Intel's Navin Shenoy on the Future of the Data Warehouse
If you think of Intel as a PC company, think again. The firm's Data Warehouse business generated more than $17 billion in revenue last year, and it will do even better this year.
Last week, Intel launched its newest and most powerful CPU at the New Lab in Brooklyn. Dubbed Xeon Scalable Processor Platform, this new CPU will enable everything from autonomous driving, artificial intelligence, and the global transition to 5G networks.
After the announcement, I sat down with Navin Shenoy, Intel's new EVP and General Manager for the Data Center Group, to get his take on our super-connected future.
Dan Costa: Let's talk about the technology itself that you launched today. The Xeon Scalable Processor Platform. What makes this a scalable platform?
Navin Shenoy: Sure. As we were thinking about this product—by the way, we've been working on this for five years; this is a long-term development product—we thought about the needs of the workloads in the data center as we look at things today. You talked about a few of them, autonomous driving ... it happens to be Prime Day today, so the workloads happening around us today, by definition, require scalability. Things that happened last year may happen, order of magnitude, faster this year. It's unpredictable.
And, the data center infrastructure needs to be able to handle that scalability. It needs to be able to scale up; it needs to be able to scale down, depending on what's happening with the workload. And so, the dynamism, the dynamic nature of the way the technology world operates today, the way that workloads are developing [led] us to think about, okay, well first of all, how do we design this thing for scalability? And, that's what we were thinking about naming it. We thought, why not just use that in the name, and signify that there's something new here?
This Xeon Scalable processor is the biggest advancement we've made in the data center in a decade. We don't take that lightly. Products like this come along maybe once or twice in your career, if you're lucky. I'm lucky enough to, and privileged enough, to have been able to launch it today.
So, when you talk about making it scalable, truly a system that's designed specifically for the data center. And, you design them a little bit differently than you would a system that was going to go in a PC or a workstation. Can you talk a little bit about what's required for a data center-specific processor?
Three things. First, we need to define it and design it for ultimate performance, a magnitude higher performance than you would have in a client device. Second, security that's built-in for workloads that matter in the data center; encryption, managing your data and securing your data at rest, in flight. And third, ensuring that you have agility and the ability to move workloads from the private cloud to the public cloud, for example. All that stuff gets translated into the way you design the product. This is [holds up wafer]—
Yeah, you brought show and tell. We usually don't have show and tell on the podcast.
Well, I'm going to violate your rule, I guess. This is Fast Forward. This is what Fast Forward's all about, right here. This is a wafer of the new Scalable Xeon processor. And, it's amazing to me that somehow all this stuff works. There are billions of transistors on each of these products. But, this is done at a scale that you would never see in a PC, or in a laptop, or in a phone—28 cores, 50 percent more PCIe and memory bandwidth. A brand-new mesh architecture that allows data to move seamlessly between all the cores.
The mesh architecture seems very intuitive, and a brilliant way of improving the chip.
It's kind of like, depending on what your view of traffic in Manhattan is, the grid system of traffic in Manhattan, where you have things that are north-south and east-west, and you're able to cut back and forth and go in any direction. [That's not] really how processors were historically designed. If you had to move data from one processor in the lower-left part of the die to the upper-right part of the die, you had to go all the way around, and go through a bunch of buffers and other things that would slow down the latency.
Now, you're able to move data seamlessly from core to core, without having to go all the way around the ring. And, that's really what the mesh was designed for. It's very unique. We think it's going to help us deliver the biggest gen-on-gen performance improvement that we've had in the last decade. We anticipate about a 65 percent improvement, generation to generation, on a broad range of workloads. On artificial intelligence workloads, up to 2x improvement, based on the hardware, and up to 100x improvement when you add software optimization on top of that hardware. So, I'm super excited about the product, and we'll see what happens from here.
And, you already have units in the field. You're launching today, but you have a number of partners that have been using this system for months now.
Yeah, you picked up on that, that's great. This is the general availability launch, but we did something different this time. We did an early ship program, starting in November of last year to get this product into the hands of our most demanding customers. About 30 customers have their hands on the product. About 500,000 units we've already shipped. And, you heard a number of those customers come on stage today and talk about the benefits they're seeing from the performance, the security, the agility that we're delivering. Companies like AT&T, companies like Google, companies like Amazon Web Services. Health care companies like Monte Fiore.
Yeah, their examples were fascinating. A totally different approach to healthcare.
Yeah, the ability to predict what's going to happen to somebody through looking at their genomic data, looking at their pattern of health, looking at their pattern of doctor visits, looking at their health history, looking at their family's history. Combining all that stuff, using data analytics, to say, "Hey listen, five years from now, your probability of getting such-and-such is X, and therefore, you might want to consider the following lifestyle changes." That's really the future of healthcare.
Anyway, it's great to have those enterprise using this ... I got to tell you, there's an insatiable appetite to get these products into the hands of customers earlier and earlier. And that requires a new way of thinking in terms of the way we develop and launch products.
Yeah, it seems like, and I keep hearing this described, as there's an infinite desire for compute. There's just no maximum, there's no cap. We just need more. And, we need more for these cutting-edge applications. And, it's across multiple industries.
One of the things I talked about today is, a lot of pundits in the industry talk about the flood of data that we're seeing, generated by all the sensors out there, and the move to video, and higher definition video, and virtual reality ... All that stuff is generating lots of data. But, I think what's more interesting is that we're literally nowhere in terms of actually doing something with all that data. Processing it, analyzing it. Our estimate is that less than 1 percent of the data generated on a daily, weekly, monthly, yearly basis is actually being used. And so, as that data comes to be analyzed and used, that's just going to drive the demand for more processing capability, and that's why we have such an insatiable appetite. Because, we're just scratching the surface of what we could do with all that data.
It seems like that's the next big challenge, is that we can capture the data, we can store the data. But, we're just now starting to be able to make sense of that data. But, that's where we need this infinite compute stack.
That's right. So, it's good for us. It's good for our customers. It's good for many others in the industry. So, we're excited about what the future holds.
Very cool. I've got some standard questions I ask all of our guests. What technology trend concerns you the most going forward? Is there anything that keeps you up at night that you're worried about?
There is a need to think about the societal impact of artificial intelligence. To think about, what does it mean for jobs? What does it mean for retraining? Retraining, so that people have the skills so that they can continue to be employed. While I worry about that, I'm also optimistic that we, as a society, will figure those things out, just like we have in previous technology deflections.
I'm also optimistic about artificial intelligence. We talked about the healthcare one. I talked about farming, and how, to feed the population on Earth, farmers need to find another 50 percent increase in production of food. There's no more farmland. In fact, there's probably less farmland, over time. So, how do we increase farm production by 50 percent with less farmland? Those problems are only going to be solved through the application of technology.
It seems like there are a number of problems that we have to deal with today. We've always been able to innovate our way out, or around, those problems. AI seems like a fundamental technology to help us do that. In terms of the optimism, you mentioned you're optimistic about AI. Is there anything that you are more optimistic about, that gives you more hope in the technology sector?
I think we are at the verge of solving some of the greatest problems in human history. And, all of that will be solved through some application of technology. Whether that be healthcare-related issues, finding a cure to cancer ... That will happen. Hopefully in our lifetime. And, that will happen through the application of technology.
Autonomous driving. It's amazing, to me, how much car infrastructure is deployed and how little it's utilized. Four percent of cars in the world are actually utilized. And, if you look at the trillions of dollars of capital that are deployed, in terms of automobiles, and the utilization rate of that, you go, well, this is insane. You would never do that in a traditional factory, for example. So, our ability to increase the utilization of that technology and improve safety. Dramatically improve safety ... I think, is something I'm optimistic about.
I lean on the side of optimism. I believe that technology flows like a river. It's never going to stop. And, it's up to us in the industry to go figure out how to take advantage of it.
Is there a single device that you wear, that you carry, that you use every day, or an app that's really transformed your life?
I'm not going to say the phone, because you told me not to, everyone else says, right? I'll tell you, I just had 30 people from my family at my house, all staying in different rooms, and the one thing that was the saving grace for us was the Sonos system I have in the house, because everyone wanted to hear something different, and I got tired of hearing Hamilton play over, and over, and over again. So, I was able to play different music in my room, and my family was able to play other things in other rooms. So, Sonos saved me, I would say. That's one that comes to mind.
Yeah. It's great. And, you've got access to your entire library. In my lifetime, we've gone from scarcity in music to being able to listen to any song ever recorded just by asking for it.
My niece wanted to hear "Fireball" over, and over, and over again, and I refused to buy it, but we could find it online for free, so it was great.
For more Fast Forward with Dan Costa, subscribe to the podcast. On iOS, download Apple's Podcasts app, search for "Fast Forward" and subscribe. On Android, download the Stitcher Radio for Podcasts app via Google Play.
This article originally appeared on PCMag.com.