Executive Forum 2018 Recap
Dr. Scott Wilkinson
*This article originally appeared on the ECI LighTALK Blog
Major Themes from the Optical Society’s pre-OFC meet up
The 2018 edition of the OIDA Executive Forum in advance of OFC has wrapped up, and it is hard to determine any one major theme of the program this year. Once again, this was a worthwhile event with perspectives given from a wide variety of industry movers and shakers. However, it seems that we are at an interesting point in the optical industry where requirements are somewhat fragmented and development is a bit scattered. There are perhaps too many major points of focus to definitively come up with a major theme.
One candidate for a major theme was mergers and acquisitions, thanks primarily to the Lumentum/Oclaro announcement on the morning of the event. For years, many in the industry have held that the components sector has too many players to be efficient and has needed consolidation. The problem, as expressed by Finisar’s Jerry Rawls, was that no one player had the capital on hand to acquire anyone else. That changed with Lumentum’s recent rollicking success in 3D sensors (the kind that allow your phone to recognize your face and turn it into an animated figure). Several speakers mentioned the enormous costs of silicon photonics – the engine behind the latest optical modulators – as they move to smaller and smaller feature sizes. The cost of running a design through to manufacture with 7nm feature sizes is more than $30million. If there are several companies all trying to make that leap at the same time with competing designs, it’s easy to understand the benefits of consolidating R&D.
Another candidate for a major theme was 5G, as it has been at every show and presentation for the last year. Verizon and China Mobile presented views on how their optical network designs are being impacted by the move to 5G (AT&T had to drop out due to a legally mandated quiet period, an absence that generated its own set of rumors). Fronthaul and backhaul are both very interesting topics with many potential solutions, and each had their own take on the subject. What was interesting is that both Verizon and China Mobile see PON technology as a significant part of their backhaul strategy. They have chosen different PON technologies (NG-PON2 and XGS-PON, respectively), but they both believe that they will be able to use their embedded PON infrastructure to enable small cell 5G deployments. This is good news for fiber and passive components companies as well as for the few remaining PON vendors in the market. It may even be a good thing for businesses and residences who will have fiber pushed even more densely into their neighborhoods. It likely spells the end of what Verizon referred to as “ugly” CPRI and a move to something more compatible with the rest of the telecommunications infrastructure (like Ethernet), which is good news to everyone except the engineers who were tasked with designing CPRI interfaces on CWDM.
Silicon photonics is always a candidate for a major theme when a bunch of optical components suppliers get together to talk. The intersection of Moore’s Law and Shannon’s Limit (hat tip to Cisco for the reference) is an interesting place to play these days. As Benny Mikkelsen of Acacia put it, we have moved from the TDM era to the WDM era and are now in the silicon era. Coherent detection and complex modulation have had as big of an impact on the optical industry as the move to multiple wavelengths did in the late 1990s – hopefully without the resultant excessive exuberance and bubble. A panel discussion that brought together optical component supplier Acacia, switching silicon supplier Broadcom, equipment vendor Cisco, and network operator Microsoft in the name of silicon photonics was perhaps the most interesting of the day. The struggle between the optics and the switching chips to keep up with each other is fascinating to watch as speeds and capacities continues to increase within the data centers. The ability of a few major operators like Microsoft to create their own standards and get their specific needs met is a challenge for the industry. The question of how close optics will move to the switching – on board or even on-chip – reminds me of my university days with its science experiment and anything-is-possible feel. There is a lot to talk about with silicon photonics, especially for data center business. But I would rather not claim that, once again, the data center needs are a major trend. They are, but it’s getting repetitive to continue to bring it up.
Another trend that is repetitive is the dominance of Asia in the optical market, although it is still very much true. The kinds of builds that are going on in China are still impressive, although not quite as impressive as a few years ago. Presentations from China, Japan, Korea, and Southeast Asia were all very interesting as those perspectives sometimes get lost in the latest Web2.0 frenzy. Several CEOs mentioned their investments in Asia and discussed concerns that China has stated that they want to be more vertically integrated in-country. Overall, however, the view is that we are all in this optical business together and can learn from each other’s experiences. So long as Asia continues to buy a lot of optics, the industry will remain in good health.
Perhaps careful optimism is the overall trend from this year’s Executive Forum. The pessimism of a few years ago is largely gone and folks are legitimately excited about optics again. There are interesting innovations in technology and applications, 5G is coming quickly, and the demand for bandwidth continues to rise. We have reasons to be glad that we work in this industry at this time and we have hopes for many years of growth in the future.
