LightRiver has long labored behind the scenes of network design and deployment for service providers across the industry. With the rise of automation and software-driven infrastructure, they have begun to move deeper into the actual operation and maintenance of those networks as well. With us today to talk about the company’s approach to the ever-moving target of network infrastructure design and operation is newly minted CEO Mike Jonas.
TR: What led to your appointment as CEO at LightRiver?
MJ: We planned it over a period of months. Glenn was looking to launch his retirement, and we were looking to launch a new phase at LightRiver that increased our focus on both the recurring revenue products that we’ve been delivering the last 18 months or so and some additional geographic expansion.
TR: What does the next phase look like for LightRiver?
MJ: We are doubling down on the commitment to support our customers in their network operations. We started off by designing networks for folks who were light on design resources. Then we started deploying networks for folks who needed to accelerate their network delivery. Then they asked us to automate those networks. Now with the growing shortage of technical workers in every part of the country, not just in telecom, they’re now asking us how to use the network intelligence and automation that we’ve designed into the network to help streamline network operations. They want to reduce their OPEX while they grow their own footprint, and so we have introduced netFLEX-as-a-Service. We have been delivering netFLEX via the classic software model to a number of large tier 1 providers. But there is an entire class of emerging carriers that just don’t have the staff, the financial wherewithal, or the desire to build and staff an internal platform for their network ops. So, we have stood up two netFLEX server instances, one here in the United States and one in Europe, for tier-2 and tier-3 network operators. They can acquire network intelligence via that model, or they can have us use that connection to monitor the network for them. Further, several are now asking us not only to monitor but to do the troubleshooting and the analysis, and to recommend corrective action. And half of those are asking if we can just fix it for them. Those are investments we hadn’t made historically, but we are making them now.
TR: Adding those capabilities will be resource intensive. How will you approach the expansion?
MJ: We are not trying to do it in the traditional way, which is to just send a truck and see if somebody can figure out what’s broken and then send another truck out with the right people to fix it. If we’re connected to the network, netFLEX is always listening for network anomalies so we’re going to be the first ones to know, even before it breaks, that a threshold has been crossed. We will get an alert that light has degraded, there is a micro bend, or there is congestion, or whatever. If under contract, we can then log in, diagnose the exact problem, the location, and the fix, and we can send the exact right tools to the exact right location to perform a preventative update to avoid an outage. That keeps us from having to deliver personnel resources in the middle of the night in an unplanned fashion. The model here is to be intelligent about it, to use network surveillance and intelligent automation to shortcut that entire legacy model of just throwing bodies at problems. Some of the folks in this frontline maintenance space have thousands of people on contract who are available to go anywhere, sometimes just to shine a flashlight on stuff. We have the tools to avoid that model and we’re going to.
TR: Fewer truck rolls still means some truck rolls, which require assets in place spread across a wide geographical area. Will you be adding those resources in-house?
MJ: We are bringing on people. We are continuing with our accelerated certified engineering program, the Young Guns program. We are now in our 17th year of recruiting at least half a dozen hardware engineers right out of university and training them in our network factory on how to do these projects correctly. We continue to use engineers whereas most people use lower-end techs. Those engineers are armed with every piece of network data that they can possibly carry. We’re still showing up, but we spend a quarter of the time or less of what some of our competitors are spending on site, both to deploy the network and to maintain it.
TR: For now you are doing this on networks you have designed and built. Do you have plans to also do it for other netFLEX-enabled infrastructure?
MJ: Probably one third of our customers today are netFLEX only. We haven’t yet initiated engineering-as-a-service to support those networks. We’re building operational profiles for those customers, and if we can get comfortable that those profiles allow us to be valuable to our clients, and effective, in this day-two role, then we’ll start doing that. But I would say you’re not going to see that until 2025.
TR: Do you foresee an inorganic path to expand your workforce to accommodate these new capabilities?
MJ: We are in the private equity world, and it’s about math. What makes the most sense? When does it make sense? And how would you achieve it? Almost anything is on the table if it makes good sense to our customers and our shareholders.
TR: What does LightRiver’s customer composition look like today? Are there any new infrastructure segments you’d like to get into?
MJ: We are still primarily focused on three large markets for fiber optic network connectivity. Our largest market cohort is still telco and carrier, followed by cloud and data center. And then we still have a very meaningful business in the regulated utility market. And that market is now actually growing faster because it is being augmented by the emerging broadband projects coming out of the cooperative and municipal utility markets. They are all good markets for us, and we haven’t really felt any compulsion to move outside of them. These are big, well-designed networks that are being operated for the right reasons. They actually care about things like uptime, a reduction in outages, and customer satisfaction. There’s no reason right now for us to thin out our focus on the business problems that they possess.
TR: In what ways is the proliferation of last mile residential fiber impacting things?
MJ: We certainly have more inbound on the last mile. We have done the long haul. We have done the regional. We’ve done the metro and the middle mile. And we’ve now become involved in our first handful of projects for that access part of the network. This has historically not been a big part of what LightRiver does, but it’s a growing part of what we’re doing today. That access edge isn’t what it used to be. We’re starting to see gigabytes to a home, and 10+ to even a small business. So now we’re talking about actively managed bandwidth going into those sites, and that works for us both on the design side and on the automation side. With network surveillance, it gives us the opportunity to diagnose the network from the giant core to the tiny edge using the same set of tools. We’re not going to be going backward into some of the legacy PON technologies, but we will be going forward with some of the new ones. At FiberConnect there was a proof of concept where we brought six vendors together under netFLEX in a fiber-to-the-home configuration that earned Most Innovative PoC award recognition at the show.
TR: Has working on access projects made it necessary to add new capabilities to netFLEX’s technology?
MJ: We continue to grow the network element library and expand the kinds of services that we’re monitoring, managing, and controlling. Some of the devices that we’re now adding are types we didn’t have before.
TR: Given all the intelligence and automation you have built into netFLEX, where does the rise of AI fit into your roadmap?
MJ: We do have an AI policy here, and we have it in our products now. It’s not generative AI like ChatGPT. We don’t talk to it yet, but we do ask natural-language questions. We do query the database that way. We have over a million pages of technical documentation from the OEMs on products that exist today and existed in the past, and that’s all fed into our own secure database. We are running our AI models on that database that we own, so by definition, all of the data is good and none of it is made up. That is shortening search and test cycles and helping us with repetitive code generation. There are a number of things that we’re doing safely with AI assisted development that are giving us more bang for the buck here with our dev team, improving their approach to how they can solve problems and giving them more time to actually define the problem and less time trying to type it out.
TR: Are you using AI to help with early detection of network faults and such yet?
MJ: Not in production. There’s no reason to jump the shark on this stuff yet. There are so many ways that we’re incrementally going to both improve the customer experience and employee satisfaction by smartly integrating AI into parts of the business. We’re going to continue to do more of that than try to do something dramatic. We don’t need to.
TR: Offering netFLEX-as-a-service means you are now also a consumer of infrastructure in addition to being a builder of it. Has that changed your perspective or affected your relationships?
MJ: Yes, we have stood those up in large commercial data centers. So far it has not given us any surprises. But some of those folks that we build networks for are thrilled that we’re now their customer as well. I’m fine washing the other hand. They’re doing a good job for us, and we’re going to continue to do a good job for them.
TR: What trends are you seeing in how your customers build out their networks?
MJ: It differs by cloud, carrier, and utility in terms of pace and scale, but all of them are moving to packet. The larger operators are moving to flex-grid, and we are seeing rapid adoption of disaggregated networking. Even if that disaggregated networking is from a single vendor, they’re still starting to buy these components individually so that they can mix and match going forward. Recent supply chain headlines have kind of hurried that along a little bit. That is fantastic for us. But the biggest question is still where they are going to get the people to do this. A lot of the oxygen in the room is being absorbed by the billions of dollars that will be going to the access edge for rural broadband. But the problems that exist today with the workforce are not going to improve when those dollars start to flow.
TR: The talent deficit is a recurring conversation in the industry, of course. Is it getting any better?
MJ: It continues to get worse. But I think folks are expecting a step function when some of the federal dollars start to flow. I’m a little wait-and-see on that, but it is a lot of money requiring a lot of people to deploy and operate a lot of networks. There are many parts of the country where they have trouble retaining and acquiring talent, and I don’t think those areas are going to necessarily enjoy that activity.
TR: How do we solve this? Can we?
MJ: We’re solving it with automation. We’re solving it with software surveillance and recommendations on networks so that the engineers we have can do the most valuable tasks at hand and not those tasks that we can, in fact, fix without human intervention.
TR: How should we think about the new LightRiver with you at the helm?
MJ: It’s not too different from the old LightRiver. You can trust us to find an innovative way to solve your network problem. Those network problems might be wider than they used to be, or they might be deeper than they used to be, but we’re wider and deeper than we used to be, too. And we’re going to be there to help.
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: Fiber Networks · Industry Spotlight · SDN · Software