Data Sustainability Summit Tech Talk: Net-Zero Carbon Emissions

Data Sustainability Summit Tech Talk: Net-Zero Carbon Emissions

Show Video

I did have a couple of quick things as takeaways before we transition to our next speaker that I think are important and are relevant to [inaudible] conversations. The first is there was a few theme throughout a lot of the comments about the importance of workforce essentially of the people who are doing the key day-to-day work, having the right skills and ability to make these really hard decisions when they're faced by all these challenges and so great news is we have a workforce panel tomorrow and they can get into all of that. The second key takeaway though is I want to challenge both the panels [inaudible] and folks for the rest of the next couple of days.

There is an important discussion about the unique mission, distinct requirements across agencies and how that sometimes ends up in different decisions being met. And that's fully understandable but at the same time, I think that's also a notion that has to be challenged. When I was [inaudible] FBI, I heard every day from somebody saying, "I'm special and different and I can't give up my system, my capability. You need to figure out a way to work around that.

Totally understandable but I think as we move forward with these challenges, we have to think differently as government and industry about how do we do that and how do we recognize we're not going to just keep doing business the way we've been if we want to be successful. I think now that we have that [inaudible] setting from that panel we're going to move on and get a little bit of the industry perspective on that and so it's really my pleasure to introduce our next speaker Jim Connaughton. Jim has one of those bios that when I read it, it makes me wonder what I've been doing all of my life but more importantly, I think his bio is a perfect fit for the Summit. Just a few items of note because we've only got half an hour, he was the Chair of the White House Council on Environmental Quality during the George W. Bush administration. He's been Executive Vice-President at energy companies like Exelon. He was the Executive Vice-President at C3AI and now he's the CEO Nautilus Data Technologies, which if you're not familiar with it, is an emerging data center infrastructure company with some really novel approaches to sustainability.

Jim, it's an honor to have you here today. >> Thank you, Gordon and thanks to everybody. And it's great to be reengaged post-COVID as we work our way through our transition away. And at a really vital time, not just for America, but for the world as we discuss this very critical topic, sustainability of data centers. Gordon and the organizers asked me if I could you know enhance and add to the scene setters, so my mission today for all of you is to amplify and build on the quite thoughtful remarks and panels we just had and then I hope that I can then help dive into [inaudible] dive into the particulars, so I'm going to start at the 50,000 foot level and then work my work way down.

I want to know much has been said about the EO, the Executive Order, I want to go back then as the scene setter go back to the top and in the original environmental statute of the modern era was the National Environmental Policy Act and it's really what drives the Executive Order and a host of federal legislation that followed it back in the end of the sixties. It has a very important phrase and this is a phrase that the White House Council on Environmental Quality operates under. It's the legislative mandate and that's to you know preserve pathways where people and nature can exist in productive harmony. And I love that expression "productive harmony". And that's really the heart of this gathering. It's the heart of what we're about.

It's not data centers, it's not sustainability, it's the combination and our contribution to the productive harmony of society. And so you know I want to be sure we're all focused on that. Secondly, I want to note that we are at an extremely exciting time in tech. We're at a generationally transformational time in tech and we are barely seeing what's coming. You know, we read about it, we got a sense of what's coming but we're on a whole new playing field when it comes to tech. So I want to put that in perspective.

So let's start you know right at the essence in the beginning. For most of humanity's existence, it was the search for water, clean drinking water to sustain life and to support, you know, the food, you know, access to naturally-provided foods. And so life was hard but water and obtaining food was essential throughout you know all of humanity. It's only been very recently in human existence that we've had access to energy okay and we use energy to enhance our lives, to prolong our lives, to keep us warm, to keep us safe, to allow us to interact. And then ultimately, you know, energy gave us the industrial revolution. But think about that, it's only been, you know, hundreds, maybe thousands of years against the arc of humanity.

Data is what enriches life, okay, and the ability to have data and share data and share information is what enriches life. Data is what has given us the modern age and so that water molecule and that electron and now that photon, okay, these are the essential elements of modern life and society. But when you put those elements together in a more thoughtful way, which is what this conference is about and I say this without any sense of glibness or understatement, the data center is the engine of sustainability. And so as we talk about the sustainability of data centers, I think it's important we, as we know why we're talking about that because the data center itself is the power plant is the water treatment plant of sustainability. It is access to data and digital infrastructure that closes the information gap.

The information gap is the education gap, the education gap is the wealth gap, the wealth gap is the piece of democracy gap. And so it's information and education and not just of men but of women that is what is going to be driving the societal decisions globally in terms of informing those decisions, alright? But just as importantly, it's the digital infrastructure that will then be illuminating and enlightening our ability to achieve leaps forward, gigantic leaps forward in sustainability, leaps forward the likes of which we have not yet experienced in society and life because it's the digitalization, not just of information and helping us think about smarter choices, but the actually operational advantages of digitalization that will cause us to finally have smart grid, smart city, smart transportation, telemedicine, you know, and tele-education, material rethinking and design. The material computational work that's being done only at national laboratories today can now be accessible to everybody. So it's very important as we talk about sustainability that we look at it in that context. That means getting hardware right and we're at a truly transformational stage in hardware with the end of Moore's Law, where there's a whole new set of young engineers guided by old engineers who are trying to figure out how we move beyond, you know, just piling more transistors on to a single silicon wafer. [inaudible] whole new series of hardware architectures that are coming.

Software, AI and machine learning, we are barely scratching the surface of that, barely and we're writing and seeing what's really there yet. And that software is wildly computational and intense but at the same time, what it delivers is wildly sustainable but beneficial. And that software has to sit on hardware and since software runs the hardware, the hardware wants to get hotter and so we have to be smarter about where we put our digital infrastructure, how we cool it and how we move our data around. So getting the fiber networks, right and getting the data centers right is essential to that four-legged chair.

And so what does it mean then to get data centers right? We have coined in my company, my colleagues have coined a new phrase, a new thing. Data centers for all of the last 20 years all the talk's been about PUE you know can we make them energy more efficient and we can buy green electrons. Okay, that's fine, that's good, it was a great first step but we've tried to rethink this, which is something we call TRUE, true, which is Total Resource Uses Effectiveness.

The idea is how do we apply a true life cycle thinking? How do we apply true sustainability thinking to the digital infrastructure? And when you're forced to confront TRUE, you're forced to confront how, in the modern area, how laughingly unsustainable the current infrastructure is. And I say that with the affection but I think I'd like to say bold things, I say with affection but if you have to start over again, we would not construct and locate the infrastructure the way we do today. If we knew then that we'd be looking at 10, 20, 50, 100-megawatt data centers and if we'd seen what's coming, which is this incredible network of distributed data center infrastructure, all of which is using tons of electrons generating, you know, massive amounts of heat, we would not design it at all this way.

So how do you get back to basics? You have to apply lifecycle thinking. And so, you know, as we think, as we all should be thinking at the data center, to be the engine of sustainability to support this hardware and the software that's going to drive all the sustainability elsewhere, we need to have our own house in order and it comes down first to performance. So just performance is going to require us to think dramatically different about the infrastructure because the machines want to and need to get hotter and [inaudible] hit a natural point where costs go up and up, it goes up the hotter the computers get. We need a way where costs go down [inaudible] goes down the more [inaudible] or higher performing the computers get. And that's available to us, that's the great thing, it's available to us. Second, environment and energy, think about a data center for a second.

It takes power from far away. There's huge line losses along the way, up to 25%. A massive amount of water is engaged in the cooling of the thermal power. And then that power is delivered into the data center to run [inaudible] and you're handling refrigerants and you're doing water treatment and you're competing with local communities for drinking water, that's just nuts, think about that, that's just nuts. And then you're dumping water in the waste water system.

And then meanwhile, the drinking water plant, energy's being used to take freshwater, convert it to water that we drink, export that to the data center with energy and then blow a whole bunch of that into the air and put the rest of it, after it's been retreated into the waste water system, energy's used to channel the wastewater, it's reprocessed to wastewater and dumped it back out again. I mean that's a lot of conversion steps, utterly unnecessary. And so we need to get rid of that.

And the refrigerants are potent greenhouse gases and ozone-depleting substances are continuously being phased out, so there's huge just regulatory risk in the essential componentry of the data center. And none of that's needed, none of that's needed and we can talk about that because we can just shift to the way everybody else cools and operates thermally-strong industrial infrastructure. That's naturally cold water. Mother Nature's given us a really good solution but we have to think that way.

You have to sort of completely reorder. Why do we put data centers on the commercial side of town? There's a community element to this, which is why don't we solve environmental justice, not only by removing the burden of air pollution, the burden of water pollution, the burden of industrial locational decay, but actually and leaving behind no jobs and instead repurpose that brown field infrastructure and move the data center infrastructure to the side of town where it belongs? You know, you have the power plants, you have the industrial facilities, you have the ports [inaudible] distant to the users, we just have data centers on the wrong side of town. And so we can do more for environmental justice than simply removing the burden. We can restore huge new value to disadvantaged communities and at the same time, add to the resiliency equation. And this opportunity, this [inaudible] point of moving outside, of moving off premise, as mentioned in the prior panel, this gives us the opportunity to use the federal dollar, to use the private sector dollar, to use the consumer dollar to stand up the greenest most powerful, most important infrastructure that the world requires today. And what's the result of that? The result of that is really kind of fun because if you combine superefficient computing with superefficient cooling with zero emission electron procurements, we can generate and take the heat from the data center and use it as a product as opposed to as a waste.

We can actually generate what I call megatons of carbon, so forget that zero per se. The data center can become an engine of net zero carbon by transforming the efficiency of other infrastructure and that's what we should be aiming toward. And by the way, all of that can occur at a net savings to the taxpayer [inaudible] savings to the private sector and why is that important? It means it will actually happen if we can just rethink and re-engineer what we're about and do it in a way that saves money rather than costs money, you don't have to regulate [inaudible] in the future, the market will grab it, absorb it and deploy it with the speed the likes of which we saw with you know, going from the cell phones to smartphones and from cable to fiber and many other innovations that we've all been through. So what it requires, though, is what the last panel just said and then I'll go to questions, go to the moderated session with Gordon, It requires bold decisions and requires decisions as bold as I have a perfectly good dog of the data center but shut it down.

If I can replace it with something that's much cleaner, much higher-performing and much more environmentally beneficial, better for economically-distressed communities, it's okay to shut it down. The math will work but it takes a bold decision to say, "I'm going to go through the headache and heartache of moving my infrastructure and another [inaudible] making great strides in this but not nearly adequate to the opportunity, make bold decisions. There's some rethinking about re-engineering that occurs and just let's stop doing things the same old way. We have winds, solar, we have increasing geothermal, we're rethinking nuclear now.

If the energy system can be making this many transformational decisions, certainly the digital infrastructure community can as well. And we should have a competition because the energy sector's really slow to move and yet they're moving at a pace that far exceeds what the digital infrastructure community's doing. So with that [inaudible] I leave you with this huge opportunity. I'm just delighted for this conversation.

And we have this, again, just this generationally-exciting moment in time to get it right and to get it right for the next century. So thanks. Gordon. >> Jim thanks.

That was great and I think, at least for me, really resonated, particularly, your closing point there about the need to make big decisions and to challenge governments to do that. But I want to get down into brass tacks of that a little bit and talk about, in particular, when you look at it in the context of what Vaughn, Ann and Sylvie are saying that they've just spent a lot of money on this legacy solution, this system that was installed and then on premise data center meeting a mission need that's been around for 20 or 30 or 50 years. And so how do we incentivize people to move off of that? How do we incentivize them to say we can do things differently but we need to step back and rethink some very fundamental principles about our organization and our mission? >> Yeah, well it's like anything else, it's a matter of inertia, okay? And at this point in time, it's only human inertia, okay? We have the technological capability [inaudible] the dollars are there. I mean, this is the, it's really exciting for in the developed world there's huge amounts of new money pouring into infrastructure upgrades and infrastructure replacements. In the developing world, you know, think about that, two-thirds of the growth of digital infrastructure's going to be happening in the developing world starting from scratch, okay? And they can either build the equivalent of the old coal fired power plant, okay, which is, you know, old data center digital technology or they can start fresh with something new, clean, high-performing and efficient and future-proof, okay, and resilient, okay? Those are the choices and right now the choices I'm seeing what's getting built globally and it's, you know, little tiny increments of success and progress when we could be making big ones. And the challenge we face is once you stand that facility up and it's, you know, it's medium performing, you're stuck with it for the next 40 years.

So this is the inflection point that we're at but it is inertia. Now what's great about inertia is once a handful, just a handful of leaders make big decisions and prove their success, the lemming effect is really powerful, okay? You know, the lemming effect is really powerful and that's what I love about the conservative engineering sectors. It's "Show me, you know, I'm not going to be the first to try, you know; I'm not going to do this. I don't believe it can really work." But the second it works and they can kick it and twist it and bend it and play with it and understand it, it's like "Oh well, why in the heck didn't I do that?" I'm going to use a light example, just going from the, you know, the landline connected phone to the wireless phone, okay. I mean that was less science fiction and yet that transformation took 5 years.

And then just when we thought "Cell phones, wow, cell phones. I got my little, you know, Nokia cell phone isn't that awesome?" Five years later, boom, cell phones gone, everyone's got smartphones and it's an entirely new set of companies, okay. So infrastructure's slower, so it's harder to get that 5-year cycle but infrastructure can turn around in 10 to 15 years. We did rural electrification in 15 years.

We built up the nuclear fleet in 15 years, okay? We built out the interstate highway system in 20 years, okay? So these big shifts are entirely within our wherewithal and capability but it requires everyone walking norms and saying "We're going to make the bold decision to try the new thing and be confident, you know, in what it can do and then go." That's what it takes. >> Yes, thanks Jim and I don't know if you knew this, it was unintentional I'm sure, but I used to work for Motorola so I'm familiar with the sort of whipsawing effect with the pace of innovation in the cell phone industry and that certainly resonates very strongly with me. I think, you know, as a follow-up to that point and there are some questions coming in around this as well, there is a [inaudible] the government to an incentives aspect to this as well, you know, are we engaging the right people and encouraging them to make the right decisions? And I know you know this from your time in government, but a lot of the leaders who are making these decisions, they're relatively short in their tenure and the time scale of what we're talking about sometimes it is like you said, 5, 10, 15 years, so how do we incentivize somebody who's in a role for a couple of years, 3 years to be willing to make that decision to commit to something that's going to take 10 years to recognize? >> Yeah, I'll use LED lighting as the best example of that, okay, You know, huge inertia to just sticking with incandescence, you know, with incremental improvements of efficiency and materials associated with that.

Of course, and all the economic value of the plan [inaudible] of that, right? So the incumbents just had zero incentive to change, okay. And yet the tech was [inaudible] reliably and quickly, albeit, you know, if you think about it in LED light 10 years ago, you know, cost 10 times as much, right? So, you know, this perception of an economic drag. There is an example, that one actually took a law by the way, the bipartisan law to just set a new standard, not a technology specification, but just a standard. But it was based on very well-informed technology understanding and scalable production understanding to set a standard that was commercially and technically viable, okay? It required 10 years, okay? Now, at the high level, the level of our prior panelists, that's where those decisions are made. To take that 2 years or 3 years on the job and make the hard decision on technological viability and economic feasibility, okay, with an informed judgment about availability, okay, that needs to get made at the top, okay? And then for the folks who just, you know, the engine of the government, you know, everybody in government that I worked with and I had so many great experiences, they're always looking for stability to say yes but their incentives are aligned to saying no because the conservatism of being held accountable and that's true for all us, you know.

So, how do you give people permission to get to yes, that's how you do it. You do it with the standard setting, technology assessments and then and the driving thing such as the executive order. It's a very powerful executive order but an executive order's only as good as the people who've read it and want to do it. There's no laws, no legal power to an executive order.

Executive order is an exhortation of better conduct and action and it still comes down to people. But we can do big things. Now, I gave an example of a law but there's equally good examples of technology choice so the leadership and procurement. So by the way, all this off-premise stuff, so I hear a lot about the effort to, you know, make the U. S. owned data centers greener but more important since more than 50%

of new procurement is off-premise is to insist on higher standards of performance and outcomes of the, you know, off-premise providers, the big public cloud providers and the private communication providers and to insist on that, to make that a procurement decision. And it's funny how the market works. We love, as providers, to give our customers what they want. The customers have to be bold and specific about what it is they want. Right now, this sustainability is not being preferenced. No one will pay more for it [inaudible] like I said, they don't have to.

But it's not being preferenced at scale across the board in terms of the volume and the size of opportunity that can be delivered. Now mission critical is the most critical thing, so I understand why it's not being preferenced but I would suggest it shouldn't be preferenced but it should be required once the other threshold conditions have been met. And so we've got to, we have a long way to go there yet but the opportunities [inaudible]. >> Yeah, I do think there's a bit of like dichotomy there, Jim, in the sense that the mission critical applications, those are the ones that they've got the resources from their agency to say it's essential that we do this function, it's a core mission agency function.

It's a mission essential function, here's the resources to go do it. And so that's why I think sometimes the incentive structure isn't as aligned because we can't stop doing this because it's so core to the agency function. So, I guess the question then is, and you already started to address this, you know, we need to speak to not just the technologist but to the acquisition professionals and others.

Who else, though, is in the audience, do you think are going to be effective as decision makers to help drive that change, to tell an agency "Yeah, it is acceptable to have sustainability be a core mission function for the agency and have that be a part of the decision-making process?" >> Yeah, this is a great one [inaudible] wrap this up on which is how do you make sustainability mission critical? >> Right. >> Okay and where I'm excited is it's the chip and server designers that are going to make that happen, okay? And they're going to make happen is the users want more AI, they want more machine learning, they want more computational capability for their research programs and we're going to want, you know, sub-millisecond operational control. I mean, I hope everyone's as frustrated as I am that we don't have smart city yet. We don't have smart grid yet. You don't have smart transportation yet. And yet we have all of the technologies in the end of the day exists.

All the technology to make that possible exists. It's the lack of integrated infrastructure and system to deliver on that that we don't have. And so how sustainability becomes mission critical is that new capability demands a more sustainable model for housing and cooling and operating those machines, okay? It requires it because conventional methods cannot support it at a declining cost, right? They can only support it at an increasing cost.

And so I'm excited because the driver is finally here as these machines are pushing, you know, 15 kilowatts [inaudible]. Well, we have to rethink this and so, for me, it's only a question of how soon not whether it's going to happen and why doesn't everybody join in? Well, let's not wait, we don't have to wait, let's make it happen now. We don't need anybody's help to make it happen; we can just decide to buy. >> Yeah, I think just one sort of additional follow-up point to that, Jim and then we probably need to wrap up. In my mind, the hardware is necessary and essential, you're absolutely right but the system design, as a whole, is equally important, right? And so we need the entire ecosystem of technologists too, the people who are building the applications, to understand the choices that they're making, too.

We're giving them their use of the cloud, all this incredible ability to have at scale and to turn on and turn off services when they want to use them and that's all great. But they need to then start thinking about how do they understand sustainability implications of the decisions that they're making as well at the software and at the application level. And I think that that's not a place where we've really gotten to in the government yet, to thinking about those choices in the design. >> Yeah, let me make sure that it's a positive statement though, which is we actually don't have, you know, we have a data center problem in that respect.

We have a new energy construction problem, okay, which is I'm all for a lot more use of digital infrastructure for the reasons I described because it's going to drive, at the macro level, you know, us hugely forward in sustainability. So I'm all for using more energy in data centers. We just shouldn't waste any and we shouldn't put any emissions into the air and we shouldn't waste any water in doing it. And so who needs chemicals? We don't need chemical to do it.

And so what I want is more cleaner energy going into data centers used much more efficiently and then the only byproduct of a data center, at that point, is heat. And we can actually turn the heat into a product, too. You can't do it very effectively, I mean It's physically possible and it works but in the air-to-air exchange methods we use but you certainly can go liquid-to-liquid.

It's just a really easy thing to do but you've got to co-locate the data center with the infrastructure that can take advantage of it and that's just a decision, that's just a choice. There's no impediment other than human imagination and a fortitude of making a sensible decision. >> Thanks Jim and I feel like that that last comment [inaudible] that was a great sort of teaser for a lot of the great work that Nautilus has been doing around the use of liquid-to-liquid transfer heat and I encourage folks who are not familiar to do a little bit of research and to look up some of the great work that Nautilus has been doing.

Jim, I know I could keep asking you questions for hours but we would probably drive the audience away. So I want to thank you for taking the time to be with us today to help with some of this in context and to really challenge some of the decision making and to encourage people to really step forward and seize the moment because the time is here but the need is also there as well. So again, thanks for joining us today and I'll ask the audience to give a nice virtual round of applause for Jim before I move on in the agenda.

2023-01-20 19:57

Show Video

Other news