Utilizing Tech - Season 7: AI Data Infrastructure Presented by Solidigm - 08x03: Data at the Edge Needs Protection and Management with Rick Vanover of Veeam

Episode Date: April 14, 2025

As the use of AI spreads outside the cloud and datacenter, data management and protection has never been more important. This episode of Utilizing Tech, sponsored by Solidigm, features Rick Vanover of Veeam discussing the importance of data protection with hosts Stephen Foskett⁠⁠ and Scott Shadley⁠. Data protection for backup and security is familiar in the datacenter but is not always considered at the edge. But as AI and sensors push data to the edge, we have to consider how to protect and manage it. Edge AI systems combine elements of endpoint, datacenter, and cloud technologies and often have limited or intermittent connectivity, creating a differentiated platform that challenges data protection software. Mixing these capabilities to build a useful platform is the key challenge for data protection at the edge. Data protection at the edge is different, with continuity and recovery from outages and attacks more important than long-term preservation of data, and IT pros are beginning to adjust their perspectives on the requirements for these systems as well.Guest:⁠Rick Vanover⁠, Vice President of Product Strategy at ⁠Veeam⁠Hosts: ⁠⁠Stephen Foskett⁠⁠, President of the Tech Field Day Business Unit and Organizer of the ⁠Tech Field Day Event Series⁠⁠⁠Scott Shadley⁠⁠, Leadership Narrative Director and Evangelist at ⁠Solidigm⁠Follow Tech Field Day ⁠⁠on LinkedIn⁠⁠, ⁠⁠on X/Twitter,⁠⁠ ⁠⁠on Bluesky⁠⁠, and ⁠⁠on Mastodon⁠⁠. Visit the ⁠⁠Tech Field Day website⁠⁠ for more information on upcoming events. For more episodes of Utilizing Tech, head to ⁠⁠the dedicated website⁠⁠ and follow the show ⁠⁠on X/Twitter⁠⁠, ⁠⁠on Bluesky⁠⁠, and ⁠⁠on Mastodon⁠⁠.

As the use of AI spreads outside the cloud in the data center, data management and data protection has never been more important. This episode of Utilizing Tech, sponsored by SolidIME, features Rick Vanover from Veeam discussing the importance of data protection. Welcome to Utilizing Tech, the podcast about emerging technology from Tech Field Day, part of the Futurum group. This season is presented by SolidIME and focuses on the challenges of AI and data at the edge. I'm your host, Stephen Foskett, organizer of the Tech Field Day event series, including our upcoming Edge Field Day event. Joining me from Solidime as my co-host today is Mr. Scott Shadley. Welcome to the show, Scott.

Hey, Stephen, how's it going? Glad to be back. Looking forward to continuing this wonderful conversation we have with some new guests. Absolutely. And new friends and some old friends, too, as we're going to see. Very much so.

I'm very excited about our guest today. It's been a long time that we've all known each other, and it's going to be a fun chat. So I'm looking forward to it. Absolutely. Yeah, before we get into that and introduce Tim, you know, we talk a lot about AI and I don't wanna make it sound negative,

but kind of escaping from the data center, escaping from the cloud and being spread out on devices everywhere. And of course, all that AI is basically associated with and pulls in a lot of data, you know? Yeah, exactly. I was recently at an event in Dallas, Distribute Tech, which is around the power utility market. And interestingly enough, AI is just starting there, right? It's a new thing for our power utilities. And right alongside it comes things like resiliency and data protection and security. So you got these big boos with kiosks

with power substation stuff going on and there's a little kiosk saying security over the monitor, so it's kind of interesting. And of course, data management and data protection is incredibly important and that's why today we are joined by Rick Vanover from Veeam. Welcome to the show, Rick. Oh, thank you, Stephen.

Thank you, Scott. So Rick Vanover is the name. Keeping data safe is the game. I'm in my 15th year of being on the product team here at Veeam in the product strategy team, part of the corporate strategy office and product team the whole time. And as you can see behind me, I'm the curator, resident historian of the museum of sorts. And like I said, I

love keeping good data safe from bad things. That's awesome. So one of the things I love about Rick is his wonderful nickname is the Rick-O-Tron. And so if you're ever curious, you just need to reach out, find the Rickatron and then maybe he'll let you have a personal tour of the museum behind him. Yeah, it's a long story. But it started from a magazine column I used to do and they took a photo and it was an avatar, you know, like that digitization situation. And I was walking around showing it to my mom, showing it to my friends. And we

went into the movie Tron Legacy in like oh seven, eight, nine, six, something like that, and my friend said, oh that's cool, that picture looks very Tron, and it happened, and I found a font and I'm good. So many people are familiar with Veeam, I mean it's been a rocket ship, I mean it's funny with Tech Field Day field day. We've watched beam go from literally zero to I believe Don't let me I think that you guys are the largest data protection company in the business. Is that right? Yeah largest by market share and revenue through 24 numbers now there was two other parties that acquired in the space but Their mid-market and below product is going to be extracted to a different company. So we don't really know what these the next set of

numbers are going to end up with and everybody's growing Veeam's growing. Another way of saying that Stephen is I started my career at Veeam making sure people pronounced it and spelled it right and now it's a household name and data protection for sure. And tell us a little bit more, you know, what does data protection mean? I mean, some people aren't still familiar with Veeam or with the whole world of data protection. In fact, in the modern cloud-first enterprise, I've actually seen a shocking lack of understanding

of data management and data protection, high availability concepts. What, you know, not what does Veeam do, but what is data protection and data protection, high availability concepts. What, you know, not what does Veeam do, but what is data protection and data management and to the rest of the world? Well, I think everyone understands data protection and backup. That's like this lack of a better word.

It's kind of like this insurance policy. You never want to have to cash when it comes to your data. But we have nothing but evidence in the form of case studies at Veeam.com that we're in the business of protecting bad things from happening to good data. And I'll say that really things like ransomware made backup cool again, or important, or gave us a seat at the table. It's kind of an unfortunate way to become popular,

but you talk to any organization that's been through a ransomware incident that ended well, and they will end it with, and thank goodness I had a good backup. That's how the story always ends. And my thought here is, how do I get in front of that? How do I get to, you know, the reality is what you do to the left of the bang will 100% predicate what you do to the right of the bang. And whether it's performance storage for recovery, whether it's protecting everything, identifying everything, I'm just in the business of making sure people are educated and have the tools and solutions

to get there, to get out of those problems. And you know what? I'll fight the flood any day. But we drive people out of these problems every single day. And it's, it is really a good feeling. So that's, that, that makes it kind of nuclear and makes it hit home a little bit on what this segment, this industry is really all about. Yeah. And you bring up an interesting point there. Cause when you talk about, you know, where everything is and how we're managing all of it, it's, it's really not about the physicality. It's about the actual zeros and ones, right? It's about the data. The data is going to be somewhere

and how you have the ability to protect the data without loss of performance, right? IT's worst nightmare is, oh, I have to shut down a server for a backup for the next nine hours or whatever the case may be. And as we start building these bigger and bigger drives and we start pushing them further and further out to the edge, that data position is changing,

but that requirement of what you guys are doing becomes even more paramount to some extent. 100%, and I love how, I like to really phrase it on data. And Scott and Steven both kind of came to know Veeam probably around virtual machines. But the reality is it's way more than that now. It's as a service data being backed up. It's unstructured data, cloud data, backing up object storage and things like that.

I talked to a lot of people. In fact, just today we were showcasing what I believe to be one of the largest Veeam deployments ever. This one individual has hundreds of terabytes in one system and he's categorizing it. And it's just this massive unstructured data and just having some visibility and more importantly control resiliency of that. It's so important because the format changes,

whether it's virtual machines, unstructured, cloud, as a service, enterprise apps, it's just all over the place. It's kind of wild. I love it. That's really cool. I think it's an interesting idea. And so you guys have come a long way, right?

You mentioned the concept that when I, even when I first met you, it was around virtual machines, right? It's kind of where the name started, if you will, to some extent. But now we're getting more and more out into the different aspects of all of this work. So how has Veeam kind of grown with that as far as, is it just simple brute force? Have you had some fun picking up some partners and friends along the way? Have you

made acquisitions, things like that that could or will or should be coming to market that we should be aware of? Yes. No, the short story is, it's a combination of all of those things. And I guess the one kind of impact moment would be 2014 and then 2016. 2014 was our first foray out of the virtual space. So at that time, it would have only been Hyper-V

and VMware. And then in 2014, we expanded to Windows and then very soon, Linux for physical platforms. And right now today, five physical platforms, Windows, Mac, Linux, AIX, Solaris, those five, those can be protected, and then all three public clouds, Kubernetes, several software as a service apps. In 2017, we went into Microsoft 365 and out of nowhere the most deployed Microsoft 365 backup solution in the world. And then,

you know, when you think about the public cloud, you think about, you know, we're going on salesforce now and things like that. It's just there's a lot to the platform. I you know, it's like one of those how it started how it's going, right? It started with virtual machines. And right now there's very little that I say no to in regards to what Veeam can protect. And and that's actually really exciting. There's a couple of like scenarios where people have obsolete operating systems that aren't supported by that maker anymore. And I hate to say it, they got bigger problems than backup.

So when it comes to data protection at the edge, though, you know, these are unusual systems. They tend to be very small, though the data tends to be very large, especially now that we're getting more and more sensors, data-driving AI applications at the edge. They have intermittent connectivity. In a way, they look almost like what we used to call endpoint protection,

you know, PCs and Macs at the edge. But yet they have this twist that many of them are running virtual machines, many of them are running Kubernetes instances, many of them are running Kubernetes instances, many of them are running containers. It's a really weird scenario when you look at the Edge. I think that a lot of the ingredients in the Veeam shelf are appropriate for data protection at the Edge.

Are you seeing more and more demand for that need? For sure. The Edge edge I like to say You know big data can exist in little places or all data is important those you know things like that And when I look at the edge in the ed in the dawn of AI and just emerging changes Whatever it is, you know kubernetes, you know, vector databases on traditional platforms, or just a data center without a data center, like sometimes defining what's a data center. Well, I had one retail organization refer to a server class system that wasn't in a data center because it was something that more than one person used and that was their test for it to be a data center type of asset, which I'm like, well, that's interesting.

But all that being said, it's all about finding the data and protecting it. And like I said, there's very little that we cannot protect now, especially when you think about training data. If you think about vector databases, and if you look at any large multinational or enterprise org, there's going to be pockets of innovation. Could it be the new shadow IT?

I don't know. Could they be looking at AI powered solutions and Kubernetes, which you can run on your PC? It's kind of scary. You know, things like that can become these new data lakes, data puddles, data pockets. And my thought is it's all about protecting everything, even if it needs, you know, and my thought is it's all about protecting everything,

even if it needs higher level protection or more powerful protection, because this is the rake that IT has stepped on forever when development was actually production and nobody knew it. I'm in the business of trying to solve that problem, right? So the edge needs love too. That's the shortest way to solve that problem, right? So the edge needs love too.

That's the shortest way to answer that question. That's a very interesting point because as we're talking about this season is all about AI and the edge and we're kind of here at SolidIME, we've got these nice large drives. We also got these super fast drives because we're trying to help figure out how to articulate around all of these same innovation pockets you're talking about because we've got petabytes of this raw and structured data and we need to turn it into petabytes of usable inference data. And how do we go through that pipeline and protect that pipeline is something that we're

kind of looking at and working out as an organization effort, partnerships, things like that to help solve those problems. And so when you talk about it, I love the comment you made, what is a data center, what is Edge? I mean, we could spend an entire season on who defines the Edge as what. So I think we did spend that season, didn't we? Exactly. Well, and I would add that it depends.

And I hate to use that phrase. But the reality is no two environments are different. One thing I can say is that even though there's this common stack or set of stacks that a lot of organizations use, it's all about how they use it. It's not what, it's how. And that is why it depends so often. But I love the fact that because of what, that also predicates how you can protect it but that also to Scott's point I'm a big fan of modern storage of any type but on Scott's you know I talked with his team kind of ahead of this and like what

they're doing on a storage system level really provides you know this what but it also influences the how so how fast can it be so the what and how you know? What is it? How do you protect it beam? What is it? How does it perform? Well solid I you live in this world when you think about it that way you're kind of ready for anything And we just got to make sure that you know big choice platform changes don't come in to mix you know I Don't want to see people kind of coming up with bad sideways crazy ideas, but AI type systems are one that are going to be very common and mainstream, so I'm ready for it. Learning data, training models, vector databases, bring it on, protect it, break it, I can help.

Yeah, and I think that, to my point back to my point about ingredients, you know, if I look at what is necessary at the edge, you know, some of the things are pretty challenging. You know, we talked about intermittent or slow connectivity. You need to have an optimized data path and optimized storage in terms of not having to dump the entire data set across the wire every day or every whatever. You know, you look at the application integration requirements.

You know, you need to be able to flexibly protect virtual machines. You need to be able to understand Kubernetes environments. You need to be able to flexibly handle, as you said, structured data, like database data, as well as unstructured data, just big blobs of files. And all of these things are things,

I think, that the data protection industry has been not struggling with, but basically attacking for decades. not struggling with, but basically attacking for decades. I remember early in my career, seeing storage and data movement optimization was a big thing. In the basically the last couple of decades, it's been all about application and platform integration and figuring out ways of creating a useful backup,

not just a dump of data. And all of those things, I think, are kind of coming together in this cauldron at the edge. And I think that that's gonna be one of the things that's gonna challenge many companies that are trying to do data management, data protection at the edge, because in many cases, they won't have those ingredients.

You know, maybe they've never encountered structured data. Maybe they've never encountered structured data, maybe they've never encountered virtual machines, but virtual machines are widespread. So how do you mix all of these different environments together and still have a useful platform? The mix is hard. And what I mean by that is organizations are rapidly growing their stacks. So back to the what are they using? I live in a world where we have this balance of we need to keep the portfolio growing across these things. Like I said, there's very few that Veeam says no to that we can protect.

But the reality is there are some, especially around niche cloud services, additional SaaS services, but that's an incredible growth area for the market. So there's this balance of let's get some horizontal coverage, but then this pressure for innovation and deeper integrations kind of vertically down, right? And platforms like the hyperscale public clouds, the virtualization platforms, and to an extent,

traditional operating systems give really good plumbing to work with APIs and things like that for data movement, especially when things go bad. But it's a balance to, you know, I hate to say check the box, but that's where you start. And when you go into some of these niche services, that's where kind of the what matters, the platform. So again, AI comes up a lot, vector databases,

where do you run those? How do you run them? What are you running them on? Well, if you run them on these other things, you can have really incredible performance when it comes to backup and recovery. Next week is Vmon. We're thinking about one of the demos will be an instant recovery

of a vector database, and guess what? That's actually, I want to say, three years old. Or we can run it in Kubernetes and things like that. We presented that at KubeCon last year. So what matters? What platform and then what plumbing is exposed to you to then predicate how you get out of these problems and things like that?

Yeah, that's a very interesting point because I mean, you think about just a month ago, we had NVIDIA GTC and you have Jensen coming out and saying, this is the Super Bowl of AI and everybody wins, but he's talking about all this processing and this concept of an AI factory But there's a lot of things that you have to do to protect the factory that weren't even brought up I mean this the show was all about how to cool things really is what it was almost like an air conditioning event

And it's like well when you think about it there's a whole other concepts of how to manage that AI pipeline and and Footprint and where it sits and how it sits. There's all this stuff going on around it, and a lot of people do kind of tend to overthink how they protect some of that information and that data. Because you think about, like in our world, we're talking about how to make drives last longer

or be reusable. And so instead of shredding a drive, how can we fix that? And those are physical issues that we deal with while you get the joy of dealing with the real-time data resiliency around certain things and platforms. You say overthink. I don't want to confuse that with overlook because I see a lot of people doing that too. And with these new platforms, I have nothing but experience of people like, you know, that whole phenomena.

It was dev, but it's actually production and now we need it, right? So just a warning to the audience, you know, as we embrace new platforms, everything from compliance and you know, that stuff, which is less fun, but just a primitive getting out of a problem. How can it break? Trust me, things can break. User error, fire, flood and blood, you name it. So I live in a world of making sure those types of things end well. Well, that's, I think, a key point too, when you're talking about Edge.

These are not friendly environments. Things can break in the data center, things will break at the Edge. You know, I mean, we like to laugh about having the servers, as such as they are, sitting underneath the fryer or in the stock room or in the windmill, in the military truck, whatever it is, this is true. And these things will break. And I think that that's another aspect.

I mean, if I wanted to make the devil's advocate one of the. One of the things that i've heard when the philosophies i've heard from the edge focused architects out there is they want to design systems that are essentially. You know disposable and automatic because they don't have the other thing they don't, because they don't have, the other thing they don't have is they don't have any personnel on site. And so they have to make their systems, you know, the out of box experience has to be absolutely perfect.

They have to make these systems that literally, you know, somebody can come off the grill, open the box, plug it in, and go right back into the kitchen. That is an unusual situation for data protection because we're used to basically having these, well, pets versus cattle. We're used to having pets.

We're used to having systems that we closely manage, closely monitor, that we love, and that we want to make sure are protected, this is a much more rough and tumble Wild West kind of environment. And yet the other thing that we're hearing about in the news is ransomware attacks, attacks on industrial IoT at the edge, that sort of thing.

I wonder if there's a new paradigm for data protection that doesn't necessarily centralize or even prize the data that much, but is really focused on responding to outages, losses and errors more than it is to preservation of data. You know, people are gonna be people, users are gonna be users. We used to track in some survey data. What causes outages and you'd actually be surprised from even, you know,

I.T. staff, the percentage. In fact, I think I.T. staff error was more prevalent than like natural disaster, you know, which that makes sense, but it wasn't more prevalent than hardware failure, but it's really close. Okay. So when you think about the people part of this, okay, you're out of box experience. That is the gold standard, but that's hard to do. So it depends. It gets specific quick. And I think what really will win the day is a flexible set of solutions to kind of get that outcome.

But people have to talk to other people to understand what is their expectation, because then there's this whole gap of expectations that we try to bust as well. But it depends as a short answer. Yeah, I think that's really interesting and kind of to both points. Right. So, Stephen, I love the concept of going to the physical physicality of it. I'm the physical storage guy dealing with the power outage at my house where a light pole went off. Someone hit it with the car and it lit on fire and it took six hours for power to come back. But because of a lack of proper backup, the actual cable and internet for the subdivision

took another 12 hours. So, it's like there are things that stuff where these backups and supports and things like that, where they're not even the event happens, but attached to where the event happens can have a lot of need for the safety and security. As I mentioned, I'm at a utility event where you have firewalls in multiple different forms from concrete to, you know, server-based or anything like that. So, given all of this that we've discussed so far, you know, I think that it's important for data protection professionals and companies

to look at Edge and to look at AI in a different way. What is the way that a data protection company is going to tackle the issues of AI and data at the edge? And how does that differ from the traditional way of looking at these questions? Well, Stephen, I'd say that it always depends, but here's some practical advice for the audience. I say that we're in a world that we can entertain a reimagined enterprise architecture standard. And enterprise, I use that loosely because if we're talking about the edge, whether it's a retail operation, industrial IoT, you know, think like an enterprise, even if it's

not. And the thought here is what is the wish list of an enterprise architecture standard, but always keep in mind what could go wrong. Okay, if I were to give five things that could go wrong, I'd put, you know, fire, flood and blood natural disaster. I'd put equipment failure. I'd put operator error. I would put cybersecurity slash ransomware, and then I would put organizational issues out of your control issues like a strike, a labor strike, things like that. So if you think about five, really, and then you could also put in that one other, okay,

whatever else could go wrong. If you look at an enterprise architecture standard and reimagine it for today, I am always thinking about the way to keep things running and not lose data. Now, here's the thing. If we're talking about industrial IoT and the facilities offline, okay, it's good that I have my data, but I need my facility to do something with that data. I understand the realities of that.

My thought here is don't let IT be the problem. Those are the types of things that are RGEs, right? Resume generating events. And when you as an IT staff and enterprise architecture standards group, my number one piece of advice is think of those five failure scenarios, but also work to five or more different parts of the business.

Now, here's the big nasty secret. With technology today, we can do anything. I'll throw it down. We can do anything that the business wants to keep it available, keep it performant. But sometimes, you know, I had a boss. In fact, when I was working at the one place when I first met Stephen back in 007, he used to say, hey, getting new storage isn't a problem. Getting the money for storage is the problem. But here's the thing, the storage is not a problem. So my advice to talk to five or more different parts of your business, get some stakeholder alignment. Now, here's the thing, if you talk to operations and say, look, we can protect it way up here, but we're only funded way down here, they can help you

meet in the middle to go to the CFO and get what you need. Talk to the other parts of the business. That is a very important thing. And I just throwing out five because it could be things like security lines of business. You could talk to some of your compliance folk. You could talk to other parts of your I.T. group. And I got plenty of stories of left-hand not talking to right-hand in ransomware incidents. And if they only did, they could have got out of ransomware issues quicker. They got out, but it kind of sat in the luck category.

You don't want that. And the last one is this is a weird one. This is going to sound crazy. I know that IT staff are overloaded. And so I say, build a new or, you know, reimagine an enterprise architecture strategy for the edge. But here's the thing, I have several examples of these really advanced cybersecurity threats

that come in that know when the lead architect is hiking in Alaska. So that's when they're gonna strike. So I really would invest in a lot of cross training of it live in a world of double smee. You know, that single point of failure. I know it's there's a little bit of a territory thing. Well, this is my area, I built it on the architect, I own it, talk to me, don't talk to anybody. But the reality is these types of

scenarios when something does go wrong, it will be when, you know, lead architect Joe is deep in Alaska. So those are just like five and fives right there. I just came up with that on the fly. So Rick, thank you so much. I knew that you could do it. You have an incredible depth of understanding of the data protection industry and the real world of this.

And I'm so glad that we could bring your expertise to the audience. Where can people connect with you and continue this conversation with you if they want to have that? Well, first of all, next week we're live in San Diego with VeeamON. It might be too late to make the trip, but you can join us at VeeamON.com. Attend the virtual experience for free, but I also want to draw everyone's attention to an award-winning community hub

at community.veem.com. We have a really active user community. Half of the stuff behind me is all from that. And little things like, hey, how do I? You know what? Someone will answer that question. They've done it before. So the thought is we have a really engaging community. So check out the Veeam Community Hub at community.veme.com.

Thanks. And Scott, what's going on with you this April? Yeah. So we're actually next week as well going to be joining Mr. Fosket and the AI infrastructure field day event coming up as well. So of course, that will be online and you're welcome to follow along. You might see my ugly mug on that one as well. So I'll be participating alongside a couple of my coworkers at that event. And then you can also find us at solidim.com slash AI. And if you're really bored, you can go check out LinkedIn and note that I recently presented on stage at DCC wearing a beer helmet inspired by having met folks like Rick in the past and things like that. I will definitely look that up and use that picture for blackmail

purposes. Thanks, Scott. Though I guess thousands of people saw you do it. So I guess that's yeah, blackmail. ransomware. I'll delete it. I'll threaten to delete it. So thanks. Thanks very much for joining us, Rick. And thank you everyone for listening. This podcast series is brought to you by SolidIME. We really appreciate the fact that they've been able to open up the Rolodex and bring in so many different perspectives on the challenges of AI and data at the Edge.

And I can't wait to hear more from them at the AI infrastructure event. We also, as I mentioned, are gonna be doing an Edge Field Day event soon. And I'm very excited to see what we get from that, because it's always such a different world. If you enjoy this podcast, you can find it in your favorite podcast applications.

Just look for Utilizing Tech. You can also find us on YouTube. We would love for you to leave us a review or a rating. It's great to hear from you as well. This podcast is brought to you by Solidim and by Tech Field Day, part of the Futurum group. If you want show notes and more episodes, head over to our dedicated website, which is utilizingtech.com. You'll also find us on social media platforms like Blue Sky, Mastodon, and XTwitter at Utilizing Tech. Thanks for listening listening and we will catch you next week.