Ebbs and flows: Subsea versus edge
In the midst of the two-day event that was Capacity Asia 2020, a topic was raised that led to a much bigger question in need of answering.
Courtesy of panel moderator Robert Schult, research director at TeleGeography, the question was, now that the cloud is closer to the end user thanks to edge computing, will this diminish the requirements for international capacity because as Schult’s put it “we’re not tromboning things back and forth”.
Well, the general consensus among the group of panellists was that this is not the case at all, in fact, the two are connected.
Tech entrepreneur, Bevan Slattery, founder of NEXTDC, SUB.CO, Cloudscene, Superloop, Megaport, and co-founder PIPE Networks, HyperOne, said that content and compute will continue to be pushed closer to the edge and that “5G will certainly enable certain things to happen which will lead to an increase in consumption” but the biggest area of growth he foresees is “AI, machine learning, interconnecting massive data lakes and IoT. Those will be a massive adoption and consumer of subsea capacity.”
“I also think alternative routes will also be a big part of that for various reasons, geopolitically, for diversity etc. So, I think the death of subsea due to edge is greatly exaggerated,” he added.
Newcomer to the subsea space is MegaFon, whose Alexander Vlasov, director of carrier services, echoed these sentiments saying that the traffic is growing and “regardless of the content moving to the edge, you still need to carry this traffic between the data centres even though its coming closer to the customer”.
Capacity reached out to Matt Walker, chief analyst at MTN Consulting about this idea of edge leading to the end of subsea for this thoughts and unsurprisingly he says, it’s pretty impossible to answer because “too many things are happening simultaneously” but that “there is always an ebb and flow of undersea investment, driven nowadays more by the webscale/Internet players than telcos”.
“Edge computing investments are rising to support new business models which are generating new traffic for the network – and some of these new services most definitely require international connectivity,” says Walker.
Most interestingly, he reminds us that these types of questions were asked before, with the advent of CDNs. Similarly, it was believed that they might lead to a decline in demand for international bandwidth, as content would always be cached locally.
“They may have had that effect in part, but they also improved network performance, broadened the range of content types, which were worth consuming online, and increased time online among users. The net effect was to keep demand kicking ever upwards for submarine capacity,” explains Walker.
Speaking to Remi Galasso, founder and chair of Datagrid New Zealand, and founder and CEO of Hawaiki Cable, he points to the exponential growth in data as the reasons why we are nowhere near done with subsea, in fact, much more is on its way.
"The volumes of data being created worldwide today are increasing at a rate few would have predicted even just a few years ago,” says Galasso. “But what we’re seeing now is really just the beginning, with the AI revolution one of several new data-led trends looming larger on the horizon.”
This sentiment is echoed by Martin Garner, chief operations officer at market research company, CCS Insight, who points to new technologies like IoT adding even more data into networks, that rely on subsea cables to get around the world.
"We do not expect to see the volume of traffic on subsea cables dropping," said Garner. "The trend of data volumes growing exponentially continues unabated. In fact IoT and edge computing are adding thousands of new data sources to this all the time, as IoT continues to provide data about areas we’ve never been able to measure before, such as air quality."
"So, although there will be a lot of processing at the edge, this is to help make the volume of new data transmitted to and from the data centres more manageable. It does not mean we’ll stop - or slow down - with sending data around the world to get to and from the cloud.”
Though Galasso acknowledges that low latency applications won't use long-haul subsea cables for obvious reasons, the interconnection of major international points of presence will require a lot more bandwidth for the foreseeable future.
As a result of this and the fact that current transmission technology is approaching the ‘Shannon Limit’ (maximum rate of error-free data that can be transferred), “the market will increasingly demand more subsea cables with more fibre pairs”, he says.
In February 2021, Google announced the Dunant subsea cable connecting the US to the French Atlantic became ready for service and with it the launch of 12-fibre pair space-division multiplexing (SDM) system delivering 250Tbps.
Though the market was all abuzz about this next-gen system, its launch supports Galasso’s view that “this kind of system will become the norm on every major route in the coming years”.
Ocean Networks president and CEO Scott J. Schwertfager takes a similar view, believing that if we continue to create subsea bandwidth, it will always be used.
“Whenever we put in a new cable system we say, they're never going to use all that capacity and five years later they're already doing an upgrade and 10 years later that technology is now obsolete and they need to put it in the new system,” he says.
“With newer technology systems upgrades are much easier and cheaper, so we expect that to go further. But even with edge that data is still going to have to be transmitted to local networks and so we're not really seeing that as a concern. If you create the capacity, they find a way to use it.”
Looking at how this relationship between edge and subsea is likely to evolve in the near future, Telstra’s Paul Abfalter, head of North Asia & global wholesale, shared his predictions saying that while he doesn’t believe the edge will mean any slowdown for subsea, “you might see a combination of edge and new technologies pushing SLTE (submarine line terminal equipment) even more into city points of presence and into data centres”.
“Perhaps less CLS’ (cable land stations) at the shore, this is definitely something we see more of within a five-year time frame. Edge will just increase use cases for subsea rather than decrease,” he adds.
So, while changes are certainly afoot for these two critical pieces of infrastructure, neither is showing any signs of slowing down at most it seems an evolution of sorts, is on the cards. One thing is clear however, “edge computing alone is not likely to suddenly curtail overall the need for subsea connectivity,” Walker says.