Transcript
I'm your host, Rob Luedemann. It is time to bring the orange, and it is the most exciting time of year, Chad Kenney. That is right. It is launch time. It is Accelerate time. By the time that we publish this, we will have gone through Accelerate. Couldn't think of anybody better to bring in to chat through all the cool, exciting things that we're doing. Lots going on right now. Yeah. Yeah. Absolutely. And it is a celebration. Launch should be that time when we take the culmination of everything everybody's worked on and really celebrate it. Super excited to do this series. We did it last year. It was a fun one. And it performed really well. We'll be going through, in this episode one, we'll be hitting the intelligent control plane and all the great things around automation that we're bringing in. You love talking about that. Episode two, we're going to come back and we're going to do unified data play, and you can't do all those great things around automation unless you have a firm foundation to support it where the data is all easy to access. And then number three, super excited to get the former CEO of OneTouch to get Ashish in here. He's awesome. And he is. And go into the ever pure data intelligence realm. How do you feel about that being added into the mix, like as far as a storyteller that you are? I love it, in fact. If you think about the story we were telling for a long time, it was this storage management kind of evolved into building, you know, data set management or autonomous operations. Not easy to say, apparently. Yeah. And then, of course, getting into data management, like the more we could delve into the understanding and meaning of data was really kind of like the longer term goal as we started to abstract away the infrastructure and build some autonomy. So it's exciting. Like, I will say I'm learning more than I ever had in data, which has been killer. And being able to see some of this in practice has been fun as well. Yeah. And it feels like it's just super complimentary to like what you said, those two or three things that we've been laser focused on the last three or four years into building up. What's driving this, right? I mean, the answer is going to be AI because it always is. But really, I love when we kind of look back in the history and look at how like applications, they were the dominant force, right? Still very much so. You deployed to apps and you kind of created silos. So what's changed now? Yeah. So if you think about like a traditional app, it holds a lot of data, you know, and context associated to that particular thing. But like if you were to just think holistically, like let's say I want to look at a customer view even here at Pure, right? You've got support data. You've got sales data. You've got, you know, stuff we've shipped. You've got licensing, like a whole bunch of things that are all their own silos. And to interconnect all those things, you used to have to kind of pull the data out and ETL it and keep it kind of continually fresh and then munch through the data to make sure it all interconnected to kind of bring that context together. But imagine if you could actually bring that context into what we kind of like to call a canonical data layer, like truly understand the meaning of the things across all of those different data sources and bring data to the forefront or make it primary in the thinking versus the application being primary in the thinking. It changes the game quite a bit if you think about it because AI thinks in this manner. They want to understand the context behind a holistic understanding of data really to be able to build those connections and build some understanding and infer answers that you're trying to get out of it. But with data being siloed, it's nearly impossible to pull that off without being able to really build a semantics layer. So it's a killer time to be in data and in tech, especially with AI, trying to drive a lot of these new solutions that we need to build out there. And it really is challenging us, right? Because we've been doing this for a long time and that notion of an app tied to that specific data set. And as you were going through that, I was thinking about some of the challenges that we've contended with here maybe over the years, like, as you said, with the customer data and then, you know, you have data science teams who have to go piece it together. And AI operates the same way. It doesn't care, but it needs to see everything. And if you have data lost in silos or not accessible, you're not going to get the right answers or you're not going to get them as quick as you would want to. Agreed. It's a fun time. I feel like the one-task acquisition really kicked off the whole data management layer. And we'll continue to see more stuff here, right? This is a fun time to be alive and in data, for sure. Yeah. And as you said, a different way to talk about it than we're accustomed to. Totally. I've been on two sales calls today, and it's funny, like, they're always like, I need to understand this better. Like, you guys are getting into semantics and ontology and things you've never talked to me about before. I really need to get my data teams involved and understand it more. And so it's bringing in a new group of individuals who are starting to understand Pure more holistically and helping us broaden out our solutions to add more value to our customers, you know, as a whole. Yeah. I think we're really opening our arms to some of those other groups and personas. It's going to be really exciting. Well, on to the topic for this launch episode, and it is one of your favorite things, if not, besides your family, besides talking about your wonderful family, the Intelligent Control Plane and just the innovations that we've been driving there. Set up the Intelligent Control Plane for people if they're just tuning in for the first time. Like, what are we talking about here, and how does that relate to what we've been doing for data and storage management for a while? Yeah. So if you think about it, data has effectively been siloed by storage systems. Pretty much storage as a whole has always been managed by those things, right? So when you bought a storage system, you deployed some applications on it and you manually managed each one of the subsequent things, right? And we have this kind of logic line that said, what if we were to break down those silos and actually enable people to really be able to build a unified data layer, but do so with intelligence and policies and really intent, right? You want to be able to define your intent and have that system be able to manage it. And so what we did is we built in a bit of software into Purity. We call it Fusion. It fuses these systems together as one holistic, nicely done. And what's killer about this is now you can actually have these systems intercommunicate with each other. I use the analogy of, like, you know, autonomous cars, right? Like today when you and I drive, we have pretty minimal communication amongst cars. Like, you got some hand signals, maybe not the best ones. You got blinkers and you got a horn, right? That's about it. You don't know what the intent of that other driver is. And that's what causes accidents and traffic and all the other kind of things that suck about driving. But if you think about it, if these cars intercommunicated with each other, there would be no traffic, no accidents. It would be super efficient, and you wouldn't even need to really pay attention. My intent would just be to go to work, and when it gets there, I get out of the car. And so we wanted to build a storage system that really delivered that same kind of model. Almost intelligence intent with policy, and allowed it to be governed as one holistic fleet. And that was very different. And the paradigm shift, I still think, for customers is kind of weird for them, right? Because they're used to, I manually manage these systems, and my team has tribal knowledge of what it is. But what we're doing now is we're taking that tribal knowledge and actually governing it as a single entity and making sure it occurs versus hoping it occurs with humans. Yeah. And doing it at scale, right? When I'm thinking about your autonomous car example, okay, let's say they're all connected together. Yeah. But let's say something changes in the conditions, like there's a weather event, and you need to drive differently. Okay. Normally, we all just as individuals start driving differently and usually poorly if there's rain or if there's snow. But here, you just push it to the entire fleet of vehicles, and they all slow down because there's weather and they drive more safely. Same thing that we're talking about here with the storage layer. It takes a new abstraction layer, though. So this was the big kind of epiphany, I guess, for me at some level to think through this. But when we looked at storage in the past, it was a bunch of volumes, and maybe you grouped those together for an application. What we're doing now is we're abstracting away the infrastructure stack. We fused these systems together, which is killer, and built this unified data plane. But this intelligent control layer actually is starting to manage things as workloads. They're a set of data with policy and service levels associated to it. And so now you're not just deploying storage, you're actually managing a holistic thing that now has policy and service levels attached to it. And the benefits of that is now you can apply intent to data. And this does change the paradigm a bit, because that workload can move to any system. And so that starts to change the thought process of things. And I can change intent, and it applies that to all relevant workloads. The power of this is spectacular. It is. And it's super fun to talk about, you probably can tell already. I love really helping customers down that logic line, because I think it's so exciting to have that kind of control and be able to showcase the benefits of that to customers very easily. And furthermore, that you need a platform to support that. So episode two of this launch, and I'll just plug because I want people to come back, is we'll be hitting the unified data plane. I won't go into a lot of detail here, but unless you have the ability to support different data protocols, different data types, and do it at scale, but do it also across the entire environment with the same levels of management, right? Think about what we do in Purity to be able to provide that same white-glove, world-class treatment of the data. If we don't have that, the intelligent control plane doesn't exist, right? It completely falls apart. Yeah, I think it'll be a fun discussion. We get a lot of cool innovation. Yeah, I know. It's going to be super fun. Hey, I always like to put a stat in the episode, and I actually looked at what we did last year for this thing. And I was like, you know what? It hasn't changed. Like, I went and looked and researched it. It was like 45 percent of teams are still spending four to five hours per week managing scripts and managing things that they do, to your point, around workflows, right? Managing workflows and doing scripting. There's a better way, but also some of it falls down because of the silos and the copies of the data and the sprawl and the proliferation. Yeah, I think when I look at just how painful storage is to manage, I kind of feel bad for a lot of folks who are dealing with it because it's not easy to do big policy changes, and it's not easy to remember even what things were configured like. And it may not even have been the team who deployed it. It could have been some other guy who was there before. And so having no understanding of the prior intent or the current intent makes it very difficult for you to manage anything. And so this intelligent control plane really took kind of a holistic view into first breaking down the silos, second, starting to truly manage the data more, right? Not just think about LUNs and volumes and file systems, but trying to group them together with policies. But open that up for visibility, too, so that people can understand what are our policies? That's a first order question you kind of have to deal with, right? What is our policy? So we're starting to build things that help them rationalize what they've done today, so they can figure out what they should do in the future. But then take that policy and then start applying it even to older, call it brownfield deployments, alongside doing it alongside the greenfield deployments, so they can get value out of this intelligent control plane as well. Yeah, it's really a shift from tactical to strategic, right? It's the type of thing that we always wrestle with in our jobs as well, right, is finding time to do the strategic. And what we're enabling is that strategic. And we've even had customers come to us. I had a couple of notes that came in to me from one of our large customers that's here that's all in on Fusion. And they'd consumed all kind of the basic level, like training and certification and things that we do. And they said, we're in, we're good. Teach me how to take this to the next level. How do I go deliver as a service to my internal business units? And then you see those connections made to your point about what are our policies? What does this business unit need? What are they trying to achieve? You start elevating IT into a true partner with those business units to help them achieve what they need to do. Yeah. One of the things we're building that was obviously not part of this launch, but something we are helping customers with already is just rationalizing what's happening today. Like, I think some people, when they look at like, what's just my snapshot policy for frequency and retention, it is all over the map. And they thought maybe they had two or three. And so part of this is taking what has been occurring and really helping them try to figure out how they rationalize that in a more strategic way to deploy this the right way in the future, I guess. And then be able to keep those compliant over time, which is the big one. Yeah. Which leads into some of the really new, exciting things. You ready to do this? You want to jump in? Because I know you love, and again, testament to all the engineering and the product management that goes in. Like, we're celebrating launch. Thank you to all those people. Thank you also to our customers for providing feedback. We get great feedback and RFEs and from customers going, oh, this is really cool. What if you could do this? And that's at the heart of engineering, but also being super, truly customer first. So relative to the intelligent control plane, this is an evolution, right? We continue to build upon this, right, when we came out and kind of did, okay, fusion is impurity and here we go. Yes, there was a lot of functionality, but we also knew there was more that you could do. So there were three different areas around which we're continuing to innovate. The openness and that interoperability that you talked about, compliance, which turns out in the age of AI, kind of super important. And then that autonomous and those agentic features. What stands out in the open and interoperable here? We're talking about, you know, better fleet management, the ability to do telemetry, things like that. Yeah. And if you think about where we were last year, we had kind of like come up with the concept of fusing these systems together and building these first versions of policies, we call them presets. And so we could monitor a fleet, but we weren't rationalizing it in like a model of what's the topology look like, what's availability group, what regions do you have, in a very cloud-like model. And so we added that into Pure 1, so now you can kind of rationalize your environment in the way you would typically think about it, instead of it just being a bunch of cards of arrays. And this allows you to actually see systems like per region and per availabilities and all that kind of fun stuff. Some people want to build their own visibility, and so we opened up OpenTelemetry for pretty much everything. So you can take logs, you can build your own reports if you like, but Pure 1's got some pretty killer ones, and I will tell you, we're doing a lot of work in that space, so that's going to be a big one. MCP, we showed off a little bit last year, we showed off doing some migrations across VMware and QBert via workflows that happened to start from Copilot. We added MCP for Fusion, which was pretty killer. So now you can actually understand what's going on in Fusion, kick off workflows and other fun stuff. You'll start to see more autonomy being built in as part of this announcement, which I'll go through in the next section, but also being able to have Copilot manage some of these capabilities for you versus you having to do it yourself. Yeah, and so if people also want to get really deep into the MCP, there's an episode that will have gone up a little bit before this with Amirul, one of your PMs, and Mike Nelson, and we went in, and Amirul did a great job too, right? For a newish PM, all of a sudden he blew my mind with something that he said at the end of the episode. He created just this amazing analogy, and so for people that want to dig into that, but really it's about these are the ways that people are operating, right? MCP is a way that people operate, so we're going to provide that connectedness and use our open API and API-first capabilities to go and do that. Okay, onto the front around both intelligence and compliance, and again, the foundation that we've always had with presets and workflow orchestration, automation with the connectors. I think we covered those things last year, but now we've built, there's been this big outcry for more around compliance and monitoring, like you have to have that, that's a must-have in the IT world now. Yeah, so one thing we kind of tipped our hat to last year was intelligence around applications, right? So it was like trying to understand what's even out there. So Appear One finally got what we call now application intelligence, so it actually helps you understand things like what apps are deployed, how much SQL, how much Oracle, all those kind of fun things, and this will get better and better over time. We're going to actually add in some other analytics, as you'll see over time, but think of these as infrastructure-related analytics that will help you in all sorts of fun ways, both understanding the environment, even getting into other personas, so you can talk to, hey, this is an Oracle database, I know more about what's actually going on in the environment. We added in some kind of anomaly detection, which was pretty cool in Appear One, so you can start to see more heuristics-based understanding of the infrastructure. The one area that I'm super passionate about was we talked a little bit about the concept of applying policy and intent. What we're building in now is capabilities of ensuring, monitoring, and also remediating violations of policy. And what's killer about this is it changes the way, if you think about the model, you would traditionally go through this. So I'll give you an example, because I just talked to a customer on this. They had a massive ransomware attack, and they were trying to rationalize a policy change. They were going to go from seven days of snapshot retention to now 30 days, right? And they wanted to keep a longer retention, so in case they found out about it later, they could recover from it. And so they started to think about, like, how do I manage the change from seven days to 30 days? And they didn't want to do it on everything, that would be pretty easy. They wanted to do it on their most critical workloads. So there was a bunch of things they had to think about. They first had to go rationalize, what are all the important things? Well, that wasn't really identified, because there's no intent or understanding, it was all manually done. They then had to go build scripts to do all the changes and test those. And then they needed to go actually do it. And that was like months of work, they were estimating to actually go deploy this change of intent. In this model, when intent is attached to data or policy, you change your intent, and it will tell you all the violations of it. So what's killer about this, you could literally go and say, I'll just make this up gold and silver workloads that I've identified for my service level, I want to change the 30 day retention, it will tell me all of the ones that are violations of it, I can go change policy for all of them. And I can validate that it actually occurred and monitor it consistently. It gives me full control over the data without really having to do any manual work whatsoever. I just changed my intent, which changes the policy, and it changes it on the actual workload. What killer is that? It's amazing, because there's also that knowledge and that comfort that it's gotten done because you get that feedback. But also, let's assume this is a customer that has 10 arrays or 50 arrays or 100. That's where you get into that months of people just having to be heads down, going from every system, every system, because part of that is you're doing sleuthing of what workload runs where and is it a gold or is it a silver, and not having knowledge around that policy, like nobody knows what's running where it's just there. What would keep me up at night, too, is what if I forgot or I missed one, and that happens to be the one that gets attacked, right? It's like, oh, man. But this will help you sleep well at night. Now I can go look at any time. Do I have violations? Fix it. Yeah, and fix it fast, right? Yeah. A click of a button kind of stuff. That's great. That's super exciting. I can see that's a great analogy that you put in there, because I can see now why you get really excited about that. And also, that it's something that really is going to help IT teams, again, with that common thread of continuing to be storage and data strategists rather than tacticians and doing mundane things. Fantastic. All right, let's hit the last one around agentic, around autonomous. We've already done things, like Copilot was super cool when that came out. People going in and just interacting directly with all the data that they have. There's a few new things that are going on here that are pretty exciting relative to automation, relative to agentic. I've been playing around with things recently. I'm kind of a late adopter. It's pretty powerful. It's super cool. It's so awesome that it has access to the telemetry for that customer, right? So you can ask simple questions that would have taken me forever, like, give me the three most performing volumes. That was the first version of Copilot. But now we've added in everything. Give me a script to do this or whatever. Give me the serial number for the arrays that are doing X. By the way, it would take me forever. It would take forever. And then I'd probably look like in the wrong doc or some totally wrong information. When you start thinking about it now, we started building in workflows. We showcased a little bit of that. We're starting to get into things like triaging via Copilot, which is pretty killer. There's some fun demos you may see at Accelerate on it. The one big feature that I heard from customers a ton in the agentic world, although I'd say it's kind of somewhat connected to compliance at some level too, is, you know, when you think about I have a service level objective, let's just say for a workload, like it's a gold workload and it needs this performance. One of the things customers are looking for is I want you to move data in an autonomous way, potentially, or just recommend to move data from one array to another array when service levels aren't being met. And that could be for performance or capacity or any number of things. And so we're building in a new capability that allows you to do rebalancing and data mobility. So if you want to move a workload to one location, you can do it manually, which you can kind of do today via workflows. But we're going to start kind of looking at things similar to what we do in compliance where it's a violation, but think about it more in a performance or capacity sense to say this is not meeting its service level and I'm going to move it now to ensure it does. It's an efficiency thing. It's killer. This feature will really open up the whole autonomy side of the house and we're going to start with kind of a bit of a slider. It's going to start off with recommendations, then you'll delegate like, okay, these certain scenarios I want you to go automatically do. At some point people are going to say, screw it, we're going full autonomous and it's just going to dynamically balance the fleet for you. And I think it's going to be killer just to see the evolution of that as we build this with both compliance as well as rebalance. And we have some other cool ideas that we're not talking about quite yet, but more autonomy that we build in. And then when you think about with Agentic, you can have a co-pilot that does all the work for you. So they may just tell you, hey, here's the four things to think about today. And you say, yeah, do it. Sounds good. And then come back in a little bit and say, I did these things. What do you think? Don't need to really know much anymore, right? You know the intent and it delivers upon it. It's kind of a fun way to think about infrastructure that used to just be so ridiculously or still is ridiculously complex. It reminds me of like the early days of Pure. We'd come in and we'd talk about, we don't have raid groups. And everyone was kind of shocked by that, right? But as they found out that they didn't need to manage performance for raid types and all the other kind of stuff they had to deal with, it was refreshing as hell. They didn't even know about it anymore. We're going through a whole nother period of education around this where we're helping people automate and not think about things that they've had to. It's like history repeating itself, but for the best possible reason. This abstraction makes ridiculous amounts of simplicity. It's great. Close a little bit because you've mentioned it a few times and I want to make sure that we really shine a spotlight on it. You've mentioned the word intent multiple times and I know that's purposeful, but that degree of separation now or the connection between intent and the data, what does that really mean? And I want to make sure that because it's been a through line through everything that you've been talking about here and I think it's super important to not have data just being managed without that understanding and intent. Again, it's setting up a little bit for the ever pure data intelligence, but it's all connected. I think when you think about policy, it is the manifestation of someone's intent. And so I kind of use them synonymously because I think intent comes with more power. Policy somewhat gets diluted by its various definitions and uses in other areas. But when you think about intent or I intend to do something, the expectation is it's very declarative in nature and it should happen. I think in a lot of cases you were doing things with intent, but there was no connection to a policy to actually make sure it was enacted upon and could be then later monitored and validated. And I think that that's what we're doing here. We're effectively taking someone's intent, remembering it as a policy and then monitoring that to ensure it continues. And that's where the power comes into play because, you know, people swap out of environments consistently. Technologies change and terminology even changes. But intent, once defined, can always be monitored. And the power of which is you can change it at any point and it deploys it widespread. I think that's what we're doing here. It's we're really changing the game in the way that declarative types of models play out in an intelligent control plane. I love that. It's a good capstone to this first launch episode. What do you recommend for people? People already using Fusion now, right, to jump in. Features, functionality, some available, some coming. What about for folks that are not yet doing this, right? I mean, we really, I mean, if you're using pure, ever pure arrays, like we want to help you get going there. Yeah. So the first thing for customers is upgrade to a release. Yeah. Most recent releases has got all sorts of killer feature functionality. So get on Fusion, build yourself a fleet. That's step number one. Simple stuff, right? And then secondarily, start building these types of policies. That policy then can deploy workloads. Workloads are this new abstraction that you now can apply, you know, that intent on top of. And then you can start taking advantage of all these little, you know, new killer technologies that we're building, that engineering is doing an amazing job building for us in rebalancing and in compliance and really just understanding your data better. And I think that that's the big value that, you know, people who administer storage today is going to up their game by starting to think about data, less about volumes and file systems and the like. And there's a lot of really cool stuff that we plan to bring to market after this. This is just kind of the basics to get them going on thinking about data instead of the storage. That's awesome. It's a fun time. Yeah. This was great. I think we covered it all, too. I mean, that was really that was really the goal here. And we had fun. You always say, oh, hell yeah, I'm only coming on. I'm only coming on if we if we have fun together. It's impossible not to do that. Oh, it's always fun. Yeah. Thanks for sharing this and just walking everybody through the intelligent control plane, the advancements, the innovations. More information for everybody out there at everpuredata.com slash launch, where you can capture everything that we're doing across this three episode series. Super fun. Are you ready to do another one? Oh, yeah. Yeah, let's do it. Let's do it. My my thanks to Chad Kenny. Thank you out there for watching, for checking out this episode. Tell a friend, tell a colleague and share this out. We would love to spread the news far and wide about all the great launch innovations that were announced at our Accelerate event in Las Vegas. If you are listening or watching and are internationally, keep an eye out as well for our Accelerate roadshows that are going to be coming to about 10 or 11 cities, as well as some satellite events in other locations. So great chance to engage with our Everpure employees and hear some of this news. And sometimes we sneak in some new things into those events, which is super exciting. So look for those. Go to the events page for that information. And thanks again for listening. Thanks to my guest, Chad Kenny. And with that, we will wrap forever pure for Chad Kenny. This is Rob Ludeman saying don't look back. Something might be gaining on you. Transcribed by https://otter.ai
