SPONSORED FEATURE: The next generation internet is all about experience. Hyperscalers like Google, Azure and Facebook epitomize the importance of both technological and customer experience. Their ability to automate and rapidly scale technology enables the smooth, latency-free movement of data and workloads, which in turn, underpins the services and applications enterprises and consumers rely on.
As hyperscalers have stretched their way across the globe, customers expect to be able to access resources whether they are in the US, the UK and Europe, or Australia and Indonesia. And naturally other service providers, as well as global enterprises operating their own infrastructure, look to them as an example as they look to deliver a similar experience to their own users and customers.
This is one reason why cloud deployments dominate spending on compute and storage infrastructure. This will hit $24.6bn in the second quarter of 2023, according to IDC, and is shifting to ever more complex workloads, including AI. And AI will lead to a tripling in hyperscale datacenter capacity over the next six years, according to Synergy Research Group.
But even as they keep up a remorseless pace of technological evolution, the mightiest of hyperscalers can still face fundamental challenges in expanding their global footprint and infrastructure.
This is partly a question of physics when it comes to latency and data round trips. This is one of the driving forces behind the increasingly hierarchical architectures such operators are adopting, with regional architectures and edge approaches moving data and compute closer to users.
But it is also a question of negotiating very different landscapes when it comes to regulation, compliance, data sovereignty, and, on some level, politics. And this is where hyperscalers – and potentially, everyone else – need to call on someone else’s hard-won experience.
As Nokia’s Webscale CTO and head of architecture, Rodney Dellinger, explains, the standard MO for hyperscalers expanding their network is to acquire dark fiber from a provider and build out their equipment on top of it.
But in many parts of the world, he says, “You run into these challenges where you’re unable to operate your network in the manner that you’re used to operating in.”
That could be because providers looking to acquire fiber in, for example, India are required to have a telecom operator’s license. Even an Apple, Google, or Azure may be unable or simply unwilling to assume such an obligation. In Europe, and other countries, regulations around data sovereignty might be at odds with the obligations US-based operators face from their home government, making them think twice about how they build out infrastructure.
In both instances, webscale operators may choose or be obliged to work with a local cloud service or colocation provider. In Europe, for example, they will look to these partners to help them construct “trusted clouds”, which can underpin the service levels expected by hyperscalers while meeting local regulatory requirements.
This is where Nokia plays a vital role, says Dellinger. The company’s technology portfolio spans access networks, submarine networks, IP networking, and optical networking.
Nokia’s networking software technology stack and silicon capabilities have been developed over decades. And it has worked hand in glove with the hyperscalers to direct its technology. “Whether it’s helping to work with the open config standard, whether it’s developing things like gRPC, or Netconf protocol alongside the hyperscalers, that’s taken years to do,” says Dellinger. “We’re getting direct input from these customers, the features they have in them are not an accident.”
And that same aggregation of networking technology, bolstered by its acquisition of Alcatel-Lucent in 2015 for example, has given it deep expertise in these countries. It is invariably a supplier to the major regional or local providers, as well as to the hyperscalers.
As Dellinger explains, when a hyperscaler lands in a country, “You’re connecting to something, this optical gear is not connecting to just other optical gear. It’s either a local data center in the region, or we are partnering with the local service providers.”
Building networks over years and years
But if the benefits of an intimate understanding of technology or regional business links might be clear, others might be less obvious.
“You need to make certain that what you do has a resiliency that will allow you to serve your customers. These hyperscalers are CSPs now, they’re just delivering a cloud service as opposed to telecommunication service.”
This means they have SLAs they have to fulfill, says Dellinger. “What they have to think about is “how do I do this in the most sustainable, reliable and economic way.”
“And the only way to do that is to work with a supplier that has some scale and with a CSP that also has some scale and a proven track record.” Invariably, he continues, Nokia and those local service providers which can fill this gap will have been doing business together for decades.
That local ecosystem is more than just business relationships with key players, he points out. “We have labs, we have support teams, we have logistics in these places to help come along. And that goes way beyond … the bits and bytes of technology.”
The importance of relationships
This includes 365-day support globally, including local language support, which adds up to a longer-term view of what it takes to construct and maintain both networks and customer relationships irrespective of their location. This all becomes increasingly important given the current drive towards longer, more sustainable life cycles for networking and compute, even within the fastest moving hyperscalers.
“If you read any 10K report from the hyperscalers, you will see in every single one they will make a note that they are increasing the depreciation cycle of their networking gear and their compute gear. And so that means he’s going to be in the ground for five years, six years, whatever the case may be.”
And while new technology comes at a breakneck pace, says Dellinger, “If you’re making an investment in your network that’s going to last for five, ten years, you need to do it with a proven company.”
Once the local difficulties are addressed, the hyperscaler or service provider can focus on replicating its infrastructure. From Nokia’s point of view, says Dellinger, “What we have attempted to do, whether it be switching, routing, or the optical network, we tried to create network blueprints.”
This means that from the hyperscaler’s standpoint, “Whether you’re in a GCC country, whether you’re in India, or whether you’re in Tokyo, it looks the same.”
It also makes for simpler forecasting and ordering. And forecasting raises the importance of understanding likely project timescales for different locations.
“A red flag to me is when I see timelines that are completely unrealistic,” says Dellinger. “When you’ve done business in Mumbai and Delhi, you start to gain a respect for the pace of business there, you start to gain a respect for the pace of acquiring whenever things are available.”
He cites one installation where a datacenter was ready to move into, except that the air conditioning equipment was yet to arrive, which meant it was “unsafe for the technicians to work.” This is precisely the sort of issue that Nokia and the local or regional partners it works with are able to anticipate and plan for when designing and building out network infrastructure. Which means the hyperscalers don’t have to.
That allows the hyperscalers to focus on what they excel at. And that is focusing on automation and ensuring the service – and the same experience – wherever their customers are. And whatever the local connectivity environment looks like.
After all, as Dellinger says, “The hyperscalers don’t do exceptions. They don’t do one offs. They want everything to be the same.”
Sponsored by Nokia.
Sign up to our Newsletter
Featuring highlights, analysis, and stories from the week directly from us to your inbox with nothing in between.