Scale and Its Impact on the Supply Chain
Top–of–Mind for Digital Infrastructure Leaders
This is the third in a series of five blog posts reflecting the top-of-mind issues discussed during the most recent Infrastructure Masons Advisory Council meeting, held in London in November. The Advisory Council is made up of end users and partners from across the digital infrastructure ecosystem.
Demand for digital infrastructure services is being fed by streams of data from a wide variety of emerging technologies. The Internet of Things (IoT), video everywhere, AI deployed across industries, individuals and organizations continuously connected. All of these things are changing the business of the data center, the place where almost all of that data lives.
That sounds like great news for data centers. And it is – until you need to scale at the speed of demand. At the most recent meeting of the Infrastructure Masons Advisory Council, members frequently brought up concerns about the impact of hyperscale on the supply chain.
Today’s conversations about scaling aren’t merely whether to scale up or scale out, adding computing resources in an existing location. The need to scale also takes businesses to new locations around the world. That, in itself, can cause supply chain issues.
“Scale is top of mind, not only in the physical infrastructure, building out at an unprecedented pace,” one ender user said. “It’s also scaling the organization in more distributed areas. It doesn’t work anymore to have your design teams based at headquarters trying to design all over the world and manage projects all over the world. That takes us to places where we don’t have an established supply chain. As we scale, we will be building in places where we haven’t before. And establishing the full ecosystem to support us.”
“As we scale, we will be building in places where we haven’t before. And establishing the full ecosystem to support us.” – Click to tweet
Another end user talked about trying to simplify as his organization expands globally. “We’ve spent a lot of time coming up with a reference architecture, which we can now stamp out in different places around the world. We’re still U.S.-centric when it comes to infrastructure. But we are rapidly trying to figure out a way to replicate that around the globe. So scale is a big challenge for us.”
Is this growth the new normal?
Emerging technologies are generating large volumes of data in ways that make it very difficult, if not impossible, to forecast demand. It’s difficult for digital infrastructure leaders to tell whether today’s phenomenal growth is primarily due to the introduction of so many new applications, and will thus even out, or whether the type of growth in data we are seeing today will continue at this pace.
“What I’m trying to rationalize is whether we’ve entered an era of a new growth slope,” one end user explained. “If you plot our infrastructure growth on a log scale, it looks pretty much like a 20% slope … a straight line with a 20% slope, which is more or less aligned with the top line revenue growth of the company, so it’s rational. So are we at the beginning of a new growth slope, or is this uptick we’re seeing now just a bubble and we’ll go back to the 20% growth?”
“What I’m trying to rationalize is whether we’ve entered an era of a new growth slope. Or is this just a bubble and we’ll go back to the 20% growth?” – Click to tweet
Take virtual reality (VR) as an example of growth drivers. VR content may need up to 20 times more storage space than today’s high-definition video. Storing VR data isn’t as simple as storing lots of HD, UHD (4K) or higher resolution video images. Storage includes full 360-degree views for each video viewpoint and requires the video to be stored on storage that is highly responsive and able to shuttle the video smoothly in any direction. Will VR innovation change that – will the next generation of our new technologies adjust how and how much data is stored? Or will VR always demand as much scale as it does today? The same questions apply to other emerging technologies.
“I don’t know whether that slope, that curve, has changed permanently or if this is just an ingestion of a lot of capacity,” one Advisory Council member said. “And on top of that we have all of the machine learning and AI, a very new business sector. Every business and industry seems to be trying to leverage that right now. So how do you model their demand for growth? Is this just huge right now because it’s so early on or is this truly a durable type of demand profile? So these are questions that I don’t necessarily have the answers to. Those are the things that certainly keep me worried.”
Business is constantly changing
Regardless of what’s coming in the future, Advisory Council members all suggested that the need to scale is creating challenges now.
“As business goes up, the scale of the infrastructure becomes hyperscale. Keeping ahead of that is difficult,” one partner said. “We have to make sure we’ve got capacity in the right place at the right time. Also the business is constantly changing. There are always new things that are rolling on our platforms – new features, acquisitions, partnerships – that just increase the volume of what’s happening on the platform.”
“We have to make sure we’ve got data center capacity in the right place at the right time.” – Click to tweet
Still – slow is the new down
The relentless pressure to deliver more data, more quickly, in more places, even as models evolve, has put pressure on data center availability. “Resilience has been quite a challenge to us,” one end user said. “How to migrate while growing – it’s like changing the engines on the plane while flying.”
He continued, “We’re in between the physical dependency and the logical dependency right now. So we’ve got our zone and region concepts. So we’ve got three zones in a region and multiple regions and we have the ability to fail over between those. But we’re also in migration. We have a lot of capacity that’s moving from one to another as we get to new regions. And so resiliency and expectations on data centers is much higher right now. We can’t have them go down.”
Pinching the supply chain
There was a consensus among Advisory Council members that pressures to scale are putting strain on an already overburdened supply chain. “When you look at the need to scale, you realize there are starting to be some real pinch points in the supply chain,” one partner explained. “Solutions can include less customized engineering from the supply chain and more standardization down to the component level. But that is incumbent on all of us to not over-engineer.”
“Solutions can include less customized engineering from the supply chain and more standardization down to the component level. But that is incumbent on all of us to not over-engineer.” – Click to tweet
The same partner also cited problems with relying on growth forecasts. “One thing that is consistent is that demand is not only growing, but that also every forecast is wrong. Forecasts indicate bigger and bigger numbers and they are coming faster. So the supply chain issue is troublesome.”
Another partner suggested it may be time to step back and rethink digital infrastructure systems. “We’ve got to ask, ‘How do we look at all this a little differently?’ Things are moving so quickly. How can we be more nimble?”
Suppliers putting pressure on themselves
Others agreed about the need to slow down and rethink. “The whole ecosystem is stretched to the limits and it’s starting to break in certain areas,” one end user said. “The equipment manufacturers are reluctant to put in new capacity because those are fixed assets on their books. But it encumbers them for many years or even decades.”
Equipment manufacturers also want to know whether today’s demand is the new norm or just a bubble. “When you have fixed capacity and what seems like ever increasing demand, you see what happens – lead times, pricing, everything else [gets constrained],” one end user said.
Everything else includes essential equipment. “Supply chain on generators is insane,” one partner said. “We’re starting to see 40-44 week lead times. Even as a small company, we’re actually starting to inventory now. And that’s tough.”
“Supply chain on generators is insane. We’re starting to see 40-44 week lead times.” – Click to tweet
One partner, a network infrastructure provider, said he tries hard to not become the bottleneck for engineers building out data centers. “It’s getting down to the point where raw materials are becoming constrained. We provide the infrastructure that you guys [end users] can’t forecast. We’re feeling the pain that you’re feeling in terms of long lead times. You’re having trouble forecasting, and it’s impossible for us to forecast what we’re going to need for you.”
Rethinking the system might include finding ways to become less customized, ways to explore standard designs. One electrical mechanical supplier said, “When you [end users] order standard stuff, it makes our lives easy. I think one of the signs of an industry’s maturity is standards. The reality is that this industry is moving too fast to take a breath and reflect on what maturity would look like.”
“The only real differentiation between our industry and other industries, from the point of view of supply chain, engineering, operations, is speed,” he said. “You move real fast.”
Come back next week for the 4th installment in our top-of-mind series and hear from the Advisory Council on Economic Impact.