Advancing Life & Work
HPE-Editor

HPE Tech Talk Podcast: Tech Predictions for 2021 and Beyond, Episode 9

What direction will tech take in the years to come? Join Hewlett Packard Lab’s, Andrew Wheeler as he shares his perspective on energy consumption, the capability gap, AI and the future of tech that will shape our lives in the years ahead.


 

Also available on:   Spotify   /   Apple Podcasts   /   Other podcast apps

 

Transcript:

Robert Christiansen:

[0:12] Welcome to HPE Tech Talk. I'm your host Robert Christiansen. …

[0:24] Few could have predicted the disruptions of 2020 thrown our way, but technology was the guiding light that saw us through. Now that we're looking forward to the 2021, when we're facing a year of continuing resilience and one of rebuilding.

[0:39] I'm so excited today to be joined by one of my colleagues, Andrew Wheeler, he's the VP and Deputy Director of Hewlett Packard Labs to delve into some of these fantastic predictions that I believe are really going to be earth-shaking coming up here for us and how the technology's going to impact us. Andrew, welcome to the show.

Andrew Wheeler:

Hey Robert. Glad to be here and great to talk to you today.

Robert Christiansen:

[1:13] Would you just give us a background on the 55 years that Labs has been around and some of the meaningful impact that it has had to, I just say the globe, as a whole?

Andrew Wheeler:

[1:25] Yeah, would love to. I'm really fortunate that I love what I do as far as being the Director for Hewlett Packard Labs. Part of that is because of the history. The Labs Organization was originally founded by Bill Hewlett and Dave Packard themselves, way back in 1966, when the company, only had 400 people at the time, but really is a testament to their foresight and leadership.

[2:30] With that kind of rich heritage in mind, there's been just a series of remarkable breakthroughs that have come through Labs over the years. It really started with, going way back to things like the first programmable scientific desktop calculator, and then the pocket calculator, which Bill Hewlett, himself, really made famous.

[2:57] But when I really think about those major breakthroughs, I'd have to throw in there, ink jet printing. We were the first to commercialize risk computing, right? That technology came out of Labs.

[6:59] Today, we support that entire vision that the company has around being an edge to cloud platform as a service company. That means we're off developing and researching technologies that underpin all of the “as a service” offerings that the company's looking to put out there. Then, from a research portfolio, that means we're looking at advanced system architectures, AI machine learning is really a hot area right now, of course. Another big area is what I'll call “computational accelerators.” Networking and distributed systems. We look at advanced interconnects, which includes Photonix, as well as security.

Robert Christiansen:

[8:38] When you dive into this, what's the first prediction that you're seeing now coming into 2021?

Andrew Wheeler:

[10:40] … Probably the first one I want to highlight is really a kind of a core global challenge, again, that influences a lot of our thinking. I refer to this as a “sustainability problem,” or we also call it the “energy gap.”

[10:57] You go to any IT seminar and, if you've been in the industry long enough, it's all very common speak right now to talk about the explosion of data. I like to refer to it as the expansion of our digital universe … We're on track to be generating on the order of a zettabyte of data every day.

[11:48] … To put that into perspective, a zettabyte of data. Let's say you've got a terabyte hard drive, probably in your own laptop or desktop computer these days, that's pretty common. You would need a billion terabyte hard drives to store one zettabyte of data. That gives you a sense of this data, this expansion of the digital universe we're talking about.

[12:16] Now, why do I mention all of that? IT infrastructure today consumes over 10% of the global energy production in the world.

[12:47] That's only going to grow. You start to put it at that scale and you can see clearly, there's an energy issue here. There's a sustainability problem that has to be addressed. You look at all the other advancements coming online, the trends with 5G, the level of connectivity and bandwidth it's going to drive. All of the autonomous things that are going to be out there. The fact that we're going to be embedding intelligence everywhere. That problem's only going to get worse.

Robert Christiansen:

[13:20] Would you say then that the prediction is in the area of investments and breakthroughs and how do we do more with less energy? Is that correct?

Andrew Wheeler:

[13:32] Correct. Absolutely.

Robert Christiansen:

[13:33] When you make that kind of prediction, at that scale, I actually thought you were being conservative at 10%. I believe the prediction is 20% of the global electrical consumption will be brought in by a computer system of some sort.

Andrew Wheeler:

[13:50] Yeah, yeah. Picking the year, whether that's two years from now, three years, five years, that's, to your point, that is clearly the trend. It's going up and to the right.

Robert Christiansen:

{14:00] Exactly. Driving that is absolutely a global issue. What's the second prediction you would make specifically around, you had mentioned earlier about AI and just the democratization of AI. Do you have any feelings about where that's heading?

Andrew Wheeler:

[14:25] The other trend, or I would say the uber trend, is doing what I call addressing, what we call, the “capability gap.” That's where things like: Heterogeneous architectures, accelerators in AI come into play.

[14:50] When you boil it down … ultimately, you're going to get to the core infrastructure that hosts all of these services and workloads.

[15:58] … Right now, our ambitions are going a little bit faster than what our core infrastructure or computers can improve upon. That's the capability gap that we talk about. That really starts with what we call heterogeneous systems. We see this already today, especially in the world of high performance computing or even hosting some of the initial AI machine learning workloads.

[16:33] This is where you're pairing kind of conventional general-purpose processors with graphics processing units or GPUs. We kind of refer to that combination as the GP GPU combination, we’re now in a system where I've got different forms of compute. Whether you want to think of these as an offload model or an accelerator model, we're at the early days of that—where you plug in these different types of accelerators, because they relate to a particular type of workload and can give you much better performance or either just give you raw performance, or it gives you better performance per watt, per energy unit.

[17:23] Heterogeneous computing shows this really great potential.

[17:58] … But over time, the holy grail of that trend is going to be more and more dynamic. AI is part of what will get us there. Today, it's kind of rule-based, and again, very kind of statically assigning things, but over time, we're going to be very dynamic in how we kind of compose these different systems, given the workload that we're throwing at it.

Robert Christiansen: [18:42] … Yeah, that's the real holy grail, isn't it?
Andrew Wheeler:

[18:44] And I'm not coming at this purely from a hardware standpoint. Really the point you're starting to get to is that's going to drive advancements in a lot of different things, whether it's the runtime software stack, kind of the deployment, how we deploy applications and the management of those. You think about virtual machines and containers being state-of-the-art today. Well, what does that mean to run on a different type of compute model? Heterogeneous systems starts to provide kind of that big trend where things are headed.

[19:56] … So, really that future of computing, relying on a very rich ecosystem of accelerators. Again, these are special purpose computing engines that are capable of performing these critical tasks or workloads much faster, using much less energy than just a general purpose solution that largely dominates the industry today.

Robert Christiansen: [21:10] That's great.
Andrew Wheeler:

[21:11] A final trend I'll hit here is, we've hit already a little bit, is just AI. AI, really being one of the keys to staying ahead of this data explosion.

[21:26] We're looking to really augment the human capability for how we derive insight from the mountains of data that's coming at us.

[21:53] The concept of trustworthy AI is a real emerging topic in the world of AI right now. What are those techniques for? How do you explain the result that you got? What's the robustness and the privacy and security around it? These are all very critical components on how we protect our data, protect privacy. It's something that's got to be part of a robust AI data platform. So, trustworthy AI is a big one.

[22:30] Another one is what we call applications of AI for IT operations or AI ops. This one, again, relates to how we really kind of manage everything, whether it's a data center resiliency and energy efficiency. But really optimizing end-to-end those data center operations. Kind of taking the “human out of the dashboard” view of things and using AI to help optimize that entire end-to-end system.

Robert Christiansen:

[23:08] Lastly here, how does this work with the way we live and work? This is key to HPE and how we walk and talk as a company, it's a cultural thing, that we improve the way people live and work. As we become a force for good in the world, how do these trends really start showing up in, I think, called the ethical computing. I think you're talking about that in the AI piece, right? When you were saying, hey, eliminating bias as best we can. Can you provide me evidence about how an answer was given like that, those type of things? So those were the trends I think, that are related to being a force for good. Wouldn't you agree with that?

Andrew Wheeler:

[23:46] I absolutely agree with that. I'm glad you hit this one. Advancing the way people live and work, absolutely core to the culture that our CEO, Antonio Neri, is cultivating.

[24:18] … As far as these trends go, and I would say, even looking for me a little closer to home, within Labs and how we're applying this today.

[25:19] … And one of the things kicked off, was something called the COVID-19 open research dataset. This was a collection of academic scholarly publications on anything and everything related to COVID-19 Coronavirus from across the world.

[26:07] … We had an individual researcher within Labs, one of our key kind of AI researchers, was already doing some work around developing a pipeline for natural language processing. He and a few others quickly kind of adapted that pipeline and that work. They stood up a question and answer research tool that allows medical professionals, other researchers to do queries against this open research dataset… It's covid19.labs.hpe.com.

Robert Christiansen: [28:22] That's fantastic, Andrew. That is really, really, really interesting predictions here for the upcoming years. I really thank you so much for joining us for here on the HPE tech talk. Thank you.
Andrew Wheeler:

[28:33] Glad to participate, Robert. Always enjoy it.

Robert Christiansen: [28:45]  And to our listeners, thank you for tuning in. Stay tuned for upcoming episodes, where we will be discussing hot topics and news of the day with the leading experts from HPE.  We will catch you next time. Bye-bye.

 

 


HPE Editor
Hewlett Packard Enterprise

twitter.com/hpe
linkedin.com/company/hewlett-packard-enterprise
hpe.com

HPE Editor
About the Author

HPE-Editor