Exploring Sustainable Data Centers – David McGuirk, QiO Technologies

This is Techstrong tv. Hello and welcome to Techstrong tv. I'm Bonnie Schneider, and joining me today is Dave McGirk.

He is the director of data center and telecoms at Q I O technologies, with over two decades of experience in data center design, architecture, and delivery. Dave is a leading expert in sustainability, AI and data centers. Today we're diving deep into the world of sustainability and predictive modeling for effective C P U Power management.

Welcome, Dave. It's great to have you here. Um, thank you very much for having me, Bonnie.

Oh, it's great. Um, you know, we wanna talk a little bit about your background as well, and kind of talk about how you got into this, uh, field of data center design and sense and sustainability. Can you walk us through that and, uh, explain how this experience is, what you're doing now, how it, how, how it evolved to what you're doing now?

Absolutely. Um, well I started my career, um, many, many years ago in, um, selling infrastructure solutions. So network switches with the likes of hp, Cisco, and the likes.

Now my career developed into, uh, data center design. I'm actually providing, uh, safe, secure, protected environments for the IT stack. Um, and gradually over the years, that's transcended into very large enterprise class data centers from a design and build perspective.

The last, um, few years of my career, um, have been focused primarily on driving sustainability and improving efficiency of data centers. So I've got hands-on experience in several different types of technology ranging from the likes of Emerge Cooling, um, all the way through to what I'm doing today with Q I O technologies, which is essentially AI software that improves the energy efficiency of data centers and really does enhance sustainability credentials of an organization. That makes sense.

So how do you leverage artificial intelligence to help organizations reduce their greenhouse gas emissions? Okay, well, Q I O technologies, we are an AI software products company. Um, we support energy intensive and asset heavy companies to rapidly deliver on their sustainability and net zero agenda.

Now we provide a suite of data center solutions, which very much address the challenges faced by DC owners and operators. What we're doing is that we ensure the accurate measurements and baselining of key sustainability metrics associated with, uh, data center operations, and also making sure that we can provide regulatory compliance as well around E S G reporting. Now, whilst our dashboarding offers some of the most powerful data visualizations of all assets, uh, across the data center, the most important aspect of our solution are the tangible benefits and results that we can achieve by using machine learning and ai.

Now, what we do do is we provide very significant operational cost savings and also energy and G H G reductions, uh, to data center operators in as little as four weeks. Now q i o we are very much at the forefront of energy, uh, efficient server management, and along with Intel, we've developed an innovative solution that leverages the power of artificial intelligence and also predictive modeling, which does significantly reduce server energy consumption. We've found that by optimizing, uh, C P U and power management associated with it through AI driven insights that we can achieve remarkable energy savings of up to 52% across a server fleet.

That's amazing. So how does predictive modeling play this role in, in managing C P U power? Can you, our audience is very technical, so if you wanna just walk through a little bit about how it works, uh, to reduce the environment, environmental footprint of data centers.

Essentially what we're doing is we're taking key, uh, C P U telemetry and from taking this data, what we're able to do is to see how the C P U is operating. Now, we don't take any data from, uh, from the patient layer. It is purely from the C P U drilling down to levels.

We're understanding workloads and how the C P U is operating. Then through predictive modeling, once we've learned, um, the plantation of, um, the workload, we are able to predictively put the C P U into the correct P state or C state, i e the power management state in order to match that time of day workload, and that's how we go about generated such large savings. That's great.

Well, what are some of the common challenges that you find organizations are facing when striving to make data centers more sustainable? Well, there are a, a number of challenges, um, and the, the very broad as well. So we've, we've got organizational challenges.

We find, um, uh, a lot of sustainability will, will attest to is that quite often KPIs around sustainability don't flow down the organization. So it's very important that we engage the likes of SecOps teams and infra teams very early in an engagement themselves to make them and to let stand what our goal is and what we're trying to achieve. Um, and that's, that's the sort of the, the first step from technology, um, issues as well in terms of driving sustainability.

Um, it's one of our sort of motivations in terms of developing the solution. The uptime institutes who recently, uh, published the latest report, uh, around energy efficiency, uh, across data centers. Interestingly, they've seen that p u e, which is generally the overriding, um, measurement of efficiency of a data center, has remained relatively static for the last two years.

And this is signifying that most of the easy wins have already been achieved within data center infrastructure, but the focus has always, uh, been on the likes of cooling optimization, for instance, or more efficient, uh, hardware. It could be more efficient, UPSs, et cetera. Now, what we've done, and whilst we provide, uh, cooling optimization in which we can save a 15%, uh, energy reduction, what we've done is we've looked to address the cause of, uh, the heat, which essentially is the IT systems themselves.

And we feel that the industry very much needs to focus on the IT equipment load Yeah. Make that more efficient and that will have a knock on effect on cooling, et cetera, but it's very critical for an organization to address. Yeah.

How does your team approach the improvement of energy efficiency in data center design and operation, and when it comes to the, and what kind of outcomes have you seen from your experience? Well, patient, well, we've had the privilege of working with many forward figured organizations, and that's across various industries. So for instance, we've worked with the likes of Rolls Royce Continental, um, British Telecom.

Um, we are also developing our partner ecosystem, so we're working very closely with the major OEMs, uh, for networking devices and servers. Um, lots of, uh, global systems integrators, the likes of W W T and s h I. Our approach is one of innovation and collaboration, and it's, it's important to note as well here is that we have taken our expertise and our heritage and industrial and manufacturing and incredibly complex environments, and we take that and we put it into the data center itself so we can make savings on a vast array of asset types.

So from automotive, uh, production lines through brick manufacturing kilns to jet engines, all the way down to a micro level of individual cause within the C P U itself. Now, here's another question. Um, it's, and you mentioned Rolls Royce, so that always comes to mind is, you know, luxury and all that, but I'm wondering about sustainability initiatives when it comes to the financial investment in, in reaching those goals.

Unfortunately, there is an, a conception that it's costly, okay, we wanna reduce emissions, but how much is this gonna cost us? But aren't there ways to show that this actual process could reduce costs when it's done properly? Absolutely, Bonnie, and that sort of goes back to my, uh, initial, uh, point, the, um, in terms of of we are delivering huge cost savings across multiple industries.

Yeah. So we, our business case is, uh, irrefutable. We have market leader ROIs and in some cases we're also seeing payback within a matter of months.

So whilst with any sustainability initiative, it does require work, it requires effort. Nobody has a magic, uh, bullet or a button that can be pressed that will instantaneously, uh, reduce G H G emissions. So we do work in collaboration with our customers.

A Rolls-Royce is an interesting case in mind. They provide, uh, uh, it's actually the aviation, uh, arm, they provide power by the hour on the jet engine fleet. So what we're all about here is orchestrating end-to-end, uh, repair workflows.

So making sure that performance of the jet engine is, uh, at peak levels and then we can, uh, order parts, for instance, to be delivered to a certain service center. So essentially what we're doing here is we are improving any supply chain issues and also we're keeping jet engines in the air for longer and making them more efficient, and that's key. That's great.

And efficiency With hybrid cloud environments becoming increasingly prevalent, how does technology adapt across both on-premises and at the, in the cloud-based data center infrastructures? It's a really good question. Um, obviously, um, it's one that we are very keen to address.

So we are looking to work with a number of the, uh, public cloud providers, uh, in providing C P U power management, um, to which that can really improve. The customers, uh, also are reported on G H G A emissions and metrics. Sadly, the industry has fallen, uh, a little bit behind in terms of reporting on metrics, um, especially from the cloud providers themselves.

But we're seeing great in interest in our solution, and obviously we can deliver significant energy savings, which then can be passed on to the customer. But within the, the hybrid environment, obviously our solution, um, is based on-prem. So data sovereignty remains with the customer, and again, we are making sometimes investment decisions into the cloud can be, uh, avoided because we can reduce the on-prem element significantly and we can provide market lead and efficiency.

That is fascinating. It seems like you're really on the, uh, cutting edge of all of this. Uh, where do you see the future of sustainable data centers going?

As I mentioned to you, we feel that the IT stack itself is critical, uh, in terms of improving efficiency. Obviously there's a software layer in the application layer, so work needs to be done there. We're seeing great advancements in, uh, chip cooling, uh, immersive cooling, et cetera.

So that's improving all of the time. What we're seeing is, uh, it's not just the data center environment. So our solution can be applied to base stations, it can be at the on-prem, the cloud environment itself.

So we can essentially provide efficiency and improve efficiency from chip to grid. Great. We'll be have time for one more question.

Uh, what would be a key piece of advice that you would offer organizations that are looking to enhance the sustainability of their data center operations? First element is to baseline accurately. There's still a massive, uh, lack of data available to data center operators.

My advice would be to take it at a granular level and build the picture up from bottom to up and up to down, and then we meet in the middle. Then from there, once that granularity of data, uh, is achieved, then targets can be set that are achievable. And as mentioned, we then offer the pathway to reduce G H D emissions, um, massively reduce energy costs.

That's great advice. Thank you so much. Dave McGuirk, uh, director of data center and telecoms at Q I O technologies for sharing your extensive knowledge on data center sustainability and AI driven solutions.

Thank you. Thank you very much, Bonnie. It's been my pleasure.

Wonderful. Well stay with us on Techstrong tv. We have more great interviews like this one coming up.