiber (MCF) was first conducted almost four decades ago. It is definitely a real and a cool technology, and a user could eventually purchase greater capacity and density, including more optical communication gigabits per second per square millimeter than with current solutions. Still, despite continuing development work on the concept, especially in recent years, there are enduring challenges over both cost and practical use. On the first matter, the expense would be much higher than a single fiber in that a common configuration is with seven singlemode cores, and so, it is hardly like dividing by n, and it could even be the equivalent of as many as five fibers. On the second consideration, there has to be more of a compelling demonstration of fan-in/fan-out devices. Today the entire infrastructure is set up for 125-micron OD fiber with a single core. Now one would have a system launching from one core of seven cores with perhaps a glass cladding diameter of 150 micron. Intuitively, it would not be a stretch to believe that there would be problems in aligning those cores up, as well as with difficulties with connectorization in mass, in order to keep the insertion losses down to a minimum. In addition, there seems to still be issues with inter-core crosstalk. So, it is not surprising to find back at OFC 2014, when Google’s Network Architecture Team gave a presentation which was on the “Future Needs of WDM Transport for Inter-data-center Interconnections,” and on the slide discussing, “Technology transitions critical for capacity scaling,” it had listed “Single core to multicore or multifiber” last with two question marks.

At the OIDA Workshop on Scalable Integrated Photonics for 5G and IoT in concert with OFC 2018, we asked an executive at Corning about MCF, and he responded that the company “is going and back and forth on the solution, and that it is not really beneficial for distance. As an alternative, the supplier is looking to increase density to accommodate more fiber in cables.”

At Sumitomo, there has been a corporate push for MCF for more than three years. It is among the suppliers, which has been working on one of the relatively few, potentially viable use-cases for Silicon Photonics (SiPh). For instance, with SiPh, the inputs and outputs can be placed at any spacing desired. In particular, the surface emitters using the diffraction gratings from Luxtera actually are effective in an MCF that is in in a hexagonal arrangement (mentioned above) in which there are six cores (second layer) and then one in the center (first layer).

Engineers could actually make that as a surface emission because they could, again, put those gratings at any particular spacing. So, light could be launched from seven different emitters on a SiPh chip, go some distance, and then be received by the complementary SiPh features on the other side.

Nevertheless, it is frequently the case that minor refinements (even quite minimal in nature) more frequently result in a significant jump in sales rather than work being done on a future offering, which has an uncertain timeframe in being widely accepted in the marketplace. There has been speculation in the industry that Sumitomo may have the best fiber for submarine cables by achieving a competitive advantage with a mere one-hundredth of a dB better on loss. Yet, for Google and other undersea network operators going thousands of kilometers, it can add up to making a noticeable difference.

Of course, there are options other than MCF for scaling capacity. For terrestrial applications, we are still partial towards the L and other bands. At the same time, Verizon’s need for such a big increase in fiber to support 5G, MCF happening in a big way in the long-term should not be totally ruled out.

As always, fibeReality does not recommend any securities, and this writer does not invest in any companies being analyzed by us.

To follow us on our totally separate, quick update company blog, which is exclusively on fibeReality’s LinkedIn page, please click here.

[written by Mark Lutkowitz]