(upbeat music) >> In 2010 Wikibon predicted that the all flash data center was coming. The forecast at the time was that flash memory consumer volumes, would drive prices of enterprise flash down faster than those of high spin speed, hard disks. And by mid decade, buyers would opt for flash over 15K HDD for virtually all active data. That call was pretty much dead on and the percentage of flash in the data center continues to accelerate faster than that, of spinning disk. Now, the analyst that made this forecast was David FLoyer and he's with me today, along with Renen Hallak who is the founder and CEO of Vast Data. And they're going to discuss these trends and what it means for the future of data and the data center. Gentlemen, welcome to the program. Thanks for coming on. >> Great to be here. >> Thank you for having me. >> You're very welcome. Now David, let's start with you. You've been looking at this for over a decade and you know, frankly, your predictions have caused some friction, in the marketplace but where do you see things today? >> Well, what I was forecasting was based on the fact that the key driver in any technology is volume, volume reduces the cost over time and the volume comes from the consumers. So flash has been driven over the years by initially by the iPod in 2006 the Nano where Steve Jobs did a great job with Samsung and introducing large volumes of flash. And then the iPhone in 2008. And since then, all of mobile has been flash and mobile has been taking in a greater and greater percentage share. To begin with the PC dropped. But now the PCs are over 90% are using flash when there delivered. So flash has taken over the consumer market, very aggressively and that has driven down the cost of flash much much faster than the declining market of HDD. >> Okay and now, so Renen I wonder if we could come to you, we've got I want you to talk about the innovations that you're doing, but before we get there, talk about why you started Vast. >> Sure, so it was five years ago and it was basically the kill of the hard drive. I think what David is saying resonates very, very well. In fact, if you look at our original presentation for Vast Data. It showed flash and tape. There was no hard drive in the middle. And we said 10 years from now, and this was five years ago. So even the dates match up pretty well. We're not going to have hard drives anymore. Any piece of information that needs to be accessible at all will be on flash and anything that is dormant and never gets read will be on tape. >> So, okay. So we're entering this kind of new phase now, with which is being driven by QLC. David maybe you could give us a quick what is QLC? Just give us a bumper sticker there. >> There's 3D NAND, which is the thing that's growing, very very fast and it's growing on several dimensions. One dimension is the number of layers. Another dimension is the size of each of those pieces. And the third dimension is the number of bits which a QLC is five bits per cell. So those three dimensions have all been improving. And the result of that is that on a wafer of, that you create, more and more data can be stored on the whole wafer on the chip that comes from that wafer. And so QLC is the latest, set of 3D NAND flash NAND flash. That's coming off the lines at the moment. >> Okay, so my understanding is that there's new architectures that are entering the data center space, that could take advantage of QLC enter Vast. Someone said they've rented this, a nice set up for you and maybe before we get into the architecture, can you talk a little bit more about the company? I mean, maybe not everybody's familiar with with Vast, you share why you started it but what can you tell us about the business performance and any metrics you can share would be great? >> Sure, so the company as I said is five years old, about 170, 180 people today. We started selling product just around two years ago and have just hit $150 million in run rate. That's with eight sales people. And so, as you can imagine, there's a lot of demand for flash all the way down the stack in the way that David predicted. >> Wow, okay. So you got pretty comfortable. I think you've got product market fit, right? And now you're going to scale. I would imagine you're going to go after escape velocity and you're going to build your moat. Now part of that, I mean a lot of that is product, right? Product is sales. Those are the cool two golden pillars, but, and David when you think back to your early forecast last decade it was really about block storage. That was really what was under attack. You know, kind of fusion IO got it started with Facebook. They were trying to solve their SQL database performance problems. And then we saw pure storage. They hit escape velocity. They drove a truck through EMC sym metrics HDD based install base which precipitated the acquisition of XtremeIO by EMC. Something Renan knows a little bit about having led development, of the product but flash was late to the NAS party guys, Renan let me start with you. Why is that? And what is the relevance of QLC in that regard? >> The way storage has been always, it looks like a pyramid and you have your block devices up at the top and then your NAS underneath. And today you have object down at the bottom of that pyramid. And the pyramid basically represents capacity and the Y axis is price performance. And so if you could only serve a small subset of the capacity, you would go for block. And that is the subset that needed high performance. But as you go to QLC and PLC will soon follow the price of all flash systems goes down to a point where it can compete on the lower ends of that pyramid. And the capacity grows to a point where there's enough flash to support those workloads. And so now with QLC and a lot of innovation that goes with it it makes sense to build an all flash, NAS and object store. >> Yeah, okay. And David, you and I have talked about the volumes and Renan sort of just alluded to that, the higher volumes of NAS, not to mention the fact that NAS is hard, you know files difficult, but that's another piece of the equation here, isn't it? >> Absolutely, NAS is difficult. It's a large, very large scale. We're talking about petabytes of data. You're talking about very important data. And you're talking about data, which is at the moment very difficult to manage. It takes a lot of people to manage it, takes a lot of resources and it takes up a lot, a lot of space as well. So all of those issues with NAS and complexity is probably the biggest single problem. >> So maybe we could geek out a little bit here. You guys go at it, but Renan talk about the Vast architecture. I presume it was built from the ground up for flash since you were trying to kill HTD. What else do we need to know? >> It was built for flash. It was also built for Crosspoint which is a new technology that came out from Intel and micron about three years ago. Cross point is basically another level of persistent media above flash and below Ram. But what we really set out to do is, as I said to kill the hard drive, and for that what you need is to get the price parity. And of course, flash and hard drives are not at price parity today. As David said, they probably will be in a few years from now. And so we wanted to, jumpstart that, to accelerate that. And so we spent a lot of time in building a new type of architecture with a lot of new metadata structures and algorithms on top to bring that effective price down to a point where it's competitive today. And in fact, two years ago the way we did it was by going out to talk to these vendors Intel with 3D Crosspoint and QLC flash Mellanox with NVMe over fabrics, and very fast ethernet networks. And we took those building blocks and we thought how can we use this to build a completely different type of architecture, that doesn't just take flash one level down the stack but actually allows us to break that pyramid, to collapse it down and to build a single system that is as fast as your fastest all flash block device or faster but as affordable as your hard drive based archives. And once that happens you don't need to think about storage anymore. You have a single system that's big enough and cheap enough to throw everything at it. And it's fast enough such that everything is accessible as sub-millisecond latencies. The way the architecture is built is pretty much the opposite of the way scale-out storage has been done. It's not based on shared nothing. The way XtremIO was the way Isilon is the way Hadoop and the Google file system are. We're basing it on a concept called Dis-aggregated Shared Everything. And what that means is that we have the media on one set of devices, the logic running in containers, just software and you can scale each of those independently. So you can scale capacity independently from performance and you have this shared metadata space, that all of the containers can see. So the containers don't actually have to talk to each other in the synchronous path. That means that it's much more scalable. You can go up to hundreds of thousands of nodes rather than just a few dozen. It's much more resilient. You can have all of them fail and you still didn't lose any data. And it's much more easy to use to David's point about complexity. >> Thank you for that. And then you, you mentioned up front that you not only built for flash, but built for Crosspoint. So you're using Crosspoint today. It's interesting. There was always been this sort of debate about Crosspoint It's less expensive than Ram, or maybe I got that wrong but it's persistent, >> It is. >> Okay, but it's more expensive than flash. And it was sort of thought it was a fence sitter cause it didn't have the volume but you're using it today successfully. That's interesting. >> We're using it both to offset the deficiencies of the low cost flash. And the nice thing about QLC and PLC is that you get the same levels of read performance as you would from high-end flash. The only difference between high cost and low cost flash today is in right cycles and in right performance. And so Crosspoint helps us offset both of those. We use it as a large right buffer and we use it as a large metadata store. And that allows us not just to arrange the information in a very large persistent right buffer before we need to place it on the low cost flash. But it also allows us to develop new types of metadata structures and algorithms that allow us to make better use of the low cost flash and reduce the effective price down even lower than the rock capacity. >> Very cool. David, what are your thoughts on the architecture? give us kind of the independent perspective >> I think it's brilliant architecture. I'd like to just go one step down on the network side of things. The whole use of NBME over fabric allows the users all of the servers to get any data across this whole network directly to it. So you've got great performance right away across the stack. And then the other thing is that by using RDMA for NASS, you're able, if you need to, to get down in microseconds to the data. So overall that's a thousand times faster than any HDD system could manage. So this architecture really allows an any to any simple, single level of storage which is so much easier to think about, architect use or manage is just so much simpler. >> If you had I mean, I said I don't know if there's an answer to this question but if you had to pick one thing Renan that you really were dogmatic about and you bet on from an architectural standpoint, what would that be? >> I think what we bet on in the early days is the fact that the pyramid doesn't work anymore and that tiering doesn't work anymore. In fact, we stole Johnson and Johnson's tagline No More Tears. Only, It's not spelled the same way. The reason for that is not because of storage. It's because of the applications as we move to applications more and more that are machine-based and machines are now not just generating the data. They're also reading the data and analyzing it and providing insights for humans to consume. Then the workloads changed dramatically. And the one thing that we saw is that you can't choose which pieces of information need to be accessible anymore. These new algorithms, especially around AI and machine learning and deep learning they need fast access to the entirety of the dataset and they want to read it over and over and over again in order to generate those insights. And so that was the driving force behind us building this new type of architecture. And we're seeing every single day when we talk to customers how the old architecture is simply break down in the face of these new applications. >> Very cool speaking of customers. I wonder if you could talk about use cases, customers you know, and this NASS arena maybe you could add some color there. >> Sure, our customers are large in data. We started half a petabyte and we grow into the exabyte range. The system likes to be big as, as it grows it grows super linearly. If you have a 100 nodes or a 1000 nodes you get more than 10X in performance, in capacity efficiency and resilience, et cetera. And so that's where we thrive. And those workloads are today. Mainly analytics workloads, although not entirely. If you look at it geographically we have a lot of life science in Boston research institutes medical imaging, genomics universities pharmaceutical companies here in New York. We have a lot of financials, hedge funds, Analyzing everything from satellite imagery to trade data to Twitter feeds out in California. A lot of AI, autonomous driving vehicles as well as media and entertainment both generation of films like animation, as well as content distribution are being done on top of best. >> Great thank you and David, when you look at the forecast that you've made over the years and when I imagine that they match nicely with your assumptions. And so, okay, I get that, but that doesn't, not everybody agrees, David. I mean, certainly the HDD guys don't agree but they, they're obviously fighting to hang on to their awesome run for 50 years, but as well there's others to do in hybrids and the like, and they kind of challenge your assumptions and you don't have a dog in this fight. We just want the truth and try to do our best to report it. But let me start with this. One of the things I've seen is that you're comparing deduped and compressed flash with raw HDD. Is that true or false? >> It's in terms of the fundamentals of the forecast, et cetera, it's false. What I'm taking is the new egg price. And I did it this morning and I looked up a two terabyte disc drive, NAS disc drive. I think it was $54. And if you look at the cost of a a NAND for two terabytes, it's about $200. So it's a four to one ratio. >> So, >> So and that's coming down from what people saw last year, which was five or six and every year has been, that ratio has been coming down. >> The ratio between the cost Delta, between HDD is still cheaper. So Renan I wonder one of the other things that Floyer has said is that because of the advantages of flash, not only performance but also data sharing, et cetera, which really drives other factors like TCO. That it doesn't have to be at parody in order for customers to consume that. I certainly saw that on my laptop, I could have got more storage and it could have been cheaper for per bit for my laptop. I took the flash. I mean, no problem. That that was an intelligence test but what are you seeing from customers? And by the way Floyer I think is forecasting by what, 2026 there will be actually a raw to raw crossover. So then it's game over. But what are you seeing in terms of what customers are telling you or any evidence you have that it doesn't have to be, even that customers actually get more value even if it's more expensive from flash, what are you seeing? >> Yeah in the enterprise space customers aren't buying raw flash they're buying storage systems. And so even if the raw numbers flash versus hard drive are still not there there is a lot of things that can be done at the system level to equalize those two. In fact, a lot of our IP is based on that we are taking flash today is, as David said more expensive than hard drives, but at the system level it doesn't remain more expensive. And the reason for that is storage systems waste space. They waste it on metadata, they waste it on redundancy. We built our new metadata structures, such that they everything lives in Crosspoint and is so much smaller because of the way Crosspoint is accessible at byte level granularity, we built our erasure codes in a way where you can sustain 10, 20, 30 drive failures but you only pay two or 1% in overhead. We built our data reduction mechanisms such that they can reduce down data even if the application has already compressed it and already de-duplicated it. And so there's a lot of innovation that can happen at the software level as part of this new direct dis-aggregated shared everything architecture that allows us to bridge that cost gap today without having customers do fancy TCO calculations. And of course, as prices of flash over the next few years continue declining, all of those advantages remain and it will just widen the gap between hard drives and flash. And there really is no advantage to hard drives once the price thing is solved. >> So thank you. So David, the other thing I've seen around these forecasts is that the comments that you can't really data reduce effectively hard disk. And I understand why the overhead and of course you can in flash you can use all kinds of data reduction techniques and not affect performance, or it's not even noticeable like put the cloud guys, do it upstream. Others do it upstream. What's your comment on that? >> Yes, if you take sequential data and you do a lot of work upfront you can write out in very lot big blocks and that's a perfect sequentially, good way of doing it. The challenge for the HDD people is if they go for that for that sort of sequential type of application that the cheapest way of doing that is to use tape which comes back to the discussion that the two things that are going to remain are tape and flash. So that part of the HDD market in my assertion will go towards tape and tape libraries. And those are serving very well at the moment. >> Yeah I mean, It's just the economics of tape are really attractive. I just feel like I've said this many times that the marketing of tape is lacking. Like I'd like to see, better thinking around how it could play. Cause I think customers have this perception tape, but there's actually a lot of value there. I want to carry on, >> Small point there. Yeah, I mean, there's an opportunity in the same way that Vast have created an architecture for flash. There's an opportunity out there for the tech people with flash to make an architecture that allows you to take that workload and really lower the price, enormously. >> You've called it Flape >> Flape yes. >> There's some interesting metadata opportunities there but we won't go into that. And then David, I want to ask you about NAND shortages. We saw this in 2016 and 2017. A lot of people saying there's an NAND shortage again. So that's a flaw in your forecast prices of you're assuming prices of flash continue to come down faster than those of HDD but the shortages of NAND could be problematic. What do you say to that? >> Well, I've looked at that in some detail and one of the big, important things is what's happening in the flash market and the Chinese, YMTC Chinese company has introduced a lot more volume into the market. They're making 100,000 wafers a month for this year. That's around six to 8% of market of NAND at this year, as a result, Samsung, micron, Intel, Hynix they're all increasing their volumes of NAND so that they're all investing. So I don't see that NAND itself is going to be a problem. There is certainly a shortage of processor chips which drive the intelligence in the NAND itself. But that's a problem for everybody. That's a problem for cars. It's a problem for disk drives. >> You could argue that's going to create an oversupply, potentially. Let's not go there, but you know what at the end of the day it comes back to the customer and all this stuff. It's interesting. I love talking about the architecture but it's really all about customer value. And so, so Renan, I want you to sort of close there. What should customers be paying attention to? And what should observers of Vast Data really watch as indicators for progress for you guys milestones and things in the market that we should be paying attention to but start with the customers. What's your advice to them? >> Sure, for any customer that I talked to I always ask the same thing. Imagine where you'll be five years from now because you're making an investment now that is at least five years long. In our case, we guaranteed the lifespan of the devices for a decade, such that you know that it's going to be there for you and imagine what is going to happen over those next five years. What we're seeing in most customers is that they have a lot of doormen data and with the advances in analytics and AI they want to make use of that data. They want to turn it from a cost center to a profit center and to gain insight from that data and to improve their business based on that information that they have the same way the hyperscalers are doing in order to do that, you need one thing you need fast access to all of that information. Once you have that, you have the foundation to step into this next generation type world where you can actually make money off of your information. And the best way to get very, very fast access to all of your information is to put it on Vast media like flash and Crosspoint. If I can give one example, Hedge Funds. Hedge funds do a lot of back-testing on Vast. And what makes sense for them is to test as much information back as they possibly can but because of storage limitations, they can't do that. And the other thing that's important to them is to have a real-time experience to be able to run those simulations in a few minutes and not as a batch process overnight, but because of storage limitations, they can't do that either. The third thing is if you have many different applications and many different users on the same system they usually step on each other's toes. And so the Vast architecture is solves those three problems. It allows you a lot of information very fast access and fast processing an amazing quality of service where different users of the system don't even notice that somebody else is accessing the same piece of information. And so Hedge Funds is one example. Any one of these verticals that make use of a lot of information will benefit from this architecture in this system. And if it doesn't cost any more, there's really no real reason delay this transition into all flash. >> Excellent very clear thinking. Thanks for laying that out. And what about, you know, things that we should how should we judge you? What are the things that we should watch? >> I think the most important way to judge us is to look at customer adoption and what we're seeing and what we're showing investors is a very high net dollar retention number. What that means is basically a customer buys a piece of kit today, how much more will they buy over the next year, over the next two years? And we're seeing them buy more than three times more, within a year of the initial purchase. And we see more than 90% of them buying more within that first year. And that to me indicates that we're solving a real problem and that they're making strategic decisions to stop buying any other type of storage system. And to just put everything on Vast over the next few years we're going to expand beyond just storage services and provide a full stack for these AI applications. We'll expand into other areas of infrastructure and develop the best possible vertically integrated system to allow those new applications to thrive. >> Nice, yeah. Think investors love that lifetime value story. If you can get above 3X of the customer acquisition cost is to IPO in the way. Guys hey, thanks so much for coming to the Cube. We had a great conversation and really appreciate your time. >> Thank you. >> Thank you. >> All right, Thanks for watching everybody. This is Dave Volante for the Cube. We'll see you next time. (gentle music)

>> Live from San Francisco, it's theCUBE, covering Micron Insight 2018. Brought to you by Micron. >> Welcome back to the Embarcadero everybody here in the heart of San Francisco. Actually at the bay of San Francisco. Golden Gate Bridge is that way, financial district over there, Nob Hill right up the street. You're watching theCUBE, the leader in live tech coverage. I'm Dave Vellante, this is David Floyer, and we're covering the Micron Insight 2018 event. People are starting to filter in. Any minute now we're going to start the keynotes from the executives. A lot of buzz going on, Derek Dicker is here. He's the corporate vice-president and general manager of the storage business unit emerging activity within Micron, great to see you again. >> Thank you very much for having me. It's a pleasure to be here. >> You're very welcome, yeah, so Micron used to be just a straight memory company. We're hearing, we heard at the investor day in May how you guys are diversifying, finding new use cases, new applications, you run the storage business, and of course David Floyer was one of the first, the first, in my opinion, to predict the demise of the hard disk, spinning disk, and it's a tailwind for you guys, but Derek, take us through your business unit, your role, and let's get into it. >> Sure, that sounds great. I appreciate the opportunity again to be here. The storage business unit within Micron is actually comprised across a couple of product areas. Primarily NAND and NAND components, and then also SSDs, solid state drives. As we like to say, and we've talked a bit more about it since Sanjay's arrival, we have a pretty material focus on accelerating what we call high value solutions. It's a big focus of ours, so not only are we developing the core technology in memory and storage, but we're attempting to build more and more products that add value to our customers in the S-System space. But that's generally the storage business focus. Within the company, we have three other business units that focus on compute and networking memory as well as the embedded business unit and then the mobile business unit. >> Talk about some of the big trends that you see, I mean, we've talked about for years, the all-flash data center. We clearly see that in the customers that we work with. Some of the spinning disk guys don't necessarily fully buy into that, but even they have been investing in flash technologies. What are you seeing? >> I tell you, there is no better time, in my opinion, than to be in the memory and storage industry. When you look at what the trends are that are coming out in time. If you go and you stare at how memory and storage has evolved just going back into the 80s or the PC era, a $35 billion average size of the total market. You get into the mobile space, when mobile era started with smart phones, we were looking at a $62 billion-ish, and then in '17 we cleared $120 billion in size of the market, and we actually see a lot of secular trends that are going to continue to take us forward. A couple of things that are particularly noteworthy for us. The first one is the emergence of artificial intelligence, and machine learning, and deep learning. We're going to hear quite a bit about it here at the event. But in terms of a value driver for the consumption of both memory, DRAM, as well as storage, we see it going phenomenally up in content in every server that's purchased out in time. That's one, I think with the evolution of 5G out in time, we're also going to see that smart phone devices are going to end up having more memory to add features like facial recognition we see today, becoming mainstream, multiple cameras, that drives more DRAM content, but then also on top of that, storage is increasing. We're seeing, even today, a terabyte being put into some of the high-end phones, and we know that that's going to waterfall out in time. So I think if you look at this combination of what's happening both in the devices, you look at what's happening in the infrastructure, then you couple that with the processing that needs to happen, it's just an awesome time to be affiliated with memory and storage. >> Yeah, well, I've been following this LAN marketplace for the last, almost 10 years isn't it? More than that. And it's just broken through completely in the last two or three years. What are your thoughts about pushing compute closer and closer to that memory, adding to, for example, the SSDs, the capability of doing smart work? It's very very close to where the data is originally going to be placed? >> It's a great area of quite a bit of R&D work that's going on right now, and I actually think I view this as kind of two stages. One is there's the proliferation of solid state, as you suggested, it's been coming over time. I actually see it increasing dramatically as we look forward, and one of the key technologies that I think is going to enable that is QLC. The fact that we're now at a point where we're putting four bits per cell into devices, SSDs are starting to show up, I think that just creates even more opportunity. And I'll talk a little bit about that in just a minute, but I want to answer your direct question as to how that's changing with AnIML. But I think the ability, once solid state is prolific, to be able to architect systems where you can actually have processing take place closer to the media is a very interesting area. It's right with a ton of research going on right now. People are just starting to implement it. I think there's quite a bit of potential sitting behind it. You know, our focus, of course, is we're deploying, and as quickly as we can, on two vectors. One is, how do we proliferate more solid state into the market as an industry, and the second is how do we add value when we build those solid state drives, so I think it's definitely very viable. >> Let's talk about the significance of QLC. David, your forecasts early on were very aggressive in terms of pricing declines for flash. We kind of, maybe got caught off, a little bit surprised by the-- >> I think we were caught off by the demand. >> Well the demand, but also the supply constraints kept prices up. >> Yeah. >> Okay so, it didn't actually happen as fast. How does QLC change that, Derek, and what's the significance of it? >> Well, the thing that I think is most exciting for us as Micron is we actually ended up delivering the world's first QLC device. It put a terabit of data on a single die, which was unprecedented, but then in addition to that, what we did was we actually built a solid state drive called the 5210 ION. This is a standard drive. It's the worlds first SSD built on the technology, and by being able to develop a solution early on, it allowed us to go engage with customers and find where the right workloads were where we could add the most value. QLC technology actually is perfectly aligned for super read intensive, very read intensive environments, and if you look at what's happening in the data center, we're actually seeing more and more workloads move into more read intensive workloads, and a good chunk of that is just because there's analytics going on. The data's being collected. It's being housed in on place, but as we've talked about quite a bit here at the event, we want to be able to deliver insight out of that data, which means we're going to be reading it quite a bit, and massaging it, and performing analytics on it. And what we're now seeing is what, in the days of the past, was a four to one read to write ratio, we're seeing as high as 5,000 to one and in some cases a million to one. So we get these heavily read intensive workloads coupled with the technology that's optimized for it. It's more power efficient than what rotating media solutions offer in certain workloads, we're starting to see these tremendous values coming out of these early engagements that we're having with customers. >> And does that have implications for longevity, or do you just make an assumption that the read/write ratio is still going to be more write intensive in terms of wear leveling and things like that? How does it change the reliability, if you will, of the technology? >> Actually the beauty is, we're able to deliver an enterprise class SSD with these read/write capabilities that are affiliated with these read intensive solutions, and we can fit within the workloads and the needs that people are talking about. So the drive writes per day that are required in a machine learning infrastructure, we believe we can address with QLC. Same thing with Hadoop style clusters and Ceph clusters. We've actually, as we've gone out and engaged each of our earlier customers, we're able to crank out reference architecture documents that we're now posting to our websites, and we're describing how we can actually leverage this technology to allow us to, in some cases, we'll better optimize where an SSD was used before. But in many, many cases we're actually in the process of displacing hard disk drives. >> So what are the limits of this QLC? How many more bits can we add? How many more layers can we add? >> So, it's actually a great question, David. In terms of what does a roadmap look like. I've been asked in the recent several hours, what the longevity for NAND looks like. And what I'll tell you is this, QLC NAND is just getting its start. What comes after that in terms of bits per cell, I don't think anybody's made any broad claims on. But from a layer stacking perspective, which is kind of the dimension upon which the industry is growing, for the foreseeable future, we see nothing that encumbers us from going substantially higher and higher layer count. Which I think is going to be great for our industry because it's going to allow us to deliver more bits in a given device, and hopefully, that'll allow us to get into markets that, historically, we haven't been able to approach. If you think about the demand elasticity dynamic that occur when we start to bring more and more costs down, the number of applications open up, not unlike the machine learning workloads I just mentioned or Hadoop workloads. We're starting to see more and more thirst and interest for replacing with solid state, just because it's more power efficient, allows for a cost structure that's better, and gives better performance too. >> I'm fond of saying that data's plentiful, insights aren't. You guys are a $30 billion company now. You're making some interesting announcements today that we're going to hear about a little later on that I won't divulge right now, but you're putting your hands in a lot of different places. When you're that size of a company, you can't help but, as you mentioned before, adding more value, becoming more of a systems focus. How do you help the industry go from just raw data to insights? What's your role in that? >> Oh, it's a phenomenal question and this is a major focus of the company. Not just in our business unit, but across all of the different business units in the company. We have a huge focus on sitting down with our customers and getting closer and closer to understanding what their workload needs are, where their paying points are, and then working with them to find solutions, and the beautiful part about it is, as Micron, we're the only company in the world that can combine together a 3D XPoint set of technologies, a NAND set of technologies, a DRAM set of technologies. We go sit down and talk about these challenges with those in mind, plus the emerging memories that we're developing to go develop better and better solutions. But after we're able to come to a solution, we put together a reference architecture, and we deploy it broadly. >> We've been trying to squint through 3D Xpoint and understand the right fit. It seems to us that one of the big advantages of flash was it had the, had this behind it. (laughs) It had the consumer volumes, thank you Steve Jobs. It's unclear whether or not 3D Xpoint will have that, maybe have the same, sort of, cost advantages, but the same time, it sounds like there's new and emerging applications. Like I said, we're trying to figure out. Have you guys figured out yet? You're obviously betting big on the technology. Help us understand where the fit is. >> Sure, I think, you know, if I look back in time, just at the storage hierarchy alone, I don't think the memory hierarchy's any different. You have these portions of the market where you build out hard disk drives, and we had DRAM before, and SSDs came along, and people started asking, not unlike several years back when we talked about the early parts. Hey, how big is this going to get? Cost structures may be prohibitive. But as innovation unfurled, the more time and investment got placed into it, we found new workloads, new use cases we were able to drive costs out, and we ended up slotting in solid state drives squarely. I think this is another tier of memory and storage. That's the beauty of the 3D XP technology. There's both memory semantics and storage semantics that are available for use. I think we're still scratching the surface on the early days, but I love what we're seeing from the customer base that we're engaging and targeting in this space. >> And people will pay up for that performance capability relative to flash. They'll pay down relative to DRAM. Is it, are you seeing a gradience for like the hyperscalers, for example, or is it, maybe the industrial internet? Where are you seeing the. >> It's fair, actually I think, you know, it's probably reasonable to say that, you know, the challenges of inserting a new memory tier into a system requires new programming algorithms, new APIs and interface. There's a lot of ecosystem that needs to be there, as well as, not to mention, you've got to have an ecosystem to go put memory products into a server, for instance, or any other platform. I think we're still early days of enabling all of this. And I also believe we're going to learn more and more where the value of this sits as we put it out there in a cost effective fashion. So I would say that people who control software environments are very, very well suited for this because they can take advantage of some of those challenges without having to have a whole ecosystem in place. I think there's going to be a continued ramp in acceleration as an industry we go build out that ecosystem. >> Well it's been amazing to watch Micron the last several years, I mean, the last several decades. When you were just a pure memory manufacturer which was diversified, you know, gorilla in this space. (laughs) You guys are really an extremely well run company. I mean, your financials have born that out. You're really transparent to the street providing great guidance and congratulations on all of the success. I'm looking forward to watching in the future. >> Oh thank you so much. It's a privilege to be part of the company, and I really appreciate your time today. >> Our pleasure, thanks for coming on theCUBE. All right, keep it right there everybody. We'll be back with our next guest right after this short break. You're watching theCUBE from Micron Insight 2018. (upbeat techno music)

>> Live from Las Vegas, it's theCUBE! Covering VMworld 2018. Brought to you by VMware and its Ecosystem partners. >> Welcome back to theCUBE, I'm Lisa Martin with David Floyer, and Dave and I are here, day three, David, of our VMworld 2018 coverage, if you can believe it. We're excited to welcome to theCUBE, for the first time, a couple of gentlemen from Micron. We have Eric Caward, business development manager, and Greg Kincaid, ecosystem enablement program manager. Welcome guys. >> Thank you, good to be here. >> Thank you very much. >> So day three, you still have voices, that's impressive, your feet are doing okay? >> Yes, yeah. >> Pretty good, pretty good. >> Good, so Greg, tell us a little bit about your role and specifically what some of the new exciting announcements from Micron with respect to flash. >> So my role is to find deployments where SSDs can improve the performance significantly. Also, any case where you can have simplicity for the system administrator. So, with the new version of VMware 6.7, we've got, we've implemented, using NVMe as our cache layer, and set as our capacity layer to get tremendous performance across the spectrum of reads and writes. >> So can you give us some examples of how good that performance is? What sort of impact have you had? >> So, take for instance using NVMe as the cache layer and as data and a capacity layer, you can get small block random reads of 500,000 for a new cluster. >> That's very impressive. >> Yeah. >> Yeah. So can you make some savings in terms of the improvements in the VM density and things like that that you can achieve-- >> Absolutely, so almost all of these, well, all of the SSDs are in a two and a half form factor, and so you can get much better density per U with those kinds of SSDs, as opposed to a hard drive where you have to go to a three-inch to get that kind of density. >> So performance density, tons of data, what are some of the things in your opinion, Greg, that differentiate Micron Solution here, versus all those other guys out there? >> Well, we don't just put together a solution. We actually do considerable amount of testing, both in benchmarking, we also do a quite a bit of application testing as well. And we publish a very thorough reference architecture that's available on our website to act as a pragmatic blueprint for those who want to implement those kinds of solutions. >> Excellent, excellent. So, Eric, you're a part of the NVDIMM brigades. >> Yes. >> Tell us what is NVDIMM. Why is it important? >> Well, NVDIMM is very exciting. It's basically a memory that doesn't forget. So it's on the memory bus, it's comprised of DRAM, a controller, and NAND, and when the power is catastrophically lost, all your data is retained. >> So you go up to, what is it, 32 gigabytes on the DIMM? >> Actually, yes we're releasing our 32 gig NVDIMM in production next month, which is right around the corner. >> Wow, and and how many DIMMs can you have in a? >> You can have up to, typically in a 24 socket system, you can have up to 22 of those can be NVDIMM should you wish to. >> That's a lot of memory. >> It is a lot, and it's very, very fast. >> Very, very fast OK, so, tell us some of the changes that need to be made in order to exploit this. This is this is different, isn't it? So, can you give some examples of how you're working with the ISVs, for example? >> Certainly, certainly. From the operating system standpoint, Microsoft Windows Server 2016 supports, natively supports persistent memory. So does the Linux kernel version 4.2 and newer. Along with that, not only that, but you also have applications that are written from the ground up to support to be persistent memory aware. You have Exchange Server, you have SQL Server 2016, and with those applications they can actually access the persistent memory in byte mode, which is much faster than block mode, but you also can more legacy applications can get benefit from block mode, also. >> Wasn't, sorry Dave, I was just going to say let's dig into a customer example. I always love to hear how are these technologies, one, being co-developed as in collaboration with the end-users, right? And two, how are you seeing them in the, in the field actually helping customers transform their businesses from the inside out? >> Well, so one example that comes to mind, actually VMware just did a study with Oracle licensing, and they took a 12 core solution, and they put the redo log onto traditional storage, and they were able to get a certain amount of performance. Let's just call it a hundred units of performance. They did the same thing with the same workload, but they only used nine cores. So, that's actually a reduction in 25% course, but because the redo log was actually put on persistent memory, which again you're accessing that storage at DRAM like speeds, it kept the CPU much, much more busy, much more active, and they actually saw about a 2% increase in performance, but because the licensing costs are tied to your core count; actually, you could potentially save on licensing cost, even though you purchased a NVDIMM to have faster persistent storage. >> What about other benefits like to a data center in terms of energy efficiency? One of the things that Pat Gelsinger said on Monday was that VMware and their Green Charter, if you will, has saved 540 million, I think, tons of CO2 emissions. What I'm hearing Eric, what you're saying, are customers seeing pretty significant like power savings, and that were like roll into cost savings with the performance in this speed that you're able to deliver? >> Yes, if you look at it one of the other use cases for the NVDIMM, persistent memory, is that they used to NAND storage to write these logs, but because of the endurance, it ends up that they would have to replace the SSDs on a three month cadence. Because of the NVDIMM, the endurance it has just natively comes with DRAM, they were able to replace the SSDs with the NVDIMM, and then continue to use that for many, many quarters. >> It's a big cost savings. >> Definitely. >> So, can I go back to the what we were talking about before in terms of implementation of this? >> Yes. >> So, what's necessary? You need the software, the ISV software. You obviously need the Micron and the DIMM. >> That is correct. >> Anything else that you need? >> Yes, the actual, the hardware that you have to have, you have to have, not necessarily a specific CPU, but if you have to have the BIOS that basically goes in and is aware of NVDIMM. >> Right. >> And, one of the reasons why is when a system boots up, that supports NVDIMM, it goes out and looks and sees, is there a valid image set to true? If so, it will load that image from the NAND, through the controller, into the DRAM. Then when it's completed, it will go on to booting up the OS. The OS is none the wiser that that data wasn't sitting in DRAM the entire time, but as you can see if your, if your bios support isn't there from the start with that, that process would never happen. >> But, you can have that BIOS is available on most, most system. >> On multiple, multiple OEM systems. Yes, that is supported. >> Great. So, that there's no requirement for anything special with other than that? >> Other than that, correct. >> That's amazing. So, you've got a pretty, are you going through other ISVs as well? Are you. >> Yes, there are multiple ISVs that we're working with to enable that, basically the performance benefit and the endurance and the low latency of NVDIMMs. >> And people like SAP, for example? >> Yes. >> Perfect. Okay, that's very excited, very, very exciting indeed. Are you doing the same thing with your, class? >> Yes, we actually work with many partners. We work with not just Vmware, but all of the enterprise partners. We do case studies, and we do cost analysis as well. So, for instance we found that if you statistically, strategically add an SSD to a 200 node cluster for Hadoop, you can get the same performance there that if you had added 80 additional nodes for the entire cluster. So, that's quite a bit of a savings of 80 nodes versus an additional 200 NVMe SSDs. >> Yeah, that's great. >> What's some of the feedback on these new advancements that you're hearing from some of the people that are coming by to visit the Micron booth here at VMworld? >> Well, I think people are a little surprised that we are so focused on systems, and making sure that they work on the performance with SSDs. I think people, sometimes they think of Micron in the early days when we were just simply a commodities broker with DRAM, but we're much, much more than that. >> So, customers are reacting to what sounds like an evolution of Micron? >> Absolutely, absolutely. >> Eric, what are some of your-- >> And to be honest, my favorite is when people come by, and they look at the numbers, and they're just like oh my gosh. (laughing) The performance is really outstanding when you look at an NVDIMM, and it's just, it's simply because it is DRAM acting as a storage device. It's sitting on the memory bus. It's sitting on the memory channel, right next to the CPU. The latency is absolutely fantastic. There are certain workloads that are really, really gain a lot of benefit by low latency for quality of service. Then you have just the raw bandwidth, and this is only with two NVDIMMs in this particular demo system. We could have, excuse me, we could have gone up to six in a CPU. So, we could have tripled our performance just with one CPU on one node. So, it's pretty exciting when when the people that are coming in the booth, they get excited too. It makes, it makes this show really fun. >> I think people also don't understand that there's more than one kind of SSD, and we just announced that QLC, a NAND based SSD, that for write once read many could actually supplant many of the hard drives that are used in secondary storage or archives. >> So, it also must be kind of fun to educate people on, hey guess what? There's not just different flavors, but look what Micron is doing. >> Right. >> Evolving our technologies and enabling them to you know, learn about things that they didn't know about. I imagine that must also be a pretty cool. >> I'm working with a software developers as well, so closely, so this is exciting. >> I mean the applications are just innumerable. I mean we're working with artificial intelligence. We're working on machine learning. Applications are other than just the standard database that most people think of accelerating with SSDs. >> Excellent. >> And, to be honest, I'm very passionate about technology, just, I love to geek out, if you will. >> I can tell. >> And, I love seeing the light bulbs come on in people that I'm talking about. It's just very rewarding. >> So we're gone, more than halfway through 2018, scary. September 1st is Saturday. (laughing) So, going towards the end of the of the calendar year, this excitement that I'm getting from both of you, what are you excited about Micron, you know going into early part of 2019, being able to surprise and delight your customers with? >> All right. >> Well, we're going to continue to, to do all of the performance testings that were done. We're going to, as we bring new SSDs to the market, we're going to continue to add tuning advice, and detailed deployment instructions for our customers. We're going continue to partner with the major players to make sure that our SSDs, their performance and their applications. >> And I think with the fact that we're releasing our 32 gig NVDIMM, actually in September. The ecosystem, as it solidifies, it becomes more robust. There's just going to be use cases that our engineers and our team haven't thought of yet. And, so it's going to be really exciting to see what new use cases are out there for super, very fast NVDIMMs. >> Well guys, thanks so much for stopping by and talking with David and me about-- >> Thanks for having us. >> The evolution of Micron, and the excitement that you get from from hearing that validation in the field, and we look forward to hearing what's coming out shortly. So, we'll have to have you back on. >> Sounds great, thanks Lisa, thanks David. >> Love to be back. >> Excellent. Greg, Eric, thanks for your time. For David Floyer my co-host, I'm Lisa Martin, you're watching theCUBE, live from Vmworld 2018. Stick around, we'll be right back with our next guests. (electronic music)