That's a very interesting video, not only because it has so many of our friends that we all know in it, the very familiar faces, but it also has a very familiar theme in it. The theme that the world is changing and this web scale architecture is becoming necessary. It's not just a cost savings. It's becoming necessary to actually achieve what we want to achieve. You're going to see that's what we're going to dig into. My name's Darren Thomas. I run the storage business at Micron, and I'm very pleased to welcome you all here to the MSSC, the Micron Storage Solution Center. That's the name of this place. We are sitting in an area that has not yet been built out. It was actually scheduled to start construction, and we decided to use it for this instead.
They're going to start construction almost as soon as we get out. This is our new software development center. There's a lot of things that occur in this building that are very unique, and our agenda is going to be very unique. We're going to talk about this scale-out architecture, web scale. There's a lot of terms used to describe this. We're going to talk about this model and how Micron is central to it. You're going to see I've got several of my team members that are going to come up, talk about the software component to it, talk about some of the announcements we're going to make today. Some of them are, to be quite honest, exactly what you'd expect us to do. We're going to announce some new SSDs.
Some of the announcements are going to be very unexpected, and I look forward to your reaction as well as maybe some conversation afterwards because this is a very different model for Micron. You're demonstrating what Micron has said over and over again, that we are transforming into a true storage company. As we do this transformation, events like this start to become second nature to us. It's going to be a fairly packed agenda. With that, I want to get right on to it. It's my great pleasure to introduce our President and CEO, Mr. Mark Durcan.
Thanks, Darren. Thank you all for joining us here today. It's really a unique gathering of different classes of people that you obviously haven't previously seen Micron bring together. I think it is indicative of the fact that Micron is changing, and we're changing in many significant ways, but really as a reaction to the industry changing. Many of you know us as a component supplier of leading-edge NAND flash and DRAM and NOR flash and SRAM components over the years. As Moore's Law is starting to slow, we really have a great opportunity to change the nature of our business.
How that works is that when we can take the time to become much more intimate with our customers and develop solutions that really help drive value for them, it drives a business that's much more about differentiated products, value-added solutions, and system solutions that really are sort of changing the nature of our business. That's what brings us here today, is to talk more about that, and in particular, as it applies to the data center, because in many ways, this is where we can really have a sea change in the way people think about memory and the way people think about memory solutions. Just think about it. In a leading-edge data center environment today with a server working on high-value workloads, probably 75% of the components in that server are going to be either a NAND component or a DRAM component.
What better opportunity for Micron to engage with our customers in a different way, and what better opportunity for us to really deliver some differentiated value?
That's a great point. The 75% number, it's something that we've tested against the industry. Some of our analyst partners have told us it's a pretty accurate number. 75% of the server, when it's configured in this new scale-out model, is NAND and DRAM and one day soon will become even 3D XPoint. With that, there's an enormous amount of complexity that we have to solve because just making the server faster, you have to also build the rest of the infrastructure that releases all that capability, and we're going to be talking about that. Some of the legacy systems today, they operate fine, but as soon as you go to these web scale systems, you can now start to achieve performance numbers and scaling numbers and affordability numbers that are so different, you actually have new problems to solve.
You actually have new software things to solve, and we're just getting started.
Micron has built here in the storage business unit, I think, a really world-class team. Here in Austin, you're seeing part of that team. Micron is a global company, and we're a global company with the broadest set of technologies available anywhere in the world and with a broad product portfolio that goes with those technologies. What better partner to work with? I've been at this now for about 31 years, coming up on 32 this summer. Micron's got a long and storied history in the semiconductor industry and in the memory industry. We were formed in 1978. We started manufacturing in 1981. I joined the company in 1984, the same year we went public. As we moved through time, we really were very DRAM-focused for the first part of that following decade. Along the way, we introduced SRAM products and NOR products.
As we moved into the 2000s, in 2004, really a very significant change for Micron as now at this point, a global company really taking a huge chunk of our resources and moving them over to focus on NAND flash. We could see then what a big opportunity this was going to be in storage, but also in many mobile applications, et cetera. We took a lot of resources and dedicated them to driving the technology with first to implement self-aligned double patterning. We moved from an introduction into the market at 90 nanometers in a two gigabit device, which at the time seemed like a wonderful density and just an incredible amount of storage on a monolithic chip. Today, as we've migrated to 3D NAND, our leading-edge products are TLC 384 gigabits of data on a monolithic chip.
That'll store 30,000 high-resolution pictures, over 20 hours of HD video. Just a phenomenal change that we've been able to drive through Moore's Law. Now we start to think about moving to the next phase. Along the way, Micron's done a lot of other innovations besides the sort of Moore's Law and the manufacturing efficiency piece. We've picked up Lexar in the middle of 2005, I believe, which was an introduction into card solutions. We've delivered SSDs through Crucial into the consumer space and the client space. Most recently started delivering enterprise SSDs that hopefully you guys are going to hear more about today in SSDs optimized for the data center.
We've also taken this system-level approach into other areas and innovated with things like Hybrid Memory Cube that also help bring high-density DRAM close to a processor to help solve some of the bottlenecks that we see with big data applications. That solution, by the way, has a high-density DRAM stack with a custom ASIC that drives built-in self-repair, built-in self-test, high-speed interface to the processor. We introduced an Automata Processor that we're working with many applications for today to drive sort of the future of bringing that memory also closer to processing, something I think that will be value add for many in the room in years to come, not quite today. Most recently, just this year, 3D XPoint.
We introduced just another phenomenal innovation that will drive the future of storage and drive the way we think about how we get maximum data close to the processor to solve some of these big problems going forward. We've got a history of innovation. We continue to focus on that on a go-forward basis, we think you're just really starting to see the beginning of great things to come from Micron.
If you think about it, Micron is in kind of a unique position. Let me just put it in terms of the uniqueness of Micron, just for a couple of very basic, simple numbers. There are only four companies in the world that create NAND. There are only three companies that create DRAM. There are only two companies that do both NAND and DRAM, and there's only one company that does NAND, DRAM, and these emerging memories like 3D XPoint. We're in a pretty unique position to not only drive this differentiation, but be at the forefront of it, recognize the challenges, and help our customers execute to them.
All that work requires investment. Micron's not new to the memory business. We understand the cycles. We understand it takes capital investment to be a valued supplier to all of you in the room. We understand it also takes R&D investment to bring these new emerging memories to you, we've done that through the years and have a real history of success and innovation. We also understand that more and more, there's a piece on the end, which is a systems-level investment to really drive differentiated solutions. This Austin design center, or this Austin software center, is really just a small piece of that overall effort, but a very, very important one in terms of changing the way you guys interface with our memory and really opening up the opportunity to drive significant value for the customer.
Right. We've been in the NAND business for 12 years, and in that period of time, it has already changed a lot. We have seen the adoption of the client devices, the consumer devices, to the point where they're pretty well down this path. They've adopted it. We're down to the few remaining devices that have not, are basically the low-cost entry-level parts. We're seeing this integration of the value of SSDs or flash into these devices at the client and consumer level. What's happening now is you're starting to see this innovation occurring at the enterprise level. We're starting to have enterprise customers say business to business, we need this differentiation. That's what you heard on the video. Every one of those speakers was speaking from the perspective of an enterprise company. It's very important that we begin to address this.
There's some interesting facts. About 2.5 quintillion bits are produced every day. That's a billion billion bits produced every day. That sounds like a great deal for those of us who store things, but it creates a significant challenge for the IT departments because that number is about 4x the growth rate of the worldwide economy. Our bits are growing faster than the worldwide economy. The worldwide economy cannot afford to just keep doing things the way they've been doing and keep up. A change is required. You also see it's not just the capacity, the volume of all this data, it's also the velocity of the data. You see customers now looking at data that's coming off of other machines, the Internet of Things. Medical records are all being digitized. Video on social media, just copied phone photos from social media.
You're seeing an explosion of data. Most of that data has a life cycle to it, very short life cycle. As a matter of fact, we use this in Singapore. We are looking at big data solutions in Singapore. Mark?
Yeah. We're no different than many of the other verticals represented out here in the room today. We have just massive streams of data coming at us faster than we can analyze without some of these new storage techniques and really a collaboration between hardware partners and software partners, and our own manufacturing guys to really bring all this data together. Just think about a $6 or $10 billion fab running in Singapore today. We've got thousands of tools with hundreds of data streams, each delivering thousands of bits every millisecond. Just massive amounts of data coming at us. We've got to go and try and correlate that with inspection streams that are taking millions and millions of pixels, and electronic metrology streams that are driving millions of bits of information. It's just a very difficult problem.
All that data, it's useful for a very brief period of time. It's useful to you if you can use it to quickly catch a defect mechanism or to understand something that's changing in the environment and take action. We're doing that. We're doing that today in our manufacturing environment. It's driving faster yield ramps and higher quality. Just one of the applications, I'm sure everybody out there in the room is doing a similar thing.
Yes, this is what's driving this need to go to higher scale at a better affordability, higher scale at higher performance. Higher scale where you're not going to spend 3 months building the higher scale. You're going to build the higher scale in sometimes hours and days. We have a tremendous challenge in front of us to break down some of the barriers. This technology that was more frame-based arrays is now going to be going inside the server. Within those servers, we are now going to be running them at much higher speeds, doing data analytics at speeds we've never done before. The NAND and the DRAM and the 3D XPoint all support this capability. What we have to do now is we have to build out the rest of the infrastructure to support it as well.
With that, what I'd like to do is invite my next speaker up. My next speaker actually runs this site and is the individual who has built most of this software capability for us. I want to tell you, this was so important to us that when Micron hired me, he was the first individual I hired, and he has been working towards this day for all this time, almost two years now, and on his clock, over two on mine. It's very important. Steve Moyer is our vice president of engineering. He's got a PhD from University of Virginia. I have known him throughout many parts of my life, and he has been key in building parallel file systems, parallel computing systems. He's done his own startup. He's by definition, a key software innovator.
I want to invite Steve up to walk us through what he's learned.
All right. Well, thank you, Darren.
Thank you.
So-
Left my water.
Good morning and welcome.
You got two of them.
We are excited to have you with us here today. As Darren said, this event marks the official launch of the Micron Storage Solution Center. This is our hub for storage innovation, centrally located right here in Austin, Texas. I'm going to take the next few minutes to talk about the work that we're doing in our three focus areas here. They are customer workload optimization, partner solutioning, and storage software innovation. Before I jump into those, though, I want to take just a couple of minutes and talk about the site itself. Let me start with our selection of Austin. Why did we choose to put the Storage Solution Center here in Austin, Texas? There's a lot of decisions that factor into doing a site build-out. For us, there's really two primary things.
First, a key tenet in all the work that we do here is collaboration. It turns out, Austin is just really centrally located for us to be able to collaborate with many of our valued customers and partners. A second important factor in our decision is that, as Darren and Mark were talking about, it's becoming all about the systems. Austin certainly has a great tradition in systems companies. One of the great things, of course, that means that you have here is a great pool of talent with system software development capabilities. Of course, that matters a lot for our work, as you'll see. A few facts about this site. In its current configuration, it is 11,000 square feet. It gives us room for 65 team members and up to about 600 square feet of lab space.
As Darren mentioned, we are quite literally days away from beginning the second phase of our build-out, and that's going to take us to approximately 17,000 square feet, space for 85 team members, and nearly 2,000 square feet of lab space, which we're very excited about. Pretty much exactly where you all are sitting is, by midsummer, we'll have about 1,200 additional square feet of lab space on top of what we already have today. By end of summer, we should have, I don't know, upwards of 100 million IOPS running right there. All right. That's enough about the site. Let me talk about the work because that's why we're all here. The first focus area is customer workload optimization. What does that really mean?
That is a collaborative effort between the engineers here at this site, working with our enterprise sales team, working with our end customers, to bring their applications into our lab, to understand the workloads they generate, to understand what type of demands that places on the storage. We spend a lot of time turning all the knobs and dials on the application, on the operating system, on the underlying storage subsystem, understanding how all those moving parts work together. Then armed with that data, we can generate the reference architectures, the white papers, the sizing tools, really all the collateral that our customers need to be able to deploy that application with Micron storage and do it in a way that is most cost optimized in order to meet their requirements.
We do a lot of these studies here in this lab, and we're going to be doing a lot more of them as we expand. There's a common theme that we see emerge. Something you see fundamental to every one of these studies that we do, and that is when you take a workload, and you transition it from hard drives to SSDs, invariably, you see not only an increase in performance, but a decrease in footprint, a decrease in power, a decrease in cooling, really a decrease in total TCO. Really the common theme here is when you transition a workload from hard drive to SSD, you drive cost out of your hardware infrastructure. Like I said, we do a lot of studies here. That is something that we see consistently again and again and again.
One of the fun things, from an engineering standpoint, when you do these, is when you get those unexpected results. The thing that surprises you. When you see something that you weren't looking for that ultimately brings benefit to the customer. I want to give you just two quick examples that we had here. We've done and continue to do a lot of work with Hadoop. In this specific study, we were looking at deploying Hadoop storage nodes. If you're familiar with deploying Hadoop storage nodes, the HDFS software itself, you know that it is a best practice to do that on a physical server.
The reason you do that is if you were to attempt to virtualize the workload, to put more than one virtual storage node onto a single physical server, you drive up the randomness in the IO, the net effect is you drive down the performance for the individual storage nodes. You drive down the performance for the server node. It's just a lose-lose situation. On the other hand, the great thing you see when you transition that workload from hard drives to SSDs is it enables a whole new use case, a whole new deployment architecture. It becomes a best practice to deploy it virtualized. We get more performance per storage node. We drive up the aggregate performance of the server. Think about this. This is a huge win-win for the customer, right?
On the one hand, we've driven the cost out of the hardware infrastructure. What we constantly see with these studies. The other thing is now, because we've gone from a physical deployment to a virtualized deployment, we're also driving the cost out of the administration side. It's a real win-win. I'll give you a second quick example. Another thing that surprised us. It was a new use case. It was something that we weren't really looking for, and this came from the database world. As you imagine, a lot of our customers care a lot about database workloads. It's a common workload to move on to SSDs. In this particular study we were working with, I think it was Microsoft SQL Server 2014. Of course, we move that workload, you put it on SSDs, we see all the performance benefits.
We take out all the cost, all the goodness we expect to see. What's interesting is when you do that, now all of a sudden, you have a lot more IOPS capabilities on the node. We realized, well, hey, how can we put these to constructive use? What's the best way we could use these additional IOPS? A common way, if you're deploying a high availability solution, which a lot of customers want, is you'd set up two database servers. You connect them to a shared storage array. One fails, you got the data available to the other. Well, when you've got these extra IOPS available to you, it turns out really your most cost-effective and your simplest deployment is to remove that storage array, to set up replication between the two database servers.
You still achieve the same level of performance, but now you do it in a simpler deployment, and you do it without the cost and without the administration overhead of the array. Again, another real win-win for the customer. All right. That's customer workload optimization. Let me talk a little bit about our second area of focus here, and that is partner solutioning. Now, that is kind of a made-up word, but we actually do mean something very specific by that. It is a collaboration between Micron and our software partners and our platform partners to put together solutions fully optimized end-to-end for the customer workload. Now, one of the things that I want to emphasize here is that this is not intended to be just a packaging exercise, right? It's just not a marketing exercise, right?
This is really about meaningful engineering collaboration between the partners so that we can evolve our products together, right, so that we can enhance them together, so that together we can deliver a solution that is better for the customer than anything we could have delivered individually. I'm going to stop there on this because when Darren Endebrock comes up, he's going to, among his announcements, talk to you about some of the great outcomes we've had from this partner solutioning effort. The last topic I'm going to talk about, our third area of focus, is storage software innovation. Now, that is a pretty broad term, so let me narrow that for you. What we're really focused on here is fundamental storage software R&D to really unlock the full economic value of memory-based storage. Think about this, right?
The storage stacks that we have today evolved for 60 years around hard drives. Ask yourself this question: If we had actually started with non-volatile memory, would the storage stacks look like they do today? I'll give you the short answer, and that is no. I've spent 25 years, my entire career, crawling around inside of these things, and I can tell you that what we do throughout the storage stack today is not what you would do if you had actually started with non-volatile memory. In fact, it is an imperative that we take a step back as an industry. We really examine, reexamine, the storage stack from first principles. We're not the only ones saying that. If you follow what's going on in the storage research community, you'll see others saying the same thing.
You may be thinking, "Okay, Steve. Yep, I see that. I believe that. That makes sense, but why would Micron focus on that aspect of it? Isn't that kind of outside of your wheelhouse?" I'll argue, in fact, it's exactly the opposite, that it is a natural extension of our core competencies. Micron is a company that has nearly four decades of leadership in memory technology, a decade in SSDs. In fact, the storage software for us is really just the missing link between the application and the storage. In fact, it is a perfectly natural thing for us to want to innovate from the application to the media because every point along that path impacts how our customers experience our memory technology. To have this kind of change, obviously is a collaborative effort. It has to involve the ecosystem.
It has to integrate naturally. Again, collaboration is an important part of everything that we do here at the Micron Storage Solution Center. In fact, I'll give you just two more examples that are relevant to this area. You may not know this, but Micron has been, and continues to be, a sponsor of The Linux Foundation. In fact, we recently became a founding member of the Center for Research in Open Source Software at UC Santa Cruz, which is the birthplace of Ceph. All right. That's all I'm going to say about storage innovation today, though I certainly hope to be talking a lot more about it in the not-too-distant future.
With that, I'm going to invite up Darren Endebrock, who's going to not only make some great product announcements but also talk to you about some of the great outcomes from these focus areas. Thank you, Darren.
Thanks a lot, Steve. Appreciate it. Thanks, all, I have the great privilege of maybe one of the exciting parts of this, where I get to announce new things. That I think is really what it really comes down to. As a product person, I love to come out with new solutions and working. Before I do that, I did want to mention, this is Enterprising. Enterprising is a capability, and it's the future of how Micron is at the core of doing things different. You've heard a lot about modernization The need for accelerating how flash adoption, how compute, how merging storage and computing together and scale out architectures really matters. All the things are very true, and what it takes is a site like this with engineers like Steve, who are collaborating with the partners in the world.
Great names like VMware, like Red Hat, like Microsoft, the kind of partners that hopefully are out there watching us live on the video stream and are here in the room. Teams that are really coming together to change how the industry is enabled. Not only that, Enterprising is bringing together and driving new things happening. Next door, I don't know if you know it, but we have the SNIA board members' meeting on their quarterly business meeting. We're participants on the board. We have Rob Peglar sitting with those teams on the SNIA teams, hosting them. That's part of what Enterprising is. Enterprising is even more. It's taking great technologies, and it's putting them together. Let's get into that. We talk a lot about our 3D technology, and I know one of the things as I move through the slides here a little bit.
One of the really interesting things, and I hear a lot of questions about is, when are you taking 3D to that next level? One of our 2016 big priorities is taking what I think is the best 3D NAND technology in the industry, using floating gate. It's unique. It's targeted at the kind of workloads that we're talking about and driving that through our portfolio. I thought it'd be appropriate to at least start out and say, let me tell you what this year looks like. Let me announce the portfolios that are going to use our 3D NAND technologies and when they're at. Let me start it off with our first one in April. Well, if I check my watch, it is April, so we are right on the cusp of what we call our MX300. This is a consumer SATA drive.
We're just launching it. We're extremely excited about it. It's using our 3D NAND and TLC. It's really targeted at customers who are enthusiasts, who want to upgrade their systems. It's going to be a very powerful solution for us. The great part about our 3D TLC, starting to get a little noise there, is that it leaves right off where our MLC solutions go. We're saturating the SATA bus, giving high performance at the right economics, that real central place where price and affordability come together. Very excited about this. Followed really quickly by our client OEM solution. We're announcing our 1100 and our 2100 series. These are client OEM drives, both SATA and PCIe. This will be one of our first client PCIe drives that we're announcing. These are just amazing products that are going to enable the next generation of thin platforms.
They're going to enable power savings, longer battery life, higher performance. Take some of our technologies that are unique to Micron, things like Dynamic Write Acceleration, that allow our systems today to boot quicker, to go into power save mode more quickly, to drive the applications that the client world demands. Very excited about these products, enabling both M.2 form factors that are single-sided up to a terabyte, so high capacity and extremely small form factors, and PCIe performance, which is moving towards that next generation for the client world. Very excited about those. Following on into the second half of the year, what you're going to see us do is roll out 3D NAND across our portfolio.
You'll see that coming out in technologies that have extremely low latency for the workloads today, have performance that saturates certainly SATA buses and then moving on to give higher performance in the PCIe space. Really exciting offerings that are coming out through the rest of the year. I know there's always been lots of questions about when will we see in the real world Micron's technology, I'm here to tell you it's here now starting in April, you'll see it rolling aggressively throughout our portfolio, and it's very exciting to us. 3D NAND technology isn't the only exciting things happening. We see a big transition from SATA into the PCIe space.
Another part of an environment in Enterprising is actually taking what customers need and matching our technology base that we have, we have access to a lot of great technologies, putting that together in new and innovative ways. One of those has been our participation in NVMe and the PCIe standards. We see a pretty rapid adoption, especially in the enterprise, starting to take hold where PCIe and NVMe technologies, we've been part of those consortiums for many years. They're really moving forward. I wanted to announce our enterprise NVMe solution set. I've got two different families here. I've got a 9100 family. I say family, it's many different types of drives. It suits read-centric solutions. It's got mixed use. It's got write-centric. It gives you a very broad breadth of performance and capability.
The 9100 is really geared at enterprise customers. It comes in half height, half length. It's also in two and a half inch. This solution is really targeted at general purpose, what I guess we would say has been the traditional market for PCIe in terms of its high performance. These drives run at 750,000 IOPS and extremely high megabytes per second. Very exciting technologies. Now let me compare that a little bit with the 7100 series, because that's part of what we've learned in this solution and part of also what Enterprising is understanding what's next and how do you further drive out costs for your customers? How do you take it to the next level? We've been working a lot with hyperscale customers, the biggest of the big, the scale-out customers, they say, "You know what?
What I don't need is one SSD that runs at 1 million IOPS. What I need is a lower power profile, extremely low latency, form factors that allow me to put more SSDs in my platform as I move forward." That's where the 7100 comes in. It's becoming more of an emerging space, but it's focused at lower power, where a traditional PCIe slot take 25 watts of power. Well, the 7100 series is half that. It runs at up to 12 watts of power. It comes in M.2 form factors, as well as two and a half. It's really meant for now we're seeing this more scale-out architectures and how they play forward. Very exciting part of filling out our enterprise portfolio, again, part of Enterprising is having the right tools that go with it.
An interesting one is PCIe and SATA isn't the only thing in the world, there's innovation, and innovation comes in a lot of different ways. One way you innovate, and one way Micron's innovated is unique and interesting partnerships. We partnered with another industry giant in the world of storage, that was Seagate. We came out with our drive, it's the S600 series. This series is a product of SAS drives and basically here to announce, those have been out shipping. We're in volume shipments. Very exciting. When you look at both our PCIe drives as well as our SAS, what I think the best part is you're starting to see the customers come towards them. Actually failed to mention on that last slide, we've got customers using our PCIe products as we speak, a company called Distillery. Distillery's exciting.
It's one of those companies that couldn't exist if flash and these new modern architectures didn't exist, you'll be hearing more from Amit later as he comes up with Mark Glasgow. Basically what they are is taking market intelligence and customer interactions, they're doing that at a scale and speed that's never been seen before. Something like 50 billion web transactions a day flow through their systems, all enabled through Micron technologies and some of these new modern architectures. Another great customer example that is using our drives, they're here and you can see them as well, they're a company called Symbolic IO. Symbolic IO is a software-defined storage company.
They're using both Micron NVDIMMs, so they're taking advantage of putting more computation and storage in memory, as well as our 9100 series SSDs to really get unheard of efficiencies of computation and move that closer to storage in a way that hasn't been seen. They do that through various schemes of data reduction. It's very exciting, I'd urge you to go look at their booth as well. On this slide, SAS, what we've seen is there's a lot of customers who aren't looking necessarily to ready to modernize, but they need to extend what they have, that's where we complete our portfolio with the SAS, where we have the Newberry County Memorial, this hospital set up. They've taken our SAS products, and we also have them in several other OEMs.
What they're doing is they're extending the number of remote connections that their doctors can do. They serve a broad population, they have doctors call in. While that's great for a hospital, they get to have more minutes and scale their resources so their doctors get more time with these patients. Even more than that, they get better treatment. They're able to provide better care. That's somewhat what the end result of all of this that we're doing. Now, having a great portfolio of technologies is wonderful. Filling out our portfolio of enterprise SSD is something you would expect from us, I hope. Very exciting and interesting to us. What we also wanted to talk about today is how do we take all of this that we've discussed with you, all these trends, we take it to the next level.
It's interesting, we talk about how do we partner together and collaborate. I'm here today, and one of my great pleasures is to announce the Micron Accelerated Solutions. This is our first entry into providing a complete system-level solution to the industry that is both collaborative, it includes software, it includes the server platform, as well as our memory technologies. Putting these all together in a way that makes and simplifies the activities of modernizing our infrastructure. We talked a lot about data growth being unrelenting, and you see that in reports. It's still the biggest thing storage companies have to deal with. I would put to you that data growth is almost a legacy answer of the storage community itself because it's the tail wagging the dog. It's not the growth that was so important, it's how you deal with that growth.
We talked about it in our own data centers. Mark and Darren referenced this. The data comes at you so quickly, and you have to do something with it so quickly to have it become a competitive advantage. Companies like Distillery that we talked about, these are the companies on the forefront of that. Moving forward, there isn't a business that I can think of that hasn't taken data analytics and the speed of data and transforming information into real decisions and using that as a competitive weapon. To do that is a complex world. To do that takes a more software-defined approach. It breaks the rigid barriers of what is storage today. Very excited about these solutions. Now let me talk to you. This is a capability that we have. Steve kind of mentioned what we can do in this site.
Our intent is to partner much broader than this. We're announcing three amazing solutions today that you can go out and purchase, or you can use our reference architecture. This is where it's hard for a person like me who's focused so much on marketing, we've actually removed the marketing hype from these solutions. What we've done is had our engineers search out the limits, the borders, and find the nexus of where platforms and solutions can make things affordable, have the right performance, have the right stability, the right durability, really bring these things together. Actually, even probably the most important one, the one that I think is most important to a lot of people, is predictability. How these things can be predictable solutions. Let me talk to you about a couple of those because they're very exciting.
I want to talk to you about all three we're announcing, in fact. Let me start off with our VMware vSAN solution. We're launching a ready node solution. This is a of- Four, six, and eight. Those are the VMware configurations, and we're doing that with our great partner, Supermicro and VMware. We've engineered these solutions to be just outstanding. When you think about what does it take to deploy this, a lot of people have a virtualized environment. They're looking for help to take their virtual investment and move it beyond what external storage does today. That's where vSAN comes in, which is an exciting technology, especially with the 6.2 release of VMware vSAN. When we put this together, what you find is we compared this against a hybrid solution, traditional knowledge.
Steve talked a lot about how do you learn things in this industry, how do you move the ball forward in modernization. One of the ways you do that is you take traditional understanding of a solution. You take SSDs as your cache drive in a vSAN, and spinning media as your storage. Then you compare it against what you can do. What we found is that actually, by using the right mix of SATA SSDs for affordability, taking a cache layer, in fact, we had to create a new cache layer that optimized and tuned our SSDs specifically for vSAN, and that's available through this solution as well. What we've done is we found we could be 40 times the IOPS of a more traditional hybrid configuration of vSAN. That's a read excuse me, that's a read view of that today.
Even it's 15x in a mixed-use workload. Really, you can see that SSDs and an all-flash array vSAN really moves the needle on it. That's not it. It's not just about the performance, it's finding that nexus of affordability. We can do that at $0.44 a gig, that's usable capacity based on this solution. You're going to be able to see that after. You get to go to a petting zoo on all of the things I'll talk about. You can get more information. You can see these in action. I think you'll be pretty impressed with it. It's very exciting. That sort of moves you forward on the VMware side of things, and that really helps a lot of customers change the game. The other side is big data and Ceph solutions.
We spend a lot of time talking to banking and finance, to media and entertainment, through these customer interactions, what we've discovered is Ceph's extremely popular in the emerging space for storage, most people think of it as a data tub environment. I want a lot of capacity, how do I store it? Well, that makes sense, in a world today which is requiring solutions that move much faster, we actually had a cry from these customers to say, "I need a performance variant of Ceph as I move forward." We're announcing our second Micron Accelerated Solution, which is around Ceph and Supermicro. Again, a complete configuration. You can order this today. What we found is with all-flash array, you can reach 1 million IOPS, something that's unheard of in the Ceph community today.
Very exciting, if you want to think about 140 gigabytes per second, really think about that's 28,000 high-definition streaming videos concurrently coming off of a platform. That's the kind of capacity, that's the kind of performance, that's the kind of solution that our customers are striving for. Very excited about our Ceph solutions. The last one is our NexentaEdge solution. We're really excited. You think about a lot of virtualized environment, I've got this big data world. Well, now also I've got software defined, the world's definitely everybody understands software defined and its impact on the industry. Well, working with a leading company, you think about somebody who's been in software defined for a long time, Nexenta is that company. They've been there, they've seen it, they're kind of riding the wave of what this future looks like.
We're excited to be working with them on their NexentaEdge solution. It's a scale-out architecture, really great for OpenStack solutions. It's object, it's block, it's whatever you need it to be. Right now, as we're talking, we have engineers working and collaborating together. We're going to have our solutions coming out towards the second half for this exciting offer. Really, this is the first of many. I talk about this capability as a platform. This is a platform for us to launch and move the industry together with companies like Supermicro, VMware, with Nexenta, with open source technologies like Ceph and Hadoop and others. It's very exciting for us. Our whole focus is how do we take out the risk? How do we move this industry forward? Out of this, we've launched sort of our SOLID Ready seal.
It's about how do we take out the risk and complexity. These are fully referenced, architected solutions. You can order them. They're complete recipes on how to deliver the performance and the optimization that we've talked about. They come, in some cases, with optimized SSDs, because part of being an SSD vendor moving forward is how do I make the right products that actually impact the solutions the customers use? The other one is it's about ease of installation. I mentioned you can order these completely built, fully serviced, predictable, ready to go. That's an important part, but if you're a build-it-your-own person or if you've got an environment you want to put together yourself, we do have the complete reference architectures ready to go as part of that. All of that's very exciting.
Lastly, I think maybe the most important part is it's not the marketing fluff. It's how do you make it predictable? How do you make your outcome equal what you were promised up front? That's what we're doing here working with these companies. Very excited about these solutions. As you think about what this announcement really holds, you can see why it's so important that Micron work with software vendors, with platform vendors to move this machine forward and come out with those. You see today, we've talked about everything from DRAM to NVDIMMs to SSDs. The reality is this work we're doing here paves the way for things like emerging memory and 3D XPoint, and it paves the way for some of the software technologies that we talked about as well. Much more to come.
I think this is the tip of the iceberg. But to talk a little bit more about the emerging memory part, I'd like to welcome to the stage Jon Carter, who's going to share a little bit about XPoint.
Thank you, Darren. It's definitely exciting times for Micron. I want to be real clear. You've met a lot of our leadership team today. We're driving a lot of solutions forward. We're changing Micron, and we're changing the industry. 3D XPoint is at the heart of that. After 10 years of development, I'd say bold development by Micron, we're turning the page on development and entering the market. I'm working specifically with a focus team on enablement of the ecosystem. It's fun. It's exciting. We're developing a captive market. The next few years are going to be huge. It's huge. Before we go any further, I just want to do a quick review. This is 3D XPoint. This is it. 1,000 times the performance of NAND, 1,000 times the endurance of NAND, and 10 times the density of NAND. This is game-changing.
I'm having some very interesting discussions as we move forward with our partners. We've been working the last year since we've announced the technology with a unique set of partners that actually are market makers, who are the infrastructure leaders of the world, who understand storage. They understand demanding workloads. To set the expectation of where we will enter the market, it's in workloads where milliseconds matter. I talk about the performance spectrum, the latency profile, 1,000 times the performance of NAND. Internalize that for a minute. We're going from batch to real-time. It's extremely game-changing. This is where reliability matters, where consistency at scale matters. This is at the heart of 3D XPoint. I guess to give you some insight, there's been a lot of questions. We had a great dinner last night, talked to a lot of analysts, a lot of burning questions.
I'm here to give you some insight to some. We're going to be giving you installments over the next year of where we're going and who we're working with. It's truly exciting. To give you some insight now to the ecosystem. I get a lot of questions around ecosystem development. What are we doing? How are we going to get this in the market? The first storage class memory, is it real? It's real. You see it here. What we've been doing over the last year is working with customers, the end customers, who understand the workloads that I've described to you, where milliseconds matter, where reliability and scale and concert matter, and they're demanding. We've worked with them. We're working with the number one, the top, I should say, infrastructure providers who understand these workloads.
They understand the development cycles that it takes to put a solution into the market. These aren't consumer devices initially. These are ones that when they come in, they have to work. They can't fail. It's uncompromising. This is exactly what 3D XPoint was developed for, so it takes time. Over the course of this year, as you see us ramp the technology and expose it to the industry through our partner with Intel, you'll start to see Micron driving differentiated solutions to take the most value out of this media and present it to the end user in 2017 and 2018, where we start to hit scale. It's extremely exciting. I'm happy to be part of this.
There's a lot of other questions I'm sure that you have, but at this time, this is what the first installment of what you're going to see over the next year as we start to get into branding, we start to expose some of our partners over time. It's extremely exciting. I guess for 3D XPoint, I'd like to leave it at that for today and bring up Mark Glasgow, our Vice President of Sales. Thank you.
Thank you, Jon. Thanks. Endebrock . All right.
Wow
exciting. My turn on? There we go. Super excited. Thank you all very much for coming out. This is a big day for Micron, many more announcements to come that we're super excited about in the future as the company continues to evolve and to address some of the more demanding challenges in the market today. I have the distinct honor of running a customer panel, so you don't take Micron's word for it. You can hear from the voice of the customer about how they're using some of our technologies to drive some of these high-performance workloads and to drive out the cost in their data center, all the while increasing and improving performance. That's not an easy task, we're super excited to have the panelists.
Before I introduce them, I just talk about something that I've observed in the last couple of decades as a career salesman. I started selling storage in 1992, since then, I have been on countless thousands of sales calls. As I think back across that career, there is a common thread through every single customer that I've been to. I can't think of one that doesn't have this imperative that it seems like everybody has. It doesn't matter the industry, the industry vertical, it doesn't matter the size of the company, it doesn't matter the workload that they're trying to solve for. Every single customer I've met with has this demand upon them, that is this requirement, this imperative to drive costs out of the data center.
In many cases, they got to do that and also improve performance, and they do it with a smaller budget than they had last year. This idea of having to do more with less is very real and very tough, and Micron stands ready to help companies of all sizes and all industries to take care and take advantage of these things. If you look across the last three or four or five decades, this has been in play in front of us the whole entire time. The 1970s, we had mainframes. Okay. The '80s came along, and we had a couple of guys like Scott McNealy and Casey Powell that really helped spur the growth of open systems. Basically, the power of a mainframe at one-tenth the cost, again, driving out efficiency in the data center.
In the '90s, we had the ERP boom and people getting ready for Y2K. Again, ERP introducing new levels of efficiencies in companies that had not been there before. As we moved into the 2000s, we saw this thing called virtualization take place, not only at the storage layer but also at the server layer. Okay. Diane Greene and a bunch of super smart engineers get together, and they go out and create this thing called server virtualization and try to somehow get their arms around this huge problem at the time, which was bare metal server sprawl. Okay. Huge inefficiencies in the data center. We saw how that changed our markets, and again, drove billion-dollar market caps and brand-new companies as these companies invented new ways to do things better, faster, cheaper.
Here we are in the 2010s, and we have this thing called cloud that's upon us. It really is amazing when you have companies that can go out and create their own private cloud, and they can burst as they need to the public cloud, again, driving efficiencies. I remember selling systems to companies that literally had to plan for the long pole in the tent, which was the holiday season, and then those systems would run at 30% capacity utilization for the rest of the year. Incredibly inefficient. Now we've solved that with hybrid cloud. These things are in play.
There's one more big, I shouldn't say one more, the current big trend that we see, again, on this theme of driving costs out of the data center, is something that Gartner said a couple of Gosh, I think this is about a year and a half ago. They predicted by the year 2017, 50% of the global data centers will be running some form of web scale architecture. Okay. If you think about the global data centers, those guys represent way more than 50% of the global IT spend. This is a massive trend. This is taking storage from frame-based arrays, typically that were built with spinning media, okay, again, very expensive, into server-attached storage, getting the data closer to the CPU, reducing costs, increasing performance, and again, this cycle continues.
Micron, I believe, is uniquely positioned to take advantage of this trend of not only increasing performance, but doing it while you're driving costs out of the data center. We're one of the few companies that has the technologies that are absolutely required to do this, as we see companies move oceans of spinning media offline, but they've got to store those bits somewhere. They're going to go to flash and some of the technologies that we've invented. We talk about many times at Micron, the four, three, two, one of Micron. There's only four makers of NAND flash on planet Earth. We are one of them. There are only three makers of DRAM on planet Earth, and we are one of them. There's only two companies on planet Earth that do both DRAM and NAND, and we are one of them.
Mark my words, as we go through time, that dynamic between volatile and non-volatile memories is going to continue to be more and more important as these workloads get more complex. We're super excited about the positioning that we have as a company at being one of two companies on the planet that can do this. We're the only company on the planet that has all three, 3D XPoint, NAND, and DRAM, and that's going to allow us to do some things that nobody else can do. That excitement is why we're here today. Without any further delay, let me go ahead and introduce our panelists. The first panelist I'd like to introduce is Amit Gupta. Please give it up. Amit is the CIO from a company by the name of Distillery. We'll have you introduce your company here in just a second.
Our next panelist is Duleep Wikramanayake, currently with Redapt. Duleep also has a very colorful and interesting past. We'll talk about that in a second. Our last panelist is Michael Peterson. Mike, there you are. Mike's with Netflix. Come on up. Mike is with Netflix, and as you all know, they are to Blockbuster and Hollywood Video as Uber is to taxis. They have reinvented some business models, again, using some very interesting technologies. We'll have him say a little bit about that. Guys, with that, let's go ahead. If you want a chair, there's three here. Mike, why don't we start with you? Take two or three minutes. We all understand, in fact, I think everybody here is a member of Netflix.
Tell us a little bit about the inner workings, what you're doing behind the scenes that allow you guys to do what you do at the scale that you do it at and still drive out costs.
Just real quickly, what we do is we make appliances that we host inside of other people's networks, a part of that is to drive some efficiency in how we deliver video over a video CDN. We do that, originally, we started with spinning disks, we've moved more and more into flash-based technologies to drive efficiency for that. We've started with SSDs then moved on to NVMe recently.
You said something to me earlier today that I thought was interesting. You care less about IOPS. They're important, but it's not your single concern. What's your biggest concern?
Our biggest concern is bus bandwidth inside of a server. Filling, originally, the 10-gig port was fine with spinning drives. We moved on to flash to try and get to around 40, moving beyond 40, we've had to move from SSDs to try and figure out how to use NVMe to do that.
You have to worry about bus speeds, latency, IOPS, driving out costs, maintaining the efficiencies. Huge job on your shoulders, thank you for that. Duleep Wikramanayake. I'm going to brag about Dil for a second. I've known Dil for a while. He's CIO. He was formerly CIO at Ancestry.com, before that, he was at Intuit, before that, he was at LexisNexis. In every single one of those companies, he took those legacy architectures chock-full with spinning media frame-based arrays. Say that 10 times fast. He had the insights to know that the right direction for those companies was to go to a server-attached storage model, improve efficiencies, drive out costs, reduce power and cooling, reduce the real estate footprint. Dil did this three different times at three different companies, Lexis, Intuit, and Ancestry.com. Did I just steal all your thunder?
Well, you did a great job. Turned out for my mark there, right? Yeah, I think the focus on innovation and modernizing infrastructures has been a consistent theme. Driving out cost, delivering agility, performance, and scalability really forced us into a thinking model where we had to basically rip all of our thinking apart and go down to pure server, memory, disk, and then figure out how that was going to evolve. Basically, we came up with make sure we have open architectures, make sure we're doing open source, and make sure we're driving scale and performance out of our systems. That's what drives us into technologies like flash and SSD, SATA, SAS, and all of the faster technologies that can then scale out horizontally.
Okay, great. Thank you for that. Amit, a quick intro on Distillery. I think it's one of the coolest business models out there, and I think somebody said it earlier, Distillery couldn't have existed 15 years ago with the technologies that existed back then. They are here today with some really interesting stuff. Please share with us.
We are a big data digital intelligence-
Make sure that is turned on. Here we go.
We are a big data digital intelligence company. We use massive data sets, proprietary technology, and human intuition to help our partners find new customers and achieve their marketing objectives. We do that by, excuse me, ingesting billions of web transactions a day, tracking hundreds and millions of devices every day, storing their history in a data store.
Excellent. Amit, thank you for that. All right, I'm going to lob up a question. Any one of the panelists can take it. As we've seen NAND flash prices drop fairly significantly over the recent years, how did that change things? Tell me about how that contributed to your driving the cost out of the data center.
Well, I think there's two factors. One is the driver factor, which is the types of workloads that we all have to deal with. Ancestry.com, as an example, two years ago, started a DNA sequencing process business. They sell a DNA test. With millions of DNA samples times 700,000 data points per sample, you have to drastically drive out cost of your infrastructure when you're trying to handle a trillion data points, and you don't want to have that all in your data center. Building out our own hybrid solution, with basically plain vanilla hardware that's loaded up with the right types of flash and SSD-type hardware, really helps our performance.
I think one thing I'll punch back to you on is, since we have to burst to the cloud in order to satisfy those workloads, how do you guys at Micron work with the big cloud providers?
Yeah. I appreciate that. That's a nice softball pitch. I will take that one. For years and years and years, when you flew into JFK, and you pulled out, there was this big, bright orange sign that said, "Eight out of 10 of the largest banks in the world all run on Oracle." Larry Ellison, some great marketing. That sign was there for a very long time, and I do not think it is there anymore. We could put up a similar sign that says, "Eight out of eight of the largest hyperscale companies on the planet all run on Micron." We do work with all of them in both DRAM and NAND. We see some pretty interesting. These guys are on that cutting edge of really driving costs out of the data center in a freakish way.
Their scale is massive. They could not do what they do today with running on a really expensive frame-based array, with all the maintenance fees and just all the many cost that come with that. Not to mention the real estate in the data center and the heating and cooling or the power and cooling. It is something they could not do. We get in conversations with them on a very interesting level from an engineering perspective.
Well, I would guess that that is one of the reasons why their cost models can be so low to make it attractive to folks like us and Netflix and others to use their technologies when we do not want to support the capital to build out a whole bunch of infrastructure and then only use it 30% of the year or something.
Exactly. No, that is great. Mike, how long have you been with Netflix?
For about four and a half years.
Okay. You've seen some pretty interesting times in just the last five-year span. Could you guys be doing what you're doing today? Could you be operating at the scale you are with just spinning media?
Yeah, no, we couldn't get to the scale we were today with just spinning media. Like I said, moving to SSDs help us grow to the point we are today, and we're definitely going to be using NVMe solutions going forward to help us scale even further.
Tell me a little bit about workloads. I get the feeling as I have these dozens and dozens of sales calls that I attend on a yearly basis, that the workloads are changing in a fairly significant way, and I think they'll change yet again when 3D XPoint comes out, and some things that we haven't even contemplated will now become available because of the new capabilities of that particular memory medium. Tell me about the workloads. Are they changing? Is that a wrong sense, Amit?
Yeah, I mean, it's increasing every day. We are adding, like I talked about this universal data store we manage. It gets hundreds of millions of devices every day, and every device could have hundreds of data points. Every day they do something different, the new data point comes in, hundreds of those data points per device coming in. We are in the business of RTB or real-time bidding, where we bid or participate in auctions of advertising inventory.
When we do that, on a large scale, I'm talking about more than 50 billion requests coming in a day, this data store, which has billions of data points, we get the request on the fly, it's a real-time access for that random device out of those billions of data points and make a decision on it. Generally, we get 100 milliseconds to respond to within. Majority of that 100 milliseconds goes to network travel time back and forth. We are generally left with 10, maybe 12 milliseconds to access that device history, make a decision on that device and bid or not bid. That directly impacts our revenue because if we don't bid, if we lose out on that opportunity, and we lose that, we don't generate revenue. Hits our top line, bottom line.
The shelf life-
Yeah
for those opportunities are like that.
Yeah.
Yeah.
Yeah.
Wow.
I think another point has to do with machine learning.
where businesses like Ancestry's and others have to accumulate an enormous amount of real-time information in order to understand how their systems are performing, how their customers are interacting with their systems, and the machine learning data requirements that go along with that require both space and performance. As those needs blow up, basically, we have to continually drive horizontally and then drive the cost per unit down. Otherwise, we wouldn't be able to provide those sorts of analytic services to our own businesses.
Dil, while you're on that, I wanted to ask you a question. When you started this at LexisNexis, that was the first time you went from a legacy architecture.
Yeah
call a web-scale IT.
These are generational, right? In LexisNexis, we're talking about two jobs ago for me.
Right.
There was moving from to distributed systems and continually driving both the process, the skills, the technologies into a more open and generic standpoint. Further on to Intuit, the same thing. Really optimize data centers, close data centers as needed, optimize their performance of systems by virtualizing and building an internal cloud, going to the lowest common denominator-type hardware that could perform very well. Then it sort of all came together at Ancestry, where so much of the SSD, SATA, SAS technologies and flash became available. That became another iteration of then driving the cost down, going further horizontal, using public clouds. It really was a game changer in terms of the last five years in terms of making all those technologies come together.
As the costs come down and the performance capabilities change and improve, it opens up more and more doors, and you continue to see that. Let me ask you this, and perhaps, Mike, you can take this. In the context that is Netflix, what do you see as the next big hill to climb for you guys?
Probably the next big hill to climb for us is to try and get to delivering 100 gigs out of a single server. That's a challenge that we're going to work on going forward. That drives a lot of scale and economics for us, and so being able to use solid-state technology to do that is going to be the big thing that we're looking for.
100 gigs out of a single server. Exceedingly dense.
Yeah.
Right?
Yeah. It's what we're aspiring for.
That's awesome. Outstanding. Dil and Amit, what about you guys? What do you see as the next big opportunity?
I think the biggest challenge and opportunity for all of the technologists that are in businesses is to continually to remain open so that they are not preventing themselves from taking up new technologies. The tendency is to close or to decide on something, and then you get locked in. You really don't want to do that. You want to maintain open architectures so that you can consistently evolve to the next level. 10 years ago, the lifespan of technologies was five to 10 years. Five years ago, it was about three years. Now, we did OpenStack two years ago, and last year, we did Docker.
Right.
It's like you have an 18-month life cycle to ingest and then roll out new technologies. The same is true at the hardware level, where you guys are inventing new things every eight, 10, 15 months. You're ready to deploy them. Customers like us have to be ready to absorb them, use them, and then we can stay in sync with the benefits. You want to get the benefits as soon as you possibly can.
Hopefully not make decisions today that accidentally paint yourself into a corner in the next 10 or 12 months.
Absolutely.
Interesting. An unenviable position to be in for sure. Amit, 50 billion requests a day, I assume that's going to continue to grow for you guys.
Yeah. This year's goal is to be able to handle 75 billion-plus a day. That's where we're working for, and we are using the latest 9100 and whatever else you guys are producing to drive that growth.
Yeah. Great. I'm glad you brought that up. We see the progression from SAS to SATA into NVMe now. Do you guys see NVMe in your futures? I think we talked a little bit about that.
Yes, we do. I think a part of it is that at this time, it's a little on the pricey side for us to scale out in a big way with it. As the 3D NAND technologies come in and that price comes down, we're definitely going to be interested in using it.
We've used all three of those disk and disk-like technologies in our own homegrown open-source software solution, storage solutions. I think whatever comes next, that's why we have to remain open, so that we can bring in that next level. Maybe there's a four-tiered storage solution. Three, and we drop the bottom one out, and it goes into a legacy mode. I think as long as we maintain ourselves as software-driven at our level, then we can employ whatever hardware as soon as it's available.
See now, with your new position as Chief Strategy Officer at Redapt, I have to think that you're seeing many, many more examples of this, and obviously you're imparting, I think, the lessons you've learned doing this several times now at different companies.
Yeah. I think many companies are looking at modernization, I call it, because it's whatever point you're at, you want to start moving from there toward the current state and then toward the future. People are in various states of having legacy, right? Whether it was last year's legacy or 20 years ago legacy.
Right.
You want to move off that. We do see a lot of customers that want to move off the dime and get into the newer technologies, and they use companies like Redapt to help advise them, consult with them, help with engineering and hardware solutions, which you guys, of course, are part of.
Yeah. Outstanding. Any parting shots before we wrap up? Any last-minute thoughts?
Keep doing new things.
We're trying. We definitely are. If we've proven anything over 37 years is that we are not sitting idle. We are thought leaders. Everything that we do is stuff that we've invented, and it's our own IP, and you can absolutely bet that you'll continue to see more cutting-edge thought leadership that'll allow folks like these to drive solutions, drive costs out of the data center while improving performance. Gentlemen, thank you very much for your time. I appreciate it. Next up, I will call Darren Thomas back up to the stage. Darren, take it away.
Thank you, sir.
Thanks.
I think we've done a great job of demonstrating that Micron has a rich history, a rich legacy of invention and developing technology and moving into that technology kind of at a pace faster than our competitors. What you're seeing now is we're not just creating the memory technologies, but we're addressing that next layer of technology that's got to get solved. You heard the customers talking about they're able to do things today they couldn't do before. To a great degree, it's because we have now reached out, Micron has reached out at a community level, and we are now reaching out to our partners. We are now reaching out to software companies. We're reaching out to some companies that you might even think are competitors to us. We intend to solve this problem so that our customers get the products they need.
When we do that, this market will grow. With that, what I want to do, we've been doing a webcast all along. I want to thank the webcast invitees for being on for this almost hour and a half. If you've made it to this end, you've gotten to see a great transformation of a great company. With that, we can end the webcast.