On Telecom Ramblings we tend to cover the more visible parts of the sector. But there are others that do a whole lot of the work behind the scenes in building out the infrastructure we depend so heavily on. With us today to talk about the realities deploying gear and building networks for operators is Michael Jonas, President Global Sales & Marketing at LightRiver Technologies. LightRiver generally flies under the radar, but recently kicked off 2016 with a newly redesigned website.
TR: How did LightRiver get started in this sector, and what kind of networks do you work on?
MJ: So we started the company in 1998 to deliver optical networks to utilities, the most robust optical networks in the country by far that were used to keep the grid in balance and keep our lights on. During those years, they all established carriers' carriers, the uTelcos, as you’ll recall. Through plenty of consolidation, those became some of the bandwidth providers of today. Interestingly, our utility customers, where we started, still represent 40% of our business, and their needs have only gone up. Our carrier business has grown over the years and has turned into very fertile ground. Most recently there are the data center and cloud players who are prolific in their growth and the number that seem to be in the market. Their networks look very much like a Tier-2 carrier, they just have a different mission.
TR: That was just as the telecom nuclear winter arrived, how did it affect the company?
MJ: I know, great timing right? Actually, for someone like LightRiver, the crash was in a perverse way actually quite good in that it brought some of what we had invested in the whole time to a real value play. Before the crash, some of the equipment manufacturers -- many of whose names are now in the dustbin of history -- they had everything: R&D, product engineering, product development, project management, deployment, FOC, NOC, first-line maintenance field resources. And they split their budget among all those activities. After the crash they had to grow up and focus on having a better product and value story than the other survivors, and they shed a lot of those non-core functions. And while they never really competed directly with LightRiver as they were focused on the Tier 1, it really elevated LightRiver's investment in those kinds of services for the Tier 2 carriers and utilities. After the crash, the opening was much wider.
Another byproduct of the crash was that as the manufacturers pulled back on their support resources, they significantly increased their focus on product and they actually all improved in that area. Better documentation, better, more reliable features, more robust roadmaps, better value per byte, and some became much more entrepreneurial than before. It made it easier for us to do what-if analysis, implementation, design, and interop simply because the documentation was better, because the software was more solid, because they paid more attention to what the market was asking for instead of what they could design. Looking back 15 years now, I think a lot of value was actually derived.
TR: How has the consolidation among your network operator customers changed things?
MJ: On the carrier side there's been a lot of consolidation, but very few networks have actually gone away and that has provided us with a couple different realities. One is that, because of all the M&A, the carriers' carrier market has had to focus a lot more on synergy savings, on EBITDA, and how to get more out of the network(s) they acquired or inherited. And they don't necessarily have the same corporate memory for those networks that they used to have. In many cases, they have relied more on us to provide links in that DNA chain. We have a multivendor lab in our Concord facility in California that has 4 or 5 generations of optical equipment still in it. Any piece of equipment our customers still have in their network we have in our lab.
TR: What kind of projects are you working on currently?
MJ: Right now, we have a national longhaul network we are building, multiple regional networks, and several metro packet-optical networks. Each one of those is probably using a different one of our vendor certifications. 95% of what we deliver are custom, turnkey DWDM carrier Ethernet or MPLS networks, built to meet a particular business goal. Our customer starts with a business challenge, and we figure out the best mix of technologies or vendors to solve that problem for them. Projects are primarily fiber-based using our Factory Built Network process we are able to deliver these networks anywhere in the US or overseas as well. For those folks who are running a network that they are able to predict and rely on.
TR: You mentioned Factory Built Networks, what is that?
MJ: It's an enhanced staging. A lot of folks will claim staging, but that's step 1 of 6 for us. As these technologies roll out, there are fewer and fewer of our customers that are versed in every new technology and every new vendor and every new product. Rather than deliver a kitted site to the field and fly out to support the implementation, we build everything in our network factory before it goes anywhere. It's more than rack and stack. We'll label it, barcode it, bring power, bring fiber, simulate the fiber distances, burn it in, test it, and provision it all in our factory. Before we tear it down, many of our customers will come out and run through a factory acceptance test in place so they'll know how to work the features they are buying as opposed to 100 pages of manual. We then build custom coffins for each of the major equipment types, and ship it pre-built. It's a way to dramatically reduce the delivery time for a new network and take a lot of the surprises and cost out.
TR: What is the biggest shift in what network operators ask you to do for them that you’ve seen in recent years?
MJ: Really it's been the migration from TDM. There was an enormous TDM/SONET plant out there, and there was a lot of comfort among a lot of people who knew how that network was going to run, how they were going to support it, and what it was going to look like. It has a deterministic nature - once you nail it up it's going to produce forever. But the ability to prioritize Ethernet traffic and make it deterministic has really been a game changer in the last few years. Almost all the new services are going to be carrier Ethernet for us, with thought to prioritization and operations and maintenance. That's improving speed, cost per byte, and flexibility. A lot of our customers were not used to that sort of dynamism, and our ability to fill in the gaps for them and empower them and backstop them has helped accelerate the growth in that area for us.
TR: How quickly have you seen different types of networks shift from TDM/SONET to Ethernet/MPLS?
MJ: For the utilities, it's been a 10 year process. In our data center/cloud business and for emerging carriers it never happened because they never saw a TDM circuit. With our carriers, the bigger the carrier, the more readily they accepted it. They are still integrating their networks, tuning their internal financial models, but are being pushed by the data center and the cloud guys to do more and do it quickly. Meanwhile the smaller carriers and independents are moving more cautiously into the new world, but some are really moving forward now.
TR: How do their priorities differ?
MJ: For the utilities, 100% of their job is risk avoidance. For the carriers, they have to make money, they have to get a return on any network they build, and they tend to be ahead of the utilities on technologies. Then we have the data center and cloud folks, who are really building world-class networks as a competitive advantage and not necessarily for resale. So we're getting to see multiple perspectives on how these technologies can be applied. The competitive environment is changing for the better.
TR: Have you seen much success with the cable MSO sector?
MJ: We have a few clients in that space, but there's an area where consolidation has really ended up with half a dozen Tier 1s, and then a lot of other folks that don't really need the big iron as much. We've had some opportunistic success, but it hasn't been a focused, proactive sales sector for us. I expect them to become a bigger part of our business in 2016 and beyond because of the way they are evolving. The dialogue we are having with them very much the same discussions as with some of our carriers' carrier customers a few years ago. I don't know how to explain the difference in timing on that, but the discussions are almost the same. They do have a rapidly evolving market. They moved from analog to digital and threw tons of bandwidth at it. Now they're getting into triple play and quad play and paying more attention to quality of service and customer experience, and that has diversified the kind of traffic they have to carry. With that, they're asking more questions of us than they used to ask.
TR: What is evolving most rapidly, and where do you expect the biggest technology shifts this year?
MJ: 100G and above everywhere for everyone is a reality that has occurred much faster than even we anticipated, and we spend all our time in the space. More powerful boxes taking less power, and providing more value per byte, those are the hardware realities. But in 2016 what we really believe is that software automation is going to come of age. It's going to come more quickly than anticipated and it's going to be more pervasive than predicted. The promise of software-based network automation becoming a reality in a compressed timeframe means that our customers are going to want support in customizing some of their SDN tools, and we're going to double down on our software technical team. We're working on that right now, and that's going to be the most interesting part of 2016. We've all talked about SDN and NFV, and rightly so. But the rubber's really going to start hitting the road in a wider cross section of the network community. At LightRiver we need a set of software tools that are a bit more thoughtful and not as ad-hoc as you get in the early stages of a new technology. What we've seen today leads me to say that we'll be able to turn some of the soon-to-be-released tools into productive use for our customers. We are starting to see the business component work itself out.
TR: Are vendors embracing a multi-vendor SDN future or working within their own silos?
MJ: It depends on the vendor, but we're going to see tools that are multi-domain for a single vendor and a variety that are built for multiple domains for multiple vendors, but just haven't developed full featured products yet at a lower, operational, level. At the higher level, many of these tools are being re-architected to take advantage of a much more comprehensive multi-domain role in the future. The development itself is additive and complementary. In the past we have seen a dozen major manufacturers doing a dozen one-offs. Now they're working on, well I wouldn't call it a common set of standards, but a much more common target. The development is benefiting from that new business reality.
TR: What role do you think standards should play for SDN right now?
MJ: It's interesting, because we do both hardware and software. It's clear to us that the hardware side of the industry has benefited from standardization, which can be pretty granular. Common guidelines and standards are going to be essential for software too, but I'm not sure they need to be quite as strict as what was beneficial to the hardware folks. It's difficult to quantify, but the software development process is a lot more idiosyncratic, creative, and tied to a particular team. I think as a result, giving folks the guidelines to live within but a wider path to get there could probably be more advantageous and allow for more innovation.
TR: What has been the most unexpected change in your business?
MJ: We originally envisioned our investment in the factory-built network process as a way to speed delivery for our customers here in the US. As we've gotten more into the cloud and data center space, an interesting unexpected result is that many of those folks may be headquartered in the U.S., but deliver content everywhere. They are using our factory to build and test and troubleshoot and train their field resources for anywhere in the world. So we might build a network for San Francisco. And then we'll build the same network right next to it for Singapore or London or Frankfurt or Amsterdam. There's almost no difference in how we do that, but it's opened up a world we hadn't expected. Where we had been building regional and metro networks in the US, we're now tying those together across the US and sometimes across a couple of oceans. Coherent optics and Ethernet have been great equalizers, taking regional variability out of these networks.
TR: Thank you for talking with Telecom Ramblings!