An increasing number of submarine cable projects are forgoing the humble cable landing station and connecting directly into data centres. Low latency, easy access to customers and cost-saving are among the benefits of doing so, but bringing the subsea cable further inland also has its own share of risks.
During the inaugural Subsea Connect Americas conference in Florida, a panel – consisting of Aqua Comms CEO Nigel Bayliff, Xsite Modular CEO Amy Marks, NJFX CEO Gil Santaliz, Digital Realty VP of commercial management and business development Don Schuett, and managing director and senior analyst at RBC Capital Markets Jon Atkin – discussed this very topic and what is driving this change.
According to Schuett, money is the biggest factor when it comes to choosing whether or not to forgo a cable landing station. “I think with the cost pressures of building new [cable] systems, nobody wants to go out and build a brand new cable landing station as they did 10-15 years ago. It’s a very costly endeavour.”
But Marks, whose Xsite Modular is the leading firm for pre-fabricated technology embedded buildings and the design-builder of modular cable landing stations, is quick to disagree.
“If you look at the cost of the cable station compared to the cost of the cable itself, it is like a decimal point. I can’t understand why people spend all this money on these cable systems and then put it in someone’s supermarket basement, instead of actually providing a cable landing station to protect the data of people on this planet. It is a pimple on the butt of the cost of these systems.”
NJFX’s Santaliz clarifies Schuett’s previous statement, saying: “The building is not the expensive part, it’s the utilities to get there, it’s the backhaul to get there, it’s the site selection and it’s all the earth you have to move to make it work.”
Capex and opex
Schuett adds: “And why would you need to go into a capex model when there’s existing cable landing stations, there’s existing data centres, that can essentially serve the same purpose?”
He continues: “Operators and owners of these new systems are not out to own assets, they’re out to build a network for their own use. They don’t want those assets on their books. They want to operate on an opex model.”
As for Bayliff, it’s all about location and diversity, he says. “I think all these things depend on where you go,” he adds.Using his own cable as an example he explains: “In Ireland (Killala) where our cable (CeltixConnect) lands, there is nothing there. There is a small town and an old aluminium bauxite smelting plant, which is why we have heavy-duty power from both directions on the grid. But it is four and half hours from Dublin and the nearest data centre is a good 200km away, as well as the fibre backhaul.”
Overall, he says: “It is about closeness to the beach. You don’t want to be taking your high voltage power, which in the case of the Atlantic about 1,200 volts DC wrapped around a fibre optic cable, too far inland because if somebody comes along and starts digging it’s not the same as your regular 110 volts.”
As the conversation progressed, a term that came up repeatedly was “hardened infrastructure”. Speaking about cables being extended inland to connect directly into data centres and being more vulnerable to risks, Atkin explains how these concerns form part of the decision making process at the investor stage.
“Physical security and reliance on hardened sites is well placed,” he says. He believes the newest cohort of data centres are resilient and up to the challenge. “During super storm Sandy there was not a single data centre in New Jersey that had a customer outage. All the outages that one hears about were the older ones where you had diesel fuel pumps in the basement that failed. All of the newer ones were fine.”
Market drivers
Speaking to Capacity, Larry Schwartz, chairman and CEO of Seaborn Networks, explains that he thinks the shift in the market is a positive thing. “Where there is a pre-established, well-populated data centre that can serve the role of landing station plus POP, then it makes sense.”
As for what’s driving it, he says there are two factors. One, “we can run over 10,000 kilometres without landing the cable in between, and can go from termination point to termination point. That can make for a more reliable system with fewer active elements.”
The second factor is “to push costs down and make these more affordable projects. So the combination of technology improvements and lower cost serves the industry very well,”
Conversely, Artur Mendes, COO of Angola Cables, thinks that the customer is the one driving this change. “In our case our cable in Angola is in a data centre. We are also building a data centre in Fortaleza and use a data centre as a cable landing station in Boca Raton and that’s because essentially our customers are directly in the data centre,” he says.
“If you are finalising at a data centre, you are at a cross-connect distance from the other guy. So it’s quite simple to just order a cross-connect and interconnect with the other customer on a different route. When you are at a landing station it is not as simple as that because most customers don’t have access to the landing station, so you need to pay for backhaul to connect. It’s a lot more complicated to do.”
But Paul Scott, president of C&W Networks, says: “I think in developed countries, there’s a greater density of data centres. But let’s face it, in some of the emerging markets there’s not an abundance of ready-to-go data centres sitting in an appropriate location suitable to land a cable in.”
He adds: “So I don’t think it’s either-or. It’s what does the landscape look like at a given country and what can I avail myself of if there’s a hardened world-class data centre that’s not too far from the beach.”
“I think combinations of that are the reality and practicality there’s not a bunch of data centres to pick and choose from in emerging markets in the areas we play in. We have a bunch of data centres of our own, which are multi-use facilities. It just all depends.”
Marks agrees. “None of us are representing the under-developed markets that actually need cable landing stations, because there isn’t a data centre there. There’s plenty of the world that isn’t in a place where they can choose which data centre to go to, because they don’t have one, and we have to provide cable landing stations for them in those locations, just for them to be able to have connectivity. So as the world changes, yes we might be moving towards this new developed model, but there’s so much of the world that’s not there yet.”
Neutral waters
It was Santaliz’s last argument that poses the most innovative solution to reviving the cable landing station, one that requires a little collaboration.
“I think the initial conversation about what’s changing in the model is not the physical attributes of the cable landing station. What’s changing is: why can’t we have a cable landing station that’s neutral? Why can’t one have two subsea cables and let each operator have independence of each other and operate in a neutral fashion?” he asks.
“And why can’t the real estate owner stay out of the way? Help them out in terms of power and cooling, but just stay out of the way. Don’t get in their business, don’t compete with them and don’t get into the business of operating a cable or connecting to other sites.”
What was most telling was Santaliz’s final point, perhaps revealing a bit of the strained relationship between the cable operators and the owners of the landing stations themselves.
“Neutral to me means you don’t compete with the guys you serve. So I don’t compete with my customers. Therefore how can I help you? I don’t have a hidden agenda, in terms of when I connect you to someone or you to someone else, I’m thinking about how this is going to hurt my other business.”
