Lighting all that fiber that connects all the world is the gear produced by the optical sector. The latest 400G, 600G, and 800G technologies are in various stages of growing up and will soon carry the bulk of that traffic. I remember when Infinera first broke onto the scene with its PICs two decades ago, and they continue to push the pace in a neighborhood dominated by much larger competition. With us today to talk about what’s going on at Infinera and in the optical industry as a whole, whether it’s about technology or the pandemic, is Rob Shore, SVP of Global Marketing.
TR: What is your background, and how did you get started with Infinera?
RS: I started with Tellabs back in 1993, where I spent about eight or nine years in engineering across multiple groups, including testing, software development and customer support. Then I moved over to the dark side into market management and then into sales. In 2013, Tellabs got acquired by Marlin Equity Partners and combined with Nokia Siemens Networks and Sycamore to form Coriant. Coriant had some really interesting concepts but didn’t own much of its own intellectual property. I think Infinera, who did own its own IP, recognized the untapped potential, which is one of the reasons they acquired Coriant. I joined Infinera as a result of the acquisition. Infinera has a really interesting story, not only because of where we came from, but because of how we are working to move the industry in a new direction.
TR: Where in its portfolio is Infinera investing development resources right now?
RS: Infinera has a broad portfolio of end-to-end transport network connectivity solutions. That includes everything from layer zero through layer three and all the way from the edge to the core and subsea. I don't think we're looking to change any of that. What we're really focused on is trying to help the industry evolve from highly proprietary, single-vendor network solutions into a more open and disaggregated approach where network operators can pick and choose the appropriate building blocks and get the best-in-class of technology at each layer for each application. We are aggressively investing not only the platforms to make that happen, but also on the transition plan of how to networks get from here to there.
TR: What benefits do you see network operators deriving from this new vendor model?
RS: There are really two primary significant benefits from this new, more open model. Not only does it enable network operators to build better networks by choosing the best-in-class for each particular function, but it enables innovative solution providers like Infinera that might not already sell everything from soup to nuts like a Cisco, Huawei, Ericsson, or Nokia to break into the business with innovation application specific solutions and improve the industry as a whole. It will lower the cost of the network, enable new features and services, and allow network operators to customize the way they build their network and the solutions they offer in order to differentiate themselves from their competitors. Networks built on one vendor tend to look very similar. With open networking, rather than fitting their service to the constraints of a single vendors solution set, network operators can start with what they want to accomplish with their network, they can, for example, put together one part from Cisco, one from Ciena, one from Infinera, and create exactly what they want in order to accomplish their particular goals. It also enables them to get access to newer, innovative things more quickly. When everything comes from a single vendor, innovation can only move at the speed of that one vendor. When multiple vendors can focus on one particular area, the innovation cycles accelerate substantially.
TR: How much of the ability to disaggregate network architecture today comes from the rise of the software side, and how much remains on the hardware side?
RS: Optics will always be very hardware-centric, unlike servers or even routers, which are more software-centric. No matter how good your software is, you can't virtualize a laser. But there are a couple of ways in which software makes a big difference. One is the idea of standardizing interface languages and data models. That enables much easier management of multiple different vendor devices because they all look and feel very similar. Therefore SDN can more easily orchestrate services across a multi-vendor network environment. However, for optical networking, the biggest enablers of disaggregation comes from the hardware side. In optical networking, there is the optical layer that moves the light around, and there’s the transponder layer that generates the light. Historically, those two things needed to be very tightly coupled together, making the concept of an alien wavelength, or multi-vendor networks, very difficult to implement. In recent years we have seen a lot of evolution on each of those layers. On the optical layer, you see the introduction of things like flex-grid ROADMs, more built-in optical sensors, and much more flexible control mechanisms. The optical layer has learned to operate more independently from the transponder. On the transponder side, the was a significant shift in capabilities with the advent of coherent based transmissions and the introduction of digital signal processors. Digital signal processors not only provide better performance, but also enable the transponder to adapt to various environments more effectively. So, advanced coherent transponders can now talk across the network, analyze the resulting performance, and automatically make the necessary adjustments for optimization, all completely independent of the optical line system. The net result is that is has become quite easy to deploy virtually any vendors transponder over virtually any vendors optical line systems effectively eliminating vendor lock-in for optical networks.
TR: Now that we are seeing 400G, 600G, and even 800G make their way toward commercial deployments, what is the next stage of technology? In what direction should we be looking?
RS: I don't want to jump the gun too much. After launching a high-speed optical technology, it takes 2-3 years to see large scale deployments. This is the same for all technology sectors. Right now we're starting to see the ramp-up now for 600G. And while 800G is just becoming available now, you're still going to see several years of ramp-up of 600G before 800 Gig really gets going and captures a large portion of the market share.
Now, what’s after 800G? For Infinera, after ICE6 800G solutions, what we're going to see is, unsurprisingly, ICE7. And while we’ve already begun the planning cycle, this technology is still several years away. However, it is interesting to know we are just starting to hit a point of diminishing returns on each successive generation of optics. There are two ways you can improve the cost effectiveness of optical transmissions: increase the baud rate and increase the modulation rate. At 64QAM, we are pretty close to end of the road for increasing modulation rates, which is the more powerful tool as it increases both capacity per transceiver and spectral efficiency. Increasing baud rate improves the capacity per transceiver but does not provide a benefit to spectral efficiency. The problem with increasing modulation is that higher modulation rates result in worse optical performance – less reach. ICE6 was a huge step forward compared to other technologies in that generation where you can have a 64-QAM signal, 800G signal reach 850 to 1,000km in real world network deployments, which is significantly better than other vendors in this space. What we see in the next generation will likely be 64-QAM for longer reach transmission with higher baud rate providing a more cost effective, if not more spectrally efficient solution. 128-QAM or higher, might see some deployments for very short reach applications.
TR: So if 64-QAM is the end for modulation, where is Infinera looking for the next stage of innovation?
RS: First – as Infinera has always done, we plan to continue to add innovative capabilities to our high-speed optical engines to improve optical performance and provide the lowest overall cost per bit solution over any distance. However, another interesting solution that we’ve been working on is a bit more out of the box type thinking. If you think about optical networking – all solutions have operated under the same base premise: optics are point-to-point connections. A laser on one end of the fiber and an identical laser on the other end of the fiber. The issue with this traditional solution is that it doesn’t match actual network traffic patterns. If you think about the way bandwidth flows in a network, particularly in a metro, you've got hundreds or thousands of endpoints and just a few locations are trying to connect to. That is definitively a point-to-multipoint type of traffic flow. So, if I’m aggregating traffic from 10,000 endpoints, with today’s optical solutions, I need 10,000 lasers at the endpoints and an equivalent 10,000 lasers in the aggregation point. That's a pretty crazy way to build a network, and yet we've been doing that for 50 years. Mobile networks realized you can't do that. A cell tower needs to talk to a thousand cell phones but there’s just one radio on the cell tower that broadcasts in a broad spectrum, and then it coordinates with each handset what portion of that audio spectrum they get. So, what we’ve done is fundamentally rethink that concept of how optics work and pioneered a technology we call XR optics which is the industry’s first coherent point to multi-point optical solution. XR optics brings the same type of concept used in mobile networks into the optical domain. If I have a laser that can generate a signal within an optical range of, say, 64 gigahertz, I can subdivide that up into smaller chunks and then distribute those to different destinations. Therefore, I can have big lasers talking to small lasers, which has never happened before. I think this is the next big evolution for optics.
TR: How has the COVID-19 pandemic affected Infinera and the industry as a whole so far this year, and how do you expect the rest of the year to play out?
RS: There were two impacts that COVID had on the telecom market. The first is the significant impact to the supply chain hindering the ability to build equipment, ship solutions, and install them. And while I think the supply chain is has largely recovered, it created a huge backlog, because demand didn't go down but the industry's to meet that demand was challenged. I think through the second half of the year you're going to see the industry in general addressing the pent-up demand that the network has experienced. One thing that we’ve seen at Infinera as a result of the supply chain disruptions is a significant up-take of our instant bandwidth feature which enables network operators to activate latent bandwidth in their network remotely through software. The second significant impact COVID has had on the market is the changes in traffic patterns. With people working from home, more video conferencing and a higher reliance on the cloud we’re seeing both an increase in traffic volume and more dynamic and shifting traffic patterns. We’ll have to see how this plays out over the long term, but it is driving network operators to reconsider certain aspect of the way they build their networks to make them more flexible, scalable, and dynamic.
TR: Do you foresee any additional disruptions over the next six months if COVID goes sideways?
RS: I really don't. I think for COVID we've largely tackled this. Even if there's a major outbreak, we've learned how to deal with it better, how to continue operating even when there's issues that require social distancing, personal protective equipment, or other safe practices. I foresee us only continuing to become more efficient in the COVID era, because I don't think this is going away for the reasonably foreseeable future.
TR: Are there changes from these pandemic times that you think will be permanent?
RS: As you know I run a marketing organization at Infinera. One of our functions is tradeshows and events. This is one area that I don’t think will ever be the same again. We will go back to in-person events at some point but they're going to have a very different look and feel, and I think maybe half of the traditional large-scale industry events will survive. This will be both a big shift toward more virtual events and toward more small-scale personalized events. I don't mind that personally too much because what I like is deep interactions with customers, and what you often get at trade shows are cursory interactions. Trade shows are good for getting big brand-type messages out there, but they are not great for promoting specific business initiatives. What we've been doing a lot more is those individual customer events – what we call Infinera Xchanges. Another impact of COVID, is that I think the industry in general, and Infinera is not an exception, is starting to realize that we don't need big giant office buildings that everybody goes to every day and where everyone has their own assigned desk. People are learning to work from home and to be effective. You still need meeting spaces, because people still do need to get together with a whiteboard and interact sometimes. And there are always things you simply can't do from home, like certain aspects of research and development and, obviously, any kind of manufacturing. But I think there’s going to be a big shift out of large commercial real estate into much smaller shared facilities -- much more open office environments where you don't have many assigned desks. You just show up and pick a desk or grab a meeting room.
TR: Thank you for talking with Telecom Ramblings!
If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!Categories: Industry Spotlight · Telecom Equipment