Published on TelecomRamblings
TR: It’s been a year now since the merger with Cross River. How has the integration gone?
RL: Honestly, it was really easy because the core values and focus of the businesses were similar in so many ways. Prior to the merger, both companies had built nearly 1,000 route miles of completely complementary network within each geography and enjoyed a long history of working closely together to develop regional deals across our combined footprint. This business is not people-intensive but rather it’s asset-intensive, so you can do it with a smaller staff. The two staffs merged to about 30 people, and it went seamlessly. We had some time (four months) leading up to the merger while waiting for regulatory approvals. But once we had that we hit the ground running, it was as if we had been working together in the same business for years. The synergies have really worked out well, we have tons of expertise in our combined staff, and we have had a really nice annual growth rate.
TR: Has the combination changed the way you do business?
RL: ZenFi Network’s fiber architecture and network architecture was really focused on solving for mobile densification, while Cross River’s was more focused on carrier wholesale and enterprise. Our thesis for the combination was to take the mobile story to the robust footprint they had built in New Jersey, and to take their enterprise and telecom services story and overlay it on our dense New York footprint. We do position ourselves as a Dark Fiber 2.0 provider. Legacy networks were built to solve a sparse networking problem. The mobile network now has an ultra-densification problem in urban environments, and they need a new kind of fiber network. The network we have built to support that is backwards compatible when it comes to supporting the sparse connectivity needs of carrier wholesale and enterprise. So the sales team that we had developed on the Cross River side simply gained some new inventory assets in New York.
TR: Should we be thinking of ZenFi Network’s business model differently than the traditional metro fiber model?
RL: The analogy I like to make is that ZenFi Networks is a horizontal tower business. The deals that we do with mobile operators are long term, larger deals. They’re unbreakable, they have escalators, and they have all the same type of terms that the tower guys have. It’s just that our towers are 10-kilometer-long fiberoptic cables with vertical extensions coming out of it. Each light pole we connect is like a RAD center on the tower, and our network edge colo is the ground space under the tower.
TR: When we last talked you were engaged in a huge buildout for LinkNYC. How has that been going?
RL: LinkNYC was a foundational deal for us. For the project, we built a dense, pervasive citywide network with a C-RAN architecture that is just like you would see for 5G. It’s a combination of fronthaul access networks, network edge colocation facilities and a backhaul network built in all five boroughs. Today there are thousands of kiosks active and connected to that network and our network is fully prepared to deploy thousands more across the City.
TR: What parts of New York City were the most difficult to build out to for that project?
RL: The historically difficult areas to build network are Queens and Brooklyn. Those are completely under Verizon’s control, so all the make-ready has to be done through the Verizon processing system to prepare underground ducts and aerial pole lines. Staten Island is also a Verizon territory and has similar challenges, but it’s mostly aerial which makes the process a little faster. Within Manhattan and the Bronx, we’re able to self-perform and do our own underground make-ready. We can build more network there quickly than we can in the other boroughs because we don’t have a third-party intermediary that we have to work through. But if you look at our map, you see we have pretty nice footprint throughout all five boroughs, with Manhattan being what we would consider ultra-dense.
TR: What does your C-RAN architecture look like today?
RL: Think of it like this: you rent a light pole to place a radio and an antenna, and you run a pair of “fronthaul” fibers from there back to a network edge colocation. Imagine a cell tower on its side, ground space colocation, fronthaul fiber up the tower and a radio head and antenna on the tower. Again, this is what we consider our horizontal tower infrastructure model. You do RF processing on the pole and you do the baseband processing in the colo. The traffic from network edge colocation facilities is then then backhauled to the mobile operator’s command and control at their Mobile Switching Centers. That was the architecture that we went to market with from day one, and that infrastructure is technology agnostic. It can run radio heads, DAS remotes, small cells, or Wi-Fi access points. One of the key points, though, is that as we migrate to 5G, latency is a concern. You have to keep every link from pole to network edge colocation less than 10 kilometers to stay within 5G’s fronthaul latency tolerances.
TR: Are you seeing any demand from 5G itself yet??
RL: The nodes going up on poles today are predominantly 4G LTE. There is some 5G testing, but all the production stuff is either DAS remotes or 4G LTE radio heads and small cells with the traditional for low- and mid-band frequencies. We are in trials with 5G antennas and 5G radios, but they’re not done in any material production environments at this point. That’s the next shoe to drop. We expect that we’ll be doing a lot of 5G, and in the high bands, 5G requires more nodes closer together. We expect to be deploying tens of thousands of nodes on our C-RAN infrastructure in the city of New York over the next 5 years to support the evolution toward 5G.
TR: Are you also seeing demand for your C-RAN architecture in New Jersey?
RL: We are seeing the same thing that happened in New York a couple of years ago now happening in New Jersey. The most costly parts of these builds are getting a node on a pole and then getting both power and fiber to it. For that infrastructure we are now winning deals for 4G nodes in the denser markets of New Jersey, which will move on to 5G nodes down the road.
TR: Are any adjacent or other geographies in your plans yet?
RL: I do expect that we’ll start moving out of Queens into Long Island and from the Bronx up into Westchester County. But our real focus is on densifying the areas where we already have substantial fronthaul or backhaul networks. If you look at our footprint in Northern New Jersey, we have a lot of territory covered, but there’s a lot of densification yet to happen. Carrier demand is pretty significant, so we have a lot of opportunity before we move on to other non-adjacent markets.
TR: Tens of thousands of nodes sounds like a lot of capex, do you anticipate raising outside funding?
RL: We have built a pretty pervasive, robust regional network with our own funding and we remain self-funded with a robust balance sheet. And because we’re a good cash-generating business, we have a lot of borrowing capacity that’s untapped. To date, we haven’t had the need to go out to the market for private equity or infrastructure funding. We’ve been doing really well just building and reinvesting in the business.
TR: What other than fiber densification have you been working on?
RL: We have really started building momentum on constructing our network edge colocation spaces around the region. In order to drive the evolving 5G ecosystem closer to the user at home and at work, a new layer of edge colocation facilities need to be built. Consequently, we think there’s a big opportunity in building and connecting these neutral network edge colocation facilities in neighborhoods rather than the legacy commodity business of interconnecting centralized data centers. The network edge colos today primarily support mobile operator baseband processing, but ultimately many of the locations will evolve to become edge data center facilities. We are spending some serious time finding locations, soliciting mobile operators as tenants, and marketing the excess capacity to provide neutral colocation services. But that’s also part of the C-RAN architectural story. I think that there’s a need now because of the latency requirement, or the short optical length that we have to be at the center of.
TR: How are you approaching the construction of those network edge colocations? Are you leveraging any of the modular technologies out there designed for colo at the tower, etc?
RL: Ultimately, we think of ourselves like a tower company. We are in the underlying infrastructure business and we want long term master lease agreements at properties within which we build our own facilities, and we’re doing that all without a data center partner. We know how to build an infrastructure with power and cooled shell and to create subdivided spaces to provide carriers with secure, environmentally-controlled co-locations. We really don’t need partners for that, so we’re doing it with our own capital and we’re doing it with our own knowledge. It’s a business that we see as core and fundamental to ZenFi Networks.
We did look at modular partnerships, but the areas that we are looking at are fairly dense with no ground space and we haven’t found an easy path to using a modular. We feel that it’s to our advantage to either buy buildings or lease space of the right character and build it ourselves. It’s really the power and cooling infrastructure that are the most costly parts, and we know how to do that.
TR: Do you foresee any additional consolidation opportunities like Cross River out there in the NYC metro area?
RL: There are only a couple of other franchises out there that are independent in our region. The challenge we have is that everybody’s network was built to solve a sparse networking problem. To support mobile from here you need ultra-dense networks which requires the unique fronthaul infrastructure that those companies don’t have. So an acquisition of another business in the region would not necessarily add anything within our geography that we can’t build ourselves more cost-effectively.
TR: Thank you for talking with Telecom Ramblings!