 Good afternoon, guys and girls. It's theCUBE, the leader in live tech coverage, live at VMware Explorer. Day two of three days of CUBE coverage. We've had a great day one. We're having a great day two so far, unpacking all of the news that came out yesterday and then some Lisa Martin and Dave Vellante. Welcoming back, one of our esteemed alumni for his ninth visit on theCUBE, Saad's Delivery Day, the CTO, Cloud Storage and Data at VMware. Great to see you in person. Thanks for coming back. Great to be here, Dave and Lisa. Good to see you again. So some big news. VSAM Max, love the name. Announced yesterday. Gives me a big announcement with VSAM A last year. A lot of momentum going on. Talk to us about the last year and how that led to the genesis of VSAM Max. Yes. It's a little bit more than one year ago. So as you guys know that HCI is a model. A lot of people think HCI is kind of like a just storage kind of thing, but actually it's a model of how we operate. Excuse me. HCI, hyper-converged infrastructure. Okay. It's helping the audience out, you may not know. It's the concept that people want to have software-defined data center and then how do you make computer, compute networking storage all together into like one easy to manage package. And in that model it's kind of where VSAM kind of took off, where computer storage was together. It kind of made a lot of things easy for a lot of people. It just was convenient. It was also a scalar architecture. But as we went along with the last many years, in the last two, three years, we noticed that the workloads are growing. Pedabytes and petabytes of data is growing everywhere. Data's getting distributed across the globe in multi-cloud. And there's also now the ransomware threats, cybersecurity threats and there's governance, like who wants my data, where's my data. So all of these problems like kind of customers kept asking us that look, this model of computing storage together is great. But we also, our databases are getting bigger. We have use cases of analytics and we have a customer who has about 200 petabytes of elastic search and they want this model of where, they want flexibility of being able to scale, compute, storage, performance, capacity, all independent of each other. So we started off on the journey like two, three years ago, like during COVID time, we have nothing else, we focused on what should we do about this. So the foundation of VSAM is like, it's all still the same. But what we did was that we changed some of the layout of the foundation. So how we lay out those data structures on to the disks itself. We want to be forward-looking for NVMe, QLC, TLC, what else comes in the future. And so we designed this new file system. By the way, log structure file system is the name of this architecture. It was invented by Mendel, the founder of VMware. Just wanted to point that out. Really, he invented log structure file system? That's right. What year was that? 1992. It's long ago, but it's used widely across the cloud. I mean, it's a very popular thing. Come on, really? Log structure file system was invented in 1992. That's right, it was a paper from 1992 by Mendel, the founder of VMware. Before he started VMware, that was the paper he had. And it's used in storage everywhere recently. But it wasn't very like a thing people knew about it. But more recently, what's happened is that it provides you a superior way to handle any device you want. And it gives you a large-scale systems. You're able to build large things and be more stable as a platform. So that's what we did. And we kind of rebuilt some of the layout of that to actually do this log structure file system. And also, it made us become a multi-purpose storage system because we have many, many use cases we want to address. And once you have this kind of layout, you can have scale out, your software defined. You can talk to any device you want. You can put many different use cases, like simple use cases, like people want to put large databases. Go for it. You want to use it for archiving, cheap and deep. Go for it. Because we have this multi-dimensional scalability of performance, capacity, whatever. You can adjust your dimensions, whatever you want. You want cheap and deep. OK, go for that. You want high performance, large-scale databases. Go for that, too. So that's the advantage we saw. And then we did the next thing, which is basically separate out compute and storage. So somewhat like a, if you want to scale them independently, elasticity. So that's what we're calling Visa and Max. The foundation of the similar, except that we are trying to be more flexible in how we offer customers. So please don't hate me for doing this. But a little one-on-one on sort of Visa and block storage and LSF. So log structure file, right? L-S-F. L-F-S. L-F-S. Yes. OK, wait. What's L-S-F? Log structured file system. Log structured. L-S-F. The S is log structured is one word. Log structured is one word. OK, fine. We'll go with that. History. So the roots are in block storage, right? Is that correct? That's right. And the way it was originally explained to me, L-F-S, what I was calling L-S-F, is that imagine you had to lay a floor down. And you had only the same exact size floor board. You couldn't cut them. And you had to, you're doing a kitchen. I am. You had to put the floor down in your kitchen, but you couldn't cut. So what do you do? So what ended to happen, you had all this wasted space. That's right. Log structured file system, you can update in place, or any place. You don't have to go back to the exact location. That's right. But now, explain how you get around because this is so ancient, what I'm describing. But the problem was in the original paper, I'm sure. You got to do garbage collection. That's right. Now, you've solved that problem, I'm sure. Because the problem would be, when the system was empty, it performed very well. But when it got full, because of all the overhead of the garbage collection, performance went to hell in a handbasket. But then you got flash, you got NVME, you got faster processors. I'm sure that's all been resolved. So log structure file system is about writing into a log. You see, every write comes in, it's in a log format. So the performance is going to be high, no matter what the incoming input data is. No matter if it's sequential, random, it doesn't matter. It all goes into a log. That's what makes it like a high performance, stable, easy to understand file system. And it's also like, yes, you're pushing off the inline processing to doing post-processing. Either you can do, see, garbage collection, you must do some form or the other, no matter how you build your system. Either you do inline, which affects performance, or you can do offline slowly in the background when there's a little bit of idleness, you can do that. Ultimately, it's a similar idea. It's just that, when do you do the work, and how stable is the algorithm? That's what log structure file gives you, it's stable. So much more efficient. Much more efficient. And then, I know we're going back to ancient history, but vSAN has evolved dramatically since then. So, and now you're, and I said, when I saw the announcement, it's about time. It's separating compute from storage. Don't hate me for saying that, but it's not a concept that's brand new, right? I mean, the Cloud guys have done it for quite some time. Snowflake actually popularized it with its architecture. So it's a very popular and sensible, as you described, so you don't have to pay for the compute and the storage to scale together. That's right. Now, what about file and object in some of the other popular formats? That's right. So vSAN already today supports files. So with this vSAN Max with files on top of it, you can get a multi-parabyte scale file server. That's one. And Block already is there. Objects currently are partnering with the partners, but our customers want us to actually have something, kind of make it available to them. They don't want to install it, they don't want to, they want a service from us. They want an object storage service. So maybe next year we'll hear something more, but right now we are investigating how to kind of do that. It's a popular thing because in the new world with the developers, right? Developers don't think of storage the same way that you know how block storage people, VMs think about it. They have two things they want, they care about databases, that's the storage, and then object storage, that's the database, that's for them is storage. But we're not underlying all that stuff, you still need a solid foundation of this vSAN Max. To actually help you. But the developer doesn't care, is what you're saying. They don't want to see that. They don't want to know about that. They don't want to know about that. They want you to take care of that complexity. That's right. But it is a thing you must solve. Oh yeah, you have to run on something. Right. The interface is different than what the block is. Right. Talk a little bit about some of the key use cases. I can imagine like cost optimization of infrastructure applications that vSAN Max is going to deliver to customers. Yes. So we are noticing that people in this, in that have a modern new new applications they're running into. They're doing analytics a lot of analytics. It is new AI world, right? There's a lot of data, a lot of analytics. They need large databases and they also like elastic search is example, elastic search. People are putting 200 petabytes of elastic search on vSAN and that's what they want us to, they want that vSAN Max desegregation of it. And besides that, they also want us to protect against ransomware problems. See, protect is the wrong word, recover. How fast can you recover from this? These are innovations we're doing, not just to support these workloads, but also to be able to kind of go and recover because this problem is not going away. Right. Oh no, we can talk about ransomware all day long. It's a household term. It's the, we've seen so much change to the cybersecurity. The landscape is changing. In fact, when you bring up ransomware, I was reminded of Ragu's keynote yesterday. He was talking about the five innovations at VMware and rapid ransomware recovery. I think he said, I think that was like 75% faster was one of them, which is critical. Actually, it's that sometimes you can't even recover sometimes in some cases, right? Right. We're actually able to recover predictably and fast. We convert it into this complex things into a workflow with AI assisted workflow. That's what is our ransom innovations in that. And we want to bring that all to all of our storage, you know, basically no matter what customers do, we want to bring that to that. And also because it's a software-defined architecture, that we send max, the customers get to choose what hardware they want to use. Latest hardware, because Intel produces, like the industry is producing the latest hardware. They can be on the latest hardware. They can dynamically adjust the scale for the dynamically adjusting like workloads they have. Yeah. So that's what we're aiming, that's what we are trying to solve is this new modern workloads. It's a dynamically scaling, large and liquid databases, and this new AI workloads are coming in as well. They also require a lot of data processing. And you need a storage platform because you can deliver on this. What's the connectivity angle? I remember reading, I saw some benchmarks a while ago where in previous versions of vSAN, you couldn't take advantage of higher speed Ethernet, but now new versions of vSAN, you can. Can you explain that? Yeah, so it's not so much to take advantage is that there's a limit of performance you could have in a cluster, right? But now the performance is so high that you can put a 100 gig pipe and it's still like, you know, it'll saturate all the entire pipe. Okay, so it didn't make sense previously to pay up for the more expensive, right? But now it does. Also, now the network can also become cheaper. The cost per port has gone down a lot. And if people want their performance, they can always add that 100 gigabytes. If they want less, they can also add a little bit less networking if they wanted to. So cost, okay, cost performance, keep business outcomes that this is gonna enable for a retailer, a manufacturer, a hospital. Cost, I think cost is important because in the scaling of computer storage together, you're paying, there is inefficiencies there, depending on what workloads you have, it can be inefficient. The right there is a cost to think. Also, Visa and Max is significantly lower cost than currently current, the current market prices for other similar kind of products in the marketplace. With higher performance, with more forward-looking architecture and more integration with vSphere. What do you mean by more forward-looking architecture? Define that. Yeah, because we are building the multipurpose file system. You don't have to reinvent anymore. We are done with, I mean the foundation is solid now. We are adding now high-level workflows. See, customers have to do a bunch of things, right? In the daily, what workflows can we do? Like service knowledge is a good example. What do they do? They have service, they have workflows for everything. Similarly for us, right? We want to, we're providing workflows for all different kinds of use cases, including we're putting database as a service on top of this Visa and Max as a storage because people want 30% of our workloads across VMware, workloads across the customer base is databases. So clearly, people use databases a lot. And so we want to help them by providing a database as a service. But still to do that, you need the foundation of this large-scale, scalable, it has all the features, workflows to actually help you lifecycle the data, lifecycle storage altogether. So who builds that database as a service? That's obviously partners or the customer. So actually, current customers are doing themselves. So we want to help them by, we are actually announcing something. Actually, we did announce something this week. It's called Data Service Manager. It is a database as a service. And it's a managed service. The customer can manage it. In the future, we can manage it also for them. And we are picking a few vendors to partner with. We don't want to build databases. There are enough databases in this world. There are a lot of different kind of database, a lot of innovation going on there. But what we want to do is provide service. Like, provide a database as a service with no platform. But you have a database. We have, yeah, interestingly, we have actually a lot of Postgres. It's like a little DNA left over from... We have a lot of Postgres expertise. EMC acquisition. Oh no, I'm talking about Green Plum. Yes, but that team uses also Postgres. We have a lot of Postgres expertise. And Postgres is one of the leading open source software out there. Well, it is the open source alternative to Oracle. Yeah, and it's pretty solid. A lot of innovation is Postgres. I'm surprised at how much innovation is going into Postgres. And that's just one. There are many other databases as well. There's like different types of database. MongoDB, there's like Databricks. All those things, we want to enable that we bring some of these database as a service onto a platform. And to do that, we need all these storage features. And they can take advantage of this large scale like you say in Macs. Exactly. So did I see in the notes right 30,000 vSAN customers? Yes. 30,000 vSAN customers. What value will vSAN Macs unlock for them? And what's the upgrade journey or migration path like? So customers already have bought into the HCI model. They're happy with it. It's just that there are a new set of workloads which are coming in for them. They don't want to disrupt this. It's working well for them. They want to be able to expand their business use cases. The outcomes are basically customers can get much more flexible architecture and cut costs. Cost is always a factor, cut costs, and be able to support this large databases and dynamically be able to adjust your scale, shrink, up and down your storage architecture depending on the use cases. Because in this modern world, things come and go all the time, like especially some of the modern workloads. It's very dynamic stuff. That's what we are able to offer. The outcomes are flexibility, agility, cost reduction, and simplicity for the customer. It's all integrated into vSphere. It's the same experience you get. You don't have to manage separately. It's a cloud experience you want to get. What if a customer is not one of those 30,000 and they want to take advantage of vSAN Macs? What do they have to do? What's the prerequisite? Obviously they need some physical server. They need some storage. What's the model of the customer? So as you know, there are a lot of VMware customers. There are like, I don't know, lots of customers. Almost everybody has a VMware customer. So for all of them, there are multiple choices. We found that some people want to do their own. They want to buy their own hardware, like from the Lenovo or EMC, Dell, and basically be able to deploy the software of that. We made it super easy to deploy. Click button, you choose. You want vSAN Macs or I want HCI model. Okay, just very easy click. It's done. It'll automatically deploy it for you. Or there are some people, some customers who want a package solution from like Dell EMC or Lenovo. We are working with all of these partners to help deliver a fully package solution for them. And so those are the models you can, both of the models. Okay. Now what about the data play? Our hallway conversation that we were having because VMware announced private AI, sort of at least one third of the content and the keynotes was AI related. And so how does, because you were talking about the developer wanting that sort of abstraction, not needing to think about the underlying storage, just think about the data platform. So does this create new opportunities for customers to have a more consolidated or coordinated data strategy and then bring AI to that data? That's right. Firstly, we want to help with the lifecycle of all data. That itself is big. By the way, you know that, that's a big problem people have that the management of the data itself is such a big burden. We want to simplify that. We're a cloud vendor. We must simplify this journey. That's what we're trying to do. And secondly, because we're having a database as a service, that now we will, we want to be able to enable new type of databases in our, because databases ultimately are kind of like, where you store all the data and how you query things, vector databases, new LLM databases, right? That's I think what we want to provide is a platform that people don't have to worry about this. They don't have to worry about deploying this, upgrading it. No, we will do it for you. We will manage it for you. Here's a service. You just run, just attach. Do your database, ODBC, whatever protocol you want to talk about. And then on top of that is a layering of another set of services which do higher level AI services. That's also kind of similar to the database service, right? Because they're all the services. And we want to also enable that use case as well, like where we want to deploy the whole package. Like if you want to enterprise the AI, we want to provide the entire package of like, here's a custom, like, here's a, you know, like a package solution AI box for you. Like private cloud AI box, something like that. And you get all the components, don't have to worry about it. Just do your business. Focus on your business rather than these components and managing all these things. That's kind of our vision of where do we want to go. And I'm sure that's from the customer feedback you've gotten from talking with so many customers. It's exactly what they're looking for. That's right. What's been in the 27 hours since VSAN Max was announced yesterday? What's been some of the feedback? What have you heard on the street here at the show? So the, we have customers now who are, I met a customer who wants to buy an exabyte of VSAN Max. An exabyte. Exabyte. Because they have something else. What industry are they in? They're shipping industry. So they want to say more names, but I think, so we're hearing this kind of story I'm hearing where people, I mean the sessions are very well attended. We had a C type on Sunday, customer advisory board, and we present some of these ideas. And we saw almost every hand go up on every question we had. Like, you know, we want to do this. They're like, yes, I would like to have this. Because they're looking for simplicity. It's getting complicated, all this data management. So they want us to deliver on that, on their promise that flexibility, you can scale. See, no other storage system is able to provide that scale, performance, compute, storage, all this independently. So every customer is different, right? So we're providing that options for them. Last question. You're in an elevator at the Palazzo. You're going up to some party. You got 30 seconds to talk to a prospective customer who's got questions about VSAN Max and what's in it for me. What's your 30-second elevator pitch? It's, it'll help you cut costs. Number one, it's easy to manage for you. Like, agility for you. Don't have to like, you know, worry about like, you know, how is this, how am I using my resources? And thirdly, there's a lot of innovation coming soon. That will help you even do even more interesting things for you. That's the foundation of what it is. I like to jump the mic. Sasala, thank you so much for coming on the program. Your next time is your 10th time. So I'll try to get you a mug. Okay, thanks a lot. Thanks, Dave, for having me over here. I appreciate you coming and kind of dissecting what's going on with VSAN, the evolution. Thanks so much. And good luck with the continued evolution. VSAN Max, yay! Yeah! All right, for our guest, I'm for Dave Vellante. I'm Lisa Martin and you're watching theCUBE. You can find all of our content from yesterday, day one. Today, day two on demand at thecube.net. All the editorial on siliconangle.com. Next guest is coming shortly. So stick around, you're watching theCUBE, the leader in live tech coverage.