 Welcome back to theCUBE's coverage of Veeamon 2022 at the ARIA. It's nice to be at the ARIA. My co-host, Dave Nicholson, you know we spend a lot of time at the Venetian Convention Center, formerly the CEN. So it's nice to have a more intimate venue. I really like it here. George Aksberg is joining us. He's the Vice President of Data Protection at Vast Data, a company that some of you may not know about. George, welcome. Hey, pleasure. Thank you so much for having me. So Vast is smoking hot, raised a ton of dough. You got great founders, hard charging, interesting tech. We've covered a little bit on the Wikibon research side, but give us the overview of the company. Yeah, if I could, please. So we're here at the Veeam show, and the theme is modern data protection. And I don't think there's any company that epitomizes modern data protection more than Vast data. The fact that we're able to do an all flash system at exabyte scale, but the economics of cloud, object-based, deep, cheap and deep archive type solutions, and an extremely resilient platform is really game-changing for the marketplace. And quite frankly, a marketplace from a data protection target space that I think is ripe for change and in need of change based on the things that are going on in the marketplace today. Yeah, so a lot of what you said is going to be surprising to people. Wait a minute, you're talking about data protection and all flash. I thought you'd use cheap and deep disk or even tape for that. Or spin it up in the cloud in a deep archive or a glacier. Explain your approach in architecture at a high level. Yeah, so great question. We get that question every day and got it in the booth yesterday probably about 40 or 50 times. How could it be all flash at an economic point that is befitting that of data protection? Yeah, what is this Ferrari minivan of which you speak? Yeah, the minivan that goes 180 miles an hour. It's really all about the architecture. The componentry is somewhat similar to what you'll see in other devices. However, it's how we're leveraging them in the architecture and design from our founders years ago in building a solution that just was not available in the marketplace. So yeah, sure, we're using all flash QLC drives. But the technology, the advanced next generation algorithms or erasure coding or rage striping allows us to be extremely efficient. We also have some technologies around what we call similarity and some advanced data reduction. So you need less capacity if you will with a vast system. So that obviously helps us out tremendously with their economics. But the other thing is I could sell a customer exactly what they need. If you think about the legacy data protection market, purpose-built backup appliances, for example, Avala, Adel, Averitas, and HP, they're selling systems that are somewhat rigid. There's always a controller and a capacity. It's tied to a model number, right? Soon as you need more performance, you buy another. Soon as you need more capacity, you buy another. It's really not modular in any way. Great cost. If you want to just keep billing the customer. Yeah, and I think at this point, Dave, the purpose-built backup appliance market is hungry for a change, right? There's not anyone that has one. It doesn't exist. And I'm not just talking about having two because of replication. It's because of organic growth, ransomware, needs to have a second unit, a second copy, and just scalability. Well, you guys saw that fatigue with that model of, oh, you need more, buy more. Oh, without a doubt. You said we're gonna attack that. Yeah, yeah, yeah. No, no, no, it's great. Without a doubt. So we can configure a solution exactly to the need because let's face it. Every single data center, every single vertical market, it's a work of art. Everyone's retention policies are different. Everyone's compliance needs are different. There might be some things that are self-mandated or government-mandated, and they're all gonna be somewhat different, right? The fact of the matter is, the way that our architecture works, disaggregated, shared everything architecture is different because when we go back to those model numbers and there's more rigid purpose-built back of appliances or maybe arrays designed specifically for data protection, they don't offer that flexibility. And I think our entry point is sized to exactly what the need is, our ease of scalability. You need more performance. We just add another compute box, what we call our C box. If you need more capacity, we just add another data box, a D box where the data resides. And especially here at Veeam, I think customers are really clamoring for that next generation solution. They love the idea that there's a low point of entry, but they also love the idea that it's easy to scale on demand as needed and as needed basis. So, Jeff, I wanna be just, I wanna go down another layer on that architecturally because I think it's important for people to understand exactly what you're saying when you're talking about scaling. There's this concept of the sort of devil's triangle, the tyranny of this combination of memory, CPU and storage. And if you're too rigid, like in an appliance, you end up paying for things you don't need. Correct. When all I need is a little more capacity. Correct. All I need is a little more horsepower. Well, you want a horsepower? No, no, you gotta buy a bunch of capacity. Exactly. Oh, you need capacity? No, no, you need to buy expensive CPUs and suck a bunch of power. All I need is capacity. So, go through that just a little more detail in terms of how you cobble these systems together. The way my brain works, it's always about Legos. So feel free to use Legos. So, with our disaggregated solution, right? We've separated basically hardware from software, right? So that's a good thing, right? From an economic standpoint, but also a design and architecture standpoint. But also an underlining, underpinning of that solution is we've also separated the capacity from the performance. And as you just mentioned, those are typically relatively speaking for every other solution on the planet, those are tied together, right? So we've disaggregated that as well within our architecture. So we again have basically three tiers, the tier's not the right word, three components that build out a vast cluster. And again, we don't sell a solution designed by a model number. And that's typically our C boxes connected via NVMe over fabric to a D box. C is all the performance, D is all the capacity. Because they're modular, you can end up like our baseline product would start out as a one by one, one C box, one D box, right? Connected again via different size NVMe fabrics. And that could scale to hundreds. When we do have customers with dozens of C boxes meeting high performance requirements. Keep in mind, when VAS data came to market, our founders brought it to market for high performance computing, machine learning, AI, data protection was an afterthought. But of those foundational things that we're able to build in that modularity with performance at scale, it befuse itself, it's perfect fit for data protection. So we see in clients today just yesterday, two clients standing next to each other in the same market, in the same vertical. I have a 30 day retention. I have a 90 day retention. I have to keep one year worth of full backups. I have to keep seven years worth of full backups. We can accommodate both and size it to exactly what the need is. Now, the moment that they need one more terabyte, we license into 100 terabyte increments. So they can actually buy it. In a sense, almost in arrears. We don't turn it off. We don't, it's not a hard cap. They have access to that capacity within the solution that they provide. And they can have access, immediate access. And without going through, let's face it, a lot of the other companies that we're both thinking of that have those traditional, again, purpose-built solutions or arrays, they want you to buy everything up front. In advance, signing license agreements. We're the exact opposite. We want you to buy for the need as and as needed basis. And also because the fact that we're multi-protocol, multi-use case, you see people doing many things within even a single, vast cluster. I want to come back to the architecture if I can and just understand it better. And I said, David Floyer's written a lot about this on our site. But I've had three key meetings in my life with Moshe and I, you've obviously know. The first we showed up in my offices at IDC in the late 1980s, said, tell me everything you know about the IBM mainframe IO subsystem. I'm like, oh, this is going to be a short meeting. And then they came back a year later and showed us Symetrix. I was like, wow, that's pretty impressive. The second one was I gave a speech at 43 South to 42 South. They came up and gave me a big hug. I'm like, wow, he knows me. And the third one, he was in my offices in Marlboro several years ago, and we were arguing about the flash versus spinning disk. And he's like, I can outperform an all flash array because we've tuned our algorithms for spinning disk. Everybody else is missing that. You're basically saying the opposite. Correct. We've tuned our algorithms for QLC. David Floyer says, Dave, there's a lot of ways to skin a cat in this technology industry. So I wanted to make sure I got that right. Basically, you're skinning the cat with different approach. Yeah, we've also changed really the approach of backup. I mean, the term backup is really legacy. I mean, that's 10, 12 years of our recovery. The story today is really about restore resiliency and recovery. So when you think about those legacy solutions, right, they were built to ingest fast, right? We want to move the data off our primary systems, our primary applications. We needed to fit within a backup window. Restore was an afterthought. Restore was, I might occasionally need to restore something. Something got lost, something got corrupted. I have to restore something. Today, with the, you know, let's face it, the digital pandemic of cyber threats and ransomware, it's about sometimes restoring everything. So if you look at a legacy system, they ingest, I'm sorry, they write very fast. They can bring the data in very quickly. But their restore time is typically about 20 to 25%. So they're reading at only 20, 25% of their write speed, you know, is their read speed. We flipped a script on that. We actually read eight times faster than we write. So I could size again to the performance that you need. If you need 40 terabytes an hour, 50 terabytes an hour, we can do that. But those systems that write at 40 terabytes an hour are restoring at only eight. We're writing at a similarly sized system, which actually comes out about 51 terabytes an hour, 54 terabytes. We're restoring at 432 terabytes an hour. So we've broken the mold of data protection targets. We're no longer the bottleneck. We're no longer part of your recovery plan going to be the issue, right? Now you got to start thinking about network connectivity. Do I have, you know, with our Veeam partners, do we have the right data movers, whether virtual or physical? Where am I going to put the data? We've really helped customer, aided customers to rethinking their whole DR plane. Because let's face it, when ransomware occurs, you might not be able to get in the building, your phones don't work, who do you call, right? By the time you get that all figured out, and you get to the point where you start, you want to start recovering data, if I could recover 50 times faster than a purpose-built backup appliance, right? Think about it, is it one day or is it 50 days? Am I going to be back online? Is it one hour or is it 50 hours? How many millions of dollars, tens of thousands of dollars will that cost us? And that's why our architecture or our thought process and how the system was designed lends itself so well for the requirements of today, data protection, not backup. It's about data protection. Can you give us a sense as to how much of your business momentum is from data protection? Yeah, sure. So I joined VAS as we were talking, chatting before I came out about six months ago. And it's funny, we had a lot of VAS customers on their own because they wanted to leverage the platform and they saw the power of VAS. They started doing that. And then as our founders decided to lean in heavily into this marketplace with investments, not just in people, but also in technology and research and development, and also partnering with the likes of VIN. We don't have a data mover, right? We require a data mover to bring us the data. We've leaned in tremendously. Last quarter was really probably our first quarter where we had a lot of marketing and momentum around data protection. We sold five X last quarter than we did all of last year. So right now the momentum is great. Pipeline looks phenomenal. And we're gonna continue to lean in here. Describe the relationship with Veeam like kind of started recently it sounds like as customer demand, but what's that like? What are you guys doing in terms of engineering integration, go to market? Yeah, so we've gone through all the traditional, verifications and certifications. And I'm proud to say that we kind of blew the roof off the requirements of the Veeam environment. Veeam was very innovative. 10, 12 years ago, they were putting flash in servers because they want a high-performing environment. A feature such as instant recovery, we've now enabled. When I talked about all those things about restore, we had customers yesterday come to us that have tens of thousands of VMs. Imagine that I can spin them up instantaneously and run Veeam's instant recovery solution while then in the background restoring those items. That is powerful. And you need a very fast high-performance system to enable that. Instant recovery is not new. It's been in the market for very long, but you can ask nine out of 10 customers to walk on the floor. They're not able to leverage that today in the systems that they have. Or it's over-architected and very expensive and somewhat cost prohibitive. So our relationship with Veeam is really skyrocketing. Actually, as part of that success in our last quarter, we did seven-figure deals here in the United States. We've done deals in Australia. We were chatting. I happened to be in Dubai. And we did a deal there with the government there. So there's no specific vertical market. They're all different. It's really driven by they have a great cyber-resilient message. And you get seen by the last couple of days today. And they just want that power, that vascism. Now, there are other systems in the marketplace today that leverage all flash, but they don't have the economic solution that we have. Your design anticipated the era that we're in right now. It anticipated the ability to scale in a variety of direction. Well, listen, anticipation, of course, coincidental architecture. It's a fantastic thing. Either way. Either way. I mean, it's a fantastic fit for today. And that's the conversations that we're having with all the customers here. It's really all about resiliency. And they know. I mean, one of the sessions, I think it was mentioned, 82% or 84% of all clients interviewed don't believe that they can do a restore after a cyber attack, or it'll cost them millions of dollars. So there's a tremendous amount of risk there. So time is ultimately equals dollars. So we see a big uptick there. But we're actually continuing our validation work and testing with Veeam. They've been very receptive, very receptive globally. Veeam's channel has also been very receptive globally, because their customers are hungry for innovation as well. And I really strongly believe VASpa brings that. George, we've got to go, but thank you. Congratulations on the momentum. Say hi to Jeff for us. We'll do so. Can I leave you with one last thought? Jeff, please do. Give us your final thought. If I could, in closing, I think it's pretty important. When customers are evaluating VAS, if I could give them three data points, 100% of customers that try VAS, test VAS, POC VAS, buy VAS, 100%. Gartner, Peer Insights recently did a survey. They do it with our blind survey, dozens of VAS customers. And never happened before where 100% of the respondents said, yes, I would recommend VAS and I will buy VAS again. It was more than two respondents? It was more. It was dozens. They won't do it if it's not dozens. It's dozens. It's not dozens. It's checking. And last but not least, our customers are speaking with their wallet. And the fact of the matter is, for every customer that spends a dollar with VAS, within a year they spend three more. So I mean, there's no better endorsement if you have a customer base, a client base that are coming back and looking for more use cases, not just data protection, but again, high performance computing, machine learning, AI for a company like VAS data. Awesome. And a lot of investment in engineering, more investment in engineering than marketing, how do I know? Because your capacity nodes aren't the C nodes. They're the D nodes somehow. So the engineers obviously won that naming. They'll always win that one. And we let them. We need them. Thank you, Sibat. Awesome. Product sales. It's the golden rule. All right. Thank you, George. Keep it right there. Veeamon 2022, you're watching theCUBE. We're right back.