At the Photonic Integration Conference 2017 at High Tech Campus Eindhoven, The Netherlands (26 September), professor John Marsh of Scotland’s University of Glasgow (whose research has focused on semiconductor lasers, integrated optics, and their commercialization) will give a presentation emphasizing how the moment for photonic integration in the access market has finally arrived. “It is no longer a question of if photonic integration is going to come to the access market, but more a question of when,” says Marsh.
“The research has been going on since the 1980s but has to wait a long time because, so far, discrete components have been able to do the job. China, Japan and Korea were the first countries to implement passive optical networks (PONs) in the access network on a large scale. Now their networks are migrating from 2.5Gb/s to 10Gb/s, and soon PONs will use multiple wavelength channels at 25Gb/s,” he adds.
“Photonic integration will be needed to do this, but integration needs large volumes to lower the cost and the major companies are in fact a barrier here, as they have not defined a long enough roadmap for the required standards,” Marsh notes. “These standards are needed so the investment can be made in the integration technology.”
The advantages for photonic networks users are huge, Marsh says. Typical PONs are able to deliver 2.5Gb/s, and this will improve to no less than 25Gb/s. Although this data rate is shared by many users, both commercial enterprises and the public will enjoy a 10-fold speed improvement in the delivery of information.
But standardization is also needed to make the required equipment for the public market cheap enough, says Marsh. This also depends on what platforms are chosen, and there is broadly a choice of three: indium phosphide (InP), silicon photonics, and a hybrid form. InP is used in the long-haul network, with Infinera Corp of Sunnyvale, CA, USA being the market leader in photonic integrated circuits. In the short-haul data market, Intel is backing silicon photonics.
These each have their technological advantages, and logically one might expect one of these platforms to be adopted for the access market. However, Marsh thinks there is a good chance that a hybrid-system might prevail. “This can make use of a rather simple and an unexpected technology, the hard-disk drive. In around 18 months, the next generation of hard-disk drives will bring photonic integration technology to the market for a very low cost – you have to implement the photonics technology for one or two dollars in each disk drive,” he adds. “This is the type of major breakthrough that will allow you to make a very complex 100Gb/s transceiver for the access market for $25.”
The global market for optical transceivers is predicted to grow rapidly, from $7bn in 2018 to $10bn by 2021, with all the growth being attributed to high-speed transceivers using photonic integration. Costs must be reduced, however, and the barriers faced by photonics can be compared with historical progress in electronics.
