 Hello and welcome to this special conversation with Steve McDowell who's the principal analyst and founding partner at NAND Research and we're talking about the next generation of storage Steve, welcome. Hey, it's good to be here Dave. Hey, so tell me first about NAND Research, congratulations for getting that up and off the ground. Yeah, so NAND Research, we're just celebrating our first anniversary. We're a small analyst firm, we focus on kind of the horizontal technologies around the data center. So everything from servers and storage to kind of the software stacks that live right above that. So it's an interesting space to play. So let's talk a little bit about IT complexity. You know, cloud was going to solve all IT complexity, but you know how it is in this industry, we solve complexity with complexity. And so where do you see the state of IT complexity today and then we'll get into how that's going to change. Wow, so it's as complex as it's ever been. I think, you know, I've been in this industry now for 30 years and every time we think we've solved the complexity problem more comes along. I mean, there was a period, you know, maybe eight, nine years ago when, you know, the world believed that cloud was going to be the answer, right? Let's take all of this, especially when we're talking about infrastructure, let's take all of this infrastructure complexity and push it onto somebody else's plate. So I can just write an Amazon to check and I can focus on the business of digital transformation, right? But I think as we all know, hasn't played out quite as well as people predict, right? There's some workloads that just naturally want to live on-prem, right? Maybe whether it's for governance issues or latency issues or whatever, right? So what we've ended up doing was creating more complexity because now we have this whole kind of hybrid cloud infrastructure, right? I think you call it super cloud where, you know, now I'm a poor IT guy, I got to manage stuff in the cloud, maybe multiple clouds and on-prem and it just gets harder and harder. But what cloud has also done is it's reset the expectations of what an IT experience can be. It's told, you know, not just the IT guys, but their CFO and finance guys that, you know, I don't have to buy all of this equipment, right? Maybe there's a self-service, maybe there's a consumption-based model. So we're living in an extremely complex world that's a blend of clouds and then even on-prem it's a blend of, I own my rack, I rent my rack, I have a consumption-based, you know, as a service. So there's a lot going on, a lot going on for the poor IT guy. So let's explain that a little bit for people because some people might say, well, you could always, you know, lease. So it was OPEX, not CAPEX, but you're talking about something different. You're talking about the actual experience, that cloud experience coming to on-prem and hybrid. How has that evolved? And then I want to get into specifically how storage is really taking more of that burden. Oh, sure. And it's, you're absolutely right. I mean, I could always lease and, you know, the rise of consumption-based coincided with a couple of things. One is they changed some tax regulations in some countries that made leasing less financially attractive, right, from a tax basis. At the same time, you know, cloud has taught us that a managed experience is a pretty good thing. You know, if I can take storage, for example, you know, install a rack of storage and not have to rack and stack it and configure it and worry about the network and, you know, all of the technical details of the physical piece of it and just worry about the data piece of it. You know, as I'm driving my data or digital transformation projects forward, it's a good thing, right? So I think we've seen the consumption-based experience evolve from kind of a lease, which still exists in some context, but what's really growing and what's really popular right now is the managed experience. So I'm getting, you know, compute as a service. I'm getting storage as a service. Okay. And so let's double click on storage a little bit. How, where's storage's role in the value equation and is storage doing more of the work? And if so, how? Storage, storage is evolving rapidly. And we could talk about any number of things from, you know, flash consuming hard drives to the role of data and how storage is consolidating. But I think at a macro level, at a macro level, what's happening is, you know, data is becoming more critical to the digital enterprise than ever before, right? It used to be about, you know, we're putting processes onto compute and that's our digital transformation. But now we're not about data transformation. You know, and some of this coincides with the rise of advanced analytics. Some of this coincides with the rise of AI, which is I have data all over the place and I need to consolidate that data, you know, whether it's a data lake or something, you know, similar. I need to consolidate that data and manage that data because that data is now driving my organization. That's A, B is the cybersecurity aspect, right? We're in an environment now and it is kind of rapidly taken over, which is, you know, ransomware and bad actors are now targeting my data in a way that they never did before. It used to be I'll lock up a machine and hit some ransomware. Now it's all encrypted all of your data. So the role of data storage in terms of protecting that data is also evolving. You know, so we're seeing things like immutable snapshots and ransomware detection kind of in the storage device itself. So there's a lot going on in storage and it's taking, you know, more front and center role. So obviously flash is not new, but it continues to evolve. Where are we at in terms of flashes impact, you know, the all flash data center? I mean, we go back, you know, last decade, but there's still opportunities for organizations to apply flash in new ways. You could do obviously a lot more. You can compress more without the performance penalty. You can share data in a more facile way. So where are we at in that whole curve and the crossover with spinning disk? Well, I think we're about to hit kind of the second evolution of flash, right? 10 years ago and pure storage was, you know, I think key in pushing the industry in the direction of all flash because legacy storage vendors really had no desire to go there, right? It's disruptive to them as much as it is IT, but the value of flash proved itself. So, you know, kind of the first explosion of flash, you know, 2014, 2015, we started pushing flash into high performance workloads, right? It was still expensive, right? But you could not beat the performance or the density and all of the stories around that. But it wasn't going to replace disk drives, right? Not as it existed. As these things do, flash has evolved, right? So we're now, we're looking at technologies like QLC NAND, for example, which less performance and has some other, you know, kind of SLA characteristics, but it makes it much more attractive versus hard disk drive. So we're starting to see now QLC flash push down into near line storage, right? And if we look at the density curves and we look at the price curves over time, right? We're not far off where, you know, by the end of the decade, maybe, where we'll see Greenfield storage installations being all flash, and that's an exciting time. We couldn't have visualized this even five years ago, but you know, the densities, the price, the performance, the reliability that we put around QLC make it a viable alternative. So I've got your report from NAND research tackling IT complexity with simplified enterprise storage. And in there, you have this nice chart and sort of laying out the workloads that are appropriate for QLC NAND, you know, near line, big data, analytics, data lake, et cetera, data protection. And then TLC NAND, OLTP, right, is the wheelhouse, and that's the right intensive workloads. And what I'm hearing from you is that flash is expanding its total available market, if you will, in terms of use cases. And that's a function of economics. Is that correct? And also the tech itself? Economics and capability, yes. So the price of NAND is, for some capacities, has already crossed over that of hard disk drives. And we can see a point in time where at the same capacity point, the NAND will be as cheap or cheaper than spending hard drives. And the other thing to think about when you think about the economics of flash storage is it's not just the acquisition cost, it's also the operational cost. Flash, when it's not serving data, doesn't consume power or it's so minuscule that it doesn't count, right? And the same can't be true of hard drives. So it's an economic point, but it's also a technology point. When we look at technologies like QLC flash, right, it arrived a little bit of baggage. They said, QLC doesn't have the endurance, doesn't have all of these things that the TLC, which is the high performance flash does have. And some storage vendors have taken some innovative approaches and said, well, we can engineer around that. And we're seeing that. If you look at pure storage in their E-Series, flash array and flash blade, which uses QLC NAND, right? They've put a lot of software engineering on top of the QLC to solve kind of the things that gave QLC a sketchy reputation to begin with. So it's proven itself as enterprise class, right? We've all been shipping QLC now for enough time where we can trust it as an industry. So it's both technology and economics and they're converging at this point. Right, thank you for that. So in your report, we have this, I want to talk about sustainability. You've got this stat data centers account for about one to 2% of global energy consumption with storage accounting for about 25% of data center usage. And most of that storage consumption of energy is from spinning disk. When you're moving disks around, you're obviously using more energy. So how does flash sort of address that problem and where do you see that headed? Well, flash is a semiconductor. So it consumes energy, but a minuscule amount relative to something mechanical. If you look inside your hard drive, you have a motor in every one of these hard drives that's spinning at 7,500 or 10,000 rotations per minute. So those things are consuming power. Even at idle, right? A hard drive doesn't completely stop. So from a sustainability perspective, right? There's no question that flash is more sustainable. I mean, just in terms of raw electric usage, we can measure this, we can look at this. So there's that, there's also the, not just the power consumed, but the heat generated. Flash doesn't generate the heat that a mechanical hard drive does. So if I'm looking at a data center scale, right? There's less that I have to heat and cool. So there's a little bit of a ripple effect. And then kind of thirdly is the density story. I can pack a whole lot more bytes into a flash array than I can a hard drive based array, just in terms of the physical footprint, which again comes to the environmental impact inside the data center. I don't have to heat it. I don't have to cool it. I don't have to have as many racks. So it's just an overall better story. Yeah, these are key points. I mean, of course the CPU and GPU are the big culprits, but with all this activity going around in AI and the world of GPUs, GPUs are not only hot from a market standpoint, they run hot. So whatever we can do to sort of attack even at 25%, because it feels like from an economic standpoint, you're really starting to, like you said, this is now the next wave. So it's sort of a no-brainer to sort of point flash at that problem. And then I'm sure there are many folks, smart people working on the heat of the CPU and the GPU and liquid cooling and the like, but it seems like flash as a replacement for spinning disk is imminent, at least in a lot of use cases. I think it's inevitable. And nobody's going to go in and wholesale rip out all of their existing drives, right? We're following the replacement cycle. So the hard drive based arrays that are being sold today, those will be replaced by flash in five years, right? And I think post, we're already in the kind of gray area now where if I'm looking in near line storage, things that's relatively hot and accessed, I'm going to consider QLC, right? I'm going to make that, I'm going to do that math and make that choice, whether I go hard drive and flash, right? It's not a choice that I would have made three or five years ago even, right? For a lot of these workloads, but I think by the end of the decade, right? If the trends hold, all new green field, right? With maybe a few exceptions is going to be flashed towards. So Pure has, Pure was always trying to be the first at. They were like the first with the evergreen model. They were the first, I think the first company to actually partner with NVIDIA as I recall. And so you've got a chart in here that shows improvement curves, capacity and over time, HDDs, SSDs, and then you've got a curve on Pure, which is much steeper. Why is it the Pure is able to do that? How are they doing in the market? Where do they fit? What's your perspective on Pure? So Pure is an interesting one. If you look at the OEMs that sell storage today, most of them have a legacy business, right? If you look at Dell, right? They consume DMC. They already had a storage business. They've got, I don't know how many product lines that are all kind of running a little bit different. They can sell you storage for whatever application you want, but it's a little disjointed. And when I'm looking at where do I evolve my product line if I'm a Dell or an HPE or a Lenovo, I have to take into account all this legacy stuff. What I like about Pure is, they have a few models, but at the end of the day, they really have one set of technologies. They have one operating experience. And I don't know that any other storage vendor can say that, certainly not a tier one storage vendor. So they've taken that and they started with a philosophy and their philosophy wasn't about, let's push flash into the market. I think they said, flash solves a real problem and we can use that as a lever to simplify the experience. And I think for as long as I've been watching Pure for a decade plus now, it's all been about simplifying. It's not just simplifying flash storage, it's simplifying the whole storage experience. So where I'm a fan of Pure is how they apply that philosophy of simplification to everything, whether it's consumption based, whether it's even procurement. There's a reason that Pure is probably the only storage company I think that publishes their MPS score, which is in the 80s. And that's insane for an infrastructure company. But it's because it's in their DNA to provide a simplified experience. And again, it comes back to where we started the conversation, IT is complex, right? And IT guy, man, he struggles every day. Anything you can do to solve this problem is gonna resonate and it's gonna be goodness. And that's where Pure's focused. And it's less about the technology and more about the approach, right? Well, that MPS score, off the shelf, I'm just looking that up, that's insane, is right. I think they're, I just Googled it. Pure is an MPS score of 86. And Apple, Apple's in the 70s, I think. That's unbelievable. So, and you're right. I mean, Pure has always been focused on simplicity. Storage, as you well know, you've been in the business as of I for a while. Storage, simple was never how you described storage back in the day. So, Steve, it was great having you on. Congratulations on your one year anniversary. I really appreciate you sharing the results of your perspective and research. No, thank you, Dave. It's good to see you always. All right, cheers. And thank you for watching the next generation of storage made possible by Pure. Now, to learn more, please check out the resources tab and the links in the description of this video. Bye for now.