Podcast Archive - StorageReview.com - Podcast #108: HCI at the Edge

Episode Date: August 2, 2022

Brian invites Jeff Ready to talk about //Scale Computing and their HCI product and… The post Podcast #108: HCI at the Edge appeared first on StorageReview.com.

Hey everyone, Brian Beeler here with the Storage View Podcast and we've got a Midwesterner on today, Jeff Reddy, who's CEO of Scale Computing. Jeff, thanks for coming in. Hey, thanks for having me. All the way from Indianapolis, I understand. That's right, that's right. Born and raised a Hoosier. So I'm here where all the big tech companies are. Yeah, that's what they say to us. Like, why are you in

Cincinnati's storage view? And I said, it's the storage central point of the Midwest. I mean, everybody knows that. You could have Boston or Austin or Denver or whatever Bay Area. I think we're coalescing some energy in the Midwest between the two of us. That's right. That's right. I agree. So yeah, we've been here since the very beginning. So we're about 14 years into the business now and we're making it work. All right. I want to hear all about scale computing. But before that, you got to talk to me about the brewery you're involved in. What's up with that? And how'd you get into, not how'd you get into beer, because beer is fantastic, but how did you me about the brewery you're involved in. What's up with that and how'd you get into, not how'd you get into beer because beer is fantastic, but how did you get into the brewery

business? Yeah, so I own a brewery here in Indianapolis called Centerpoint Brewing Company. Only distributed throughout Indiana right now, but we're working on doing more. But, you know, in my spare time from running my own startups, I help a lot of other people with their startups. I'm just a glutton for punishment that way. And I had a couple of young guys approach me about six, seven years ago now, trying to do a business plan around starting a brewery. And I myself was a big beer fan and a home brewer. And, you know, what they didn't know is before I started scale computing, I had written a business plan to start a brewery and decided, I better do tech. So I stuck with the tech, but then I had a soft spot for these guys and have been involved ever since.

Well, I mean, the waves of micro brews, breweries, I suppose, and how that's gone, sort of, there was that initial wave. But really the one where you guys jumped in six, seven years ago was a great time to get serious about it from a business standpoint. I mean, down in Cincinnati, we've got all sorts of small breweries and now little ones like Rheingeis that have gotten really quite robust, but the timing for them was about perfect, right? Yeah, no, I mean, it's been good. Obviously, you know, the COVID brought some challenges in that industry for sure. But, you know, it's funny, I've been doing, you know, startups for 30 years and, you know, almost all of my direct experience is in tech. But, you know, starting a business is not wholly undifferent from one industry to the next. I mean, instead of tech support, we have bartenders, instead of engineers, we have brewers,

we still have sales guys, you still have marketing, you know, stuff goes wrong. I mean, instead of tech support, we have bartenders. Instead of engineers, we have brewers. We still have sales guys. We still have marketing. You know, stuff goes wrong. I mean, it's all kind of the same. Well, wait, what's the tech equivalent of a bar fight? I don't know. System crash? I don't don't bring me all that hazy stuff no i mean we yeah we'll have uh you know 10 12 beers on tap at any time our best-selling beer by far though is actually a uh a blood orange ipa so it's a traditional traditional ipa fairly low alcohol it's almost session style so you can have more than more than a few um delicious uh that's that's our top seller all right well you're only two hours away and i've been known to to travel longer than that for good beer so maybe it's worth it it's totally worth it yeah all right

so perhaps we'll get back to the point of the podcast is less beer talk and more tech talk. We've known you guys for a long time. I've worked with your marketing team for, gosh, probably six, seven years now. We've done several projects with you. But for those that don't know, give us the brief hit on who is Scale Computing. Yeah, and I'm the founder of Scale. So, you know, when Scale started 14 years ago, I was in my basement and I've been at it ever since. You know,

Scale is really all about simplifying IT infrastructure. You know, our flagship product is, you know, generally bucketed as hyperconvergence. I know you guys refer to it that way, but hyperconvergence is a storage technology, as you well know, right? And that's certainly a part of the solution, but ultimately it's about running applications, right? And having a platform that's easy to run applications. Some people have said we've been doing edge computing

since before it was cool. When I started the company, I had a vision in my head to use effectively machine intelligence to detect and correct the problems that commonly happen in IT infrastructure. Right. It's patent recognition. And, you know, we know in IT, you know, you have an IT person and, you know, when somebody, a user calls into help desk or opens a ticket or opens a ticket and they say something's wrong with the database is running slow. Well, you well know, we all know, listening to this podcast, it might be the database. It might be the network. It's probably the network, right? It might be the server.

It might be anything, right? And so what happens is IT, you know, it's like going to the doctor and saying my knee hurts and demanding the solution, right? And so what happens is IT, you know, it's like going into the doctor and saying, my knee hurts and demanding the solution, right? Well, maybe you need aspirin or maybe you need orthoscopic surgery. Like, you don't know, it's a symptom, right? And so that's what IT does. I mean, a lot of their time, right, is spent diagnosing these kinds of problems. And, you know, we learned that if you were monitoring an entire system, not just any piece in isolation, but, you know, the storage, the network, the applications, et cetera, right, you can effectively see what's going on, just like an IT person would, right?

I mean, so how much, you know, here's the weird thing that happens in IT, right? You say how much free space is on the drive in server number two. And sometimes that answer comes back and it's, you know, five gig, right? And that's fine, right? And sometimes the answer comes back and it's like less than one meg

and that's really bad, right? We all know. And sometimes the answer comes back and it's like negative e right and you're like what what the hell am i supposed to do with that right and but that that's it right you get bogus data and so you know we saw if you could like filter out the bad data and filter out you know sometimes it doesn't respond right and see what's going on holistically with system you could detect many problems in it right um and take corrective. And so the whole idea of scale was to use

that technology to provide a system which would effectively heal itself as best it could. I mean, it can't always, right, but as best it could, all in the name of keeping the applications up and running. And the inspiration for this came back to many, many years ago when I was a very young lad. My dad was what we would now call an IT guy. It was called Information Systems back then, but he was an IT guy. I thought for sure you were going to say a brewmaster and tie this back to beer. No, that won't work on that. My dad only started drinking beer after I owned a brewery, so we brought him around. But he was managing, he was managing mainframes, and, you know, I would go into the office with him on the weekend because that's when stuff breaks because it's IT, right? And I'd see the mainframes.

I'd play with the punch card machine. I'd do all these kinds of things. But, you know, as I saw that world evolve over years, I remember him, you know, talking about two things, right? One is he had a tremendous faith in IBM, right? As a company, generally, he liked the sales guys, he liked the technology, every interaction, like it was very professional. And that always stuck with me. And then the other thing was, the IBM systems themselves were super resilient. And ultimately, right, those mainframes kind of evolved into the AS400s, right? And the AS400 was known.

Like, you could hit a thing with a sledgehammer and the apps would keep running, right? And so when I set out to start Scale, and I knew we could use this technology, this machine intelligence technology, to use basically off-the- shelf hardware, commodity x86 hardware. And I wanted to build a system that was as reliable as that AS400 for running applications. Right. And then also build a company that my dad would think was as as good as as IBM. Now, I haven't asked him. I don't know if he thinks that, but that's been the goal, right, is to be that holistic solution. And knowing that,

and all along the way, we've, you know, 14 years, right? So we've seen the rise of the cloud and everything, but not everything runs in the cloud, right? And nothing was ever going to run in the cloud any more than, I mean, people still use mainframes, right? I mean, there's always,

there's a certain application. And that was sort of the distance. So scale, you know, it's a long-winded answer, but the backstory I think is relevant, right? Scale is a platform for running your applications, right? And it's a combination of hyperconvergence, of virtualization, of this kind of automation and self-healing, and really, you know, meant to be super easy to manage.

I mean, I think if there's one thing that stands out about scale is it's that manageability from sort of day one out of the box to what happens when something goes wrong. Well, talk about a little bit more on the timeline of your evolution, because as you sit today, I think most really know you as an HCI company and for better or worse but that's kind of how you guys have been categorized I think in in the market 14 years ago HCI wasn't a thing and it wasn't really a serious thing until maybe two years into to v-san or maybe a year and a half into Nutanix where the industry looked at that and said, OK, well, maybe it makes sense, especially at Edge or many small businesses and other spots to mush all this together so it's easier to manage.

And now we've gone on to looking at entire data centers that are software defined and there's all sorts of other concepts. But in that 14 year to maybe six or seven years ago window, when you're out there talking about some of these convergence topics and the only thing that's being sold are giant converged infrastructure solutions from VCE back in that window or early FlexPod, I suppose might be close. But how were you even trying to communicate what you guys were doing when the notion didn't really exist? Yeah, that's a great question, right? When we started out, again, this foundational technology around this machine intelligence and so forth has been,

I mean, that was the original tech, right? and it was designed to do this self-healing the the first product um that we brought to market was at that time um when this was all spinning disc world right this is a long time ago right all spinning disc it was a scale out ice guzzy SAN, NAS product, right? And it was storage, just pure storage. And so, you know, people would look at that and they'd say, well, for the first, you know, four years of the company's existence, I had the stupidest name for a storage company ever conceived, right?

Scale computing, the storage company, right? Like why, but you know, again, the vision was always this kind of full integrated infrastructure, right? I mean, again, the vision was always this kind of full integrated infrastructure. Right. And we didn't there was no converged infrastructure wasn't even a term then. Right. It was just this idea that we could do the self healing. And, you know, I admittedly thought, right, that, you know, when we started out, well, we might, you know, the storage, storage is hard, right? And so that was where we started with the tech, and it was very useful to have this kind of self-healing storage system, right? And as things evolved, and I saw the opportunity, admittedly, maybe 10 years early, but I saw that, you know, I mean, I saw that VMware wasn't going to be the dominant player forever, right? In terms of effectively, you know, I'll use the term loosely, but an operating system that runs your apps, right?

I mean, that's basically what it is. And I knew that would be the case. And I saw KVM as an open source alternative. And as that started to come on the scene, that's when i saw the opportunity to say okay well now that this is open source right we can actually integrate that into the storage product um and and have full control over it right that was the key right there's there's nothing wrong with vmware but my self-healing system like i don't know what's vmware is a black box right to a developer on the outside

and use apis and you get to it but that that's all you can get to so you don't know what's going on inside with kvm and open source i could see everything plug that into the system and now we said okay well we should just run the we should run the apps on the same devices that are running the storage right and that you know it's, you always think in hindsight, well, that seems obvious. But if you rolled the clock back 20 years, right, or the reason you had specialized storage systems, and effectively networking boxes and stuff is because, you know, a single server with a CPU at that time, that CPU couldn't handle all these different things. And so you basically had a special server that did just storage, and we called them SANs and NASs, or same thing on networking.

Well, what I saw and we're talking 2011, maybe, timeframe, most storage systems, CPUs are sitting there doing nothing. They're sitting at 98 percent free cycles most of the time. I said, well, then why buy a server? You can just run the apps right on this thing. Like they're sitting at like 98% free cycles most of the time, right? I said, Well, then why buy a server, right? You can just run the apps right on this thing. And hence, you know, this idea of hyper convergence was, was born as a funny story, right? And it doesn't matter anymore. But the loss of the, you know, the realms of storage history when we launched our first product, which later became known as

Hyperconvergence, right, I, along with an analyst firm called Tanaja Group, we had coined the phrase hyperconverged infrastructure. And it was intended to differentiate, you'll laugh at this, right, it was intended to differentiate what we were doing at scale at this time, which was a combination of the hypervisor and converged infrastructure, hyper-converged infrastructure, right? Differentiating that from what FlexPod and VCE and specifically Nutanix were doing, right?

Because Nutanix at the time didn't have Acropolis and stuff. Nutanix was storage for VMware, pretty much, right? And so, okay, well, it's that storage. This is the hypervisor and the storage, hence hyper-converged infrastructure, right? And everybody liked the term and everybody just ran with it, and that's cool, right? So it's all good. Well, here we are now. And it's funny because now we're

getting back into this you know as it goes everything goes in circles and we're back to now wanting to disaggregate things and and take all that stuff away and be able to scale things independently and i suppose even that modality is a is an interesting opportunity for you guys and maybe having your base in open source gives you some more flexibility. Talk about that in terms of if you want to adopt new technology, whether it's GPUs or accelerators

or this disaggregated notion, how much flexibility do you have? Do you have to wait till there's a project created and enough energy around that? Do you sponsor your own projects in the open source community? How does that work? Yeah, I mean, it's, you know, I mean, it's an all of the above kind of answer, right?

I mean, there's obviously a lot of stuff in the open source community. Well, you know, we, at this point, we're big enough, you know, we work with partners that have closed source, you know, type stuff. And obviously we do a lot of our own software development, whether that's, you know, type stuff. And obviously, we do a lot of our own software development, whether that's, you know, proprietary internal stuff, or whether it's our contributions back to the open source projects, right? Ultimately, right, it's, you know, what open source software development, you know, is beyond everything else is it's just, you know, hence the name open, right? It's just out there for anyone to contribute to and see and run with

and build new projects off of what's already there, right? And so, you know, for us, it really solved this problem that we had this architecture for the self-healing, and now we can start plugging other things into it. And, you know, we've continued to evolve the business accordingly. I mean, 14 years in IT is a long, long time, right? And, you know, like I said, we were, I mean, literally like, you know, SATA spinning disks, right, was the original product.

And, you know, those are long gone and here we are. But, you know, what hasn't changed is the constant evolution of workload shifting from centralized to distributed to centralized to distributed. And you think about the data center and the cloud, which, of course, is a fancy word that means somebody else's data center, right? Or the edge, which is a fancy word that means not in the data center, right? I mean, ultimately, these are physical locations. You're going to run the applications wherever they make the most sense. And by make the most sense that, you know, that

means different things, different people, sometimes it's just what's practical to manage. Sometimes it's because you have technical reason, we all like to talk about the technical reasons, the latency and the bandwidth and this kind of thing. And that's a driver. But, you know, sometimes it's a lot more basic than that, right? I mean, yeah, I see customers who are deploying applications that are 30 plus years old. Right. I mean, the application wasn't cloud native. Right. There's nowhere near it. Right. And so, you know, if they didn't make the leap, you know, to the cloud 10 years ago, well, now it's probably not the time. Right. And so, you know, you see all these sorts of things.

So we try to embrace all of that and you know as things have happened and thankfully like you know there's you know open source is pretty much embraced by everybody now so you know whether it's nvidia doing stuff on gpus and they've got you know projects are sponsoring an open source that allow us to you know have you know better controls over how those gpus work or what have you um you know the networking stacks and everything i mean it, it's all, it all works, you know, works really well. I think the, you know, the thing that is that open source projects, I'm going to, you know, speak very stereotypically here, but are notorious for is they can

be very difficult for the average enterprise to manage, right? Like if you're not, if you're not a developer, right, these things aren't always for you. And if you think about what scale started doing back in 2010 and 2011, you know, basically unleashing KVM, clustered based KVM solutions into at that time, small and medium business, right? That sounds like a recipe for disaster, right? Like that is these are guys who, you know, are probably Windows only in terms of their own, you know, backgrounds would not, you know, have a lot of Linux experience or command line experience. And there we were out with this thing. And so, you know, the management and turning that,

the complexity of that open source stuff into something which is sort of turnkey, easy to use, I mean, that's kind of scale, right? I mean, that's what we do. And so, you know, we're always out there as these new technologies come out, like I thought about the GPUs and so forth. You know, we may not be the very first ones to get something out there with GPU because we have to fold that into our overall ease of use, right?

And so, you know, we'll go where the customer needs are, of course, but ultimately, you know, when someone uses scale, I mean, it should sort of just take care of itself, right? And, you know, point and click kind of easy. I mean, you know, we've had, you know, I've had guys publish, you know, little YouTube videos of, you know, their five-year-old daughter deploying, you know, databases on scale, right? You know, it takes her 30 seconds or something to, you know, to spin that up. She has no idea what she's doing, but, you know, it seems all good, right? But that's the idea, right?

And so now, you know, as the world, you know, I no longer have to stand on a soapbox and yell that not everything's going to the cloud, right? There was a period of time where I had to do that, and people would think I was, you probably wouldn't know, but other people would have thought I was nuts, right? Like, you know, it's all going to the cloud. Well, now we know it's not, right? And so, you know, and so edge computing is a thing. And, you know, we joke here that now, right, the one IT guy that lingered around who was good with hardware and was like, you know, the disregarded for many years is now cool again, right? The guy who knows that stuff and can deploy it, you know, on-prem.

And, you know, an interesting thing is happening now that we see when customers start deploying infrastructure on-prem, maybe it's because of edge computing or what have you, they often find that they have applications that they may have moved to the cloud almost artificially, right? Like there's a movement to take applications to the cloud. I mean, a lot of times, I mean, you know, the IT folks that listen to this show will know, you know, some decisions to move stuff to the cloud was driven by like the CFO, right? In the early days, that was almost all of them.

There were very few IT practitioners saying, I trust the cloud, take this stuff, rid me of this gear. No, like managing this gear. No, managing this gear and figuring it out was fundamental to their being. So you're right, you get typically a new CFO comes in or a CFO with a new directive to change where the money is counted and how it is spent. And so you go to this OPEX model. I mean, we're seeing some of it now, too, with the large infrastructure providers wanting to recast how they sell solutions as a service because now it's OpEx instead of CapEx. And I think the industry has not. I mean, you talk about the zeitgeist of everyone going to the cloud. The current zeitgeist is everyone's going to as a service.

But the industry really hasn't thought through that all the way yet, I don't think. And they really miss the mark on where you spend a lot of your time in the mid-market or SMB, right? Where there might be some of that, but very few are just going all as a service. That model doesn't necessarily make sense everywhere. Yeah, I think that's a great analogy because this kind of one-size-fits-all seldom actually happens. And so what we see is, if you imagine, say I arbitrarily had 100 applications, and the CFO said move to the cloud 15 years ago. So the first things you move to the cloud are the apps that make the most sense there, right? Like email.

We don't run exchange servers anymore, right? Like, thank God, right? I mean, that's… I've administered some of those. That's the worst. Oh, right. So, okay. Email's gone to the cloud.

It's never coming back, right? Like, that's off it goes, right? And other applications, there's other things that fit very well like that. And then in the middle, you've got applications where maybe it makes sense in the cloud, maybe it would make sense on-prem, you can go either way, but you move that to the cloud. And then maybe these last 20 applications, right, you started effectively forcing into the cloud because there was this fictitious utopia that you were going to get rid of all of your data center and all your on-prem gear forever, right?

And so you started doing that. Well, then all of a sudden edge computing happens, you're deploying things on-prem. And so there's these applications which they're probably more expensive to run in the cloud. Maybe they're more awkward for the users in the cloud, whatever it might be. So we now see those being repatriated, right?

Back on-prem, not because they're for edge computing, just because, well, I've got the infrastructure, and instead of paying $50,000 a month to run this database in the cloud, I can run it on a server that was $5,000 forever, right? So, yeah, we see a lot of that, and it just, it makes sense, and I think the, you know, what's interesting in our kind of deep experience in the mid-market is that, you know, often because of budget reasons or expertise reasons, right, the mid-market lags some of these trends a bit, right? So when, when, quote-unquote, everything's moving to the cloud, well, that, you know, you may not have had

the budget to even entertain that, right? And so, but here's the situation, and I'm sure this happens a lot in IT, right, where that a little bit of a laggard mentality, not because you didn't find the technology interesting, but just from a practical standpoint, you couldn't get there. Well, now turns out you are running stuff on prem again. Right. And that's that's part of what I, you know, I think is often overlooked. And again, I credit our you know, we have thousands and thousands of customers in these classic mid-market environments and you know the the reasons even for edge computing i mean i just talked about a second ago right the technical reasons the low latency and data collection and aggregation all all are real all really happen right this is why people deploy this stuff on the other hand right my my reference to the 30-year-old application database or what have you, right, there's a certain kind of blue-collar reality of IT when you're out in the trenches, which is, you know, you've got this mix, right?

Here's something you can imagine, right? You could be a construction company, a 1,000-person construction there in cincinnati right they've got 20 active job sites they've got trailers and job sites they're you know running some applications there they've got some crummy internet connection into the trailer right they've got modern you know they probably have modern kubernetes containerized app of some kind running in the cloud, that's great. Right next to their, you know, job estimating software, which is, you know, green screen, 40 years old. Right. Like and probably three IT guys. Right. Like they've got to deal with all with that mishmash. Right.

And so, you know, these kind of, you know, major, you know, one size fits all movements seldom work. I wonder what you call it now. Right. And so, you know, these kind of, you know, major, you know, one size fits all movements seldom work. I wonder what you call it now. Right. So there's the word du jour when everything was going to the cloud was all digital transformation. Right. So what do you call when you move back out of the cloud? Untransformation? No, I mean, we were just talking about this in a piece last week. I was at HPE Discover, their annual customer conference, and the tagline was something, I'll butcher it now because I don't remember, but like edge to core to cloud or something like that. And it's like, it was the same thing six, seven years ago. Like that was it.

And now I think it's just making the most of the tools that are available. And maybe cloud in your world isn't a whole, all your applications in the cloud, wholesale data center moved to the cloud, but maybe it's some degree of on-prem. Maybe it's an S3 connection to, you know, some sort of object layer for backup and recovery. Maybe you're using it for DR, but there's definitely a spectrum. And I think customers are getting smarter at picking out the pieces that are important to them and figuring out how do I use the cloud where it works. And like you said, the easy apps, fine.

Send those off and pay a seat license for your email or whatever, because you have to be a masochist to want to administer email. But it's not, the wholesale rip and replace with cloud is not, it's just not happening at a big scale, I don't think. And there are though some cases where cloud makes a ton of sense. If you've got some new AI ML thing you're working on and you don't want to stand up a quarter million dollar infrastructure in your data center because you know it's expensive the GPUs are hard to get then you have to utilize the GPUs

no one wants to spend 30 grand on a GPU and then not have it run less than you 98% of the time right so I could go to Amazon spin up my workload there learn and then tear it down without having to make that investment so there's certainly a lot of point use cases where it makes a ton of sense no that's exactly right right and you just you just put the application wherever it makes the most sense right and if there's one thing that right the cloud the rise of the cloud era did teach us all right is that there's a desire to be able to manage your quote-unquote infrastructure without you know that resulting in you busting out a screwdriver and performing surgery on hardware every other day right like that's not and so you

know when we look at you know what's different about edge computing now versus where we just had client server computing before, right? You're looking for at least some level of abstraction on that hardware, right? You want the cloud-like experience, right? Oh, I have resources that my applications will consume, right? I mean, that's what the cloud is, right? If you can create that pool of resources out of infrastructure, which is running local, without now creating a new nightmarish headache of, you know, managing hardware and all of that, I mean, that's really where, you know, scale comes to be, right?

I mean, the whole point of our software stack is to create that abstraction to allow you to have a, you know, you can think of it a cloud like experience of managing your infrastructure with, you know, but knowing that those apps are running local, right? I mean, there's a, you know, there's a company, you know, here in Indiana, not super far from from our office. And they're a specialty steel parts manufacturer, a client, one of these classic SMB company, you know, hundreds of employees, maybe 1000 employees, I'm not exactly sure. But, you know, they have these, these robotic production cells, right? So so they got these it's super cool right these robots they make steel and iron parts and they do all kinds of things and they all the

robots have all these sensors and data and they can produce massive amounts of data one robot in that factory can produce so much data that it will saturate the land right not much less like send the data to the cloud. Like, the local LAN can't handle the data that the robots can generate, right? And so... Well, the worst part about that is it's probably a bunch of text files, too.

It's not even, like, imagery or some bulky files. No, no, yeah, no. I mean, it's crazy, right? And, you know, and then it gives my blue-collar reality when I was talking to the IT director about this particular implementation. Because what they ended up doing is they deployed scale in an edge environment. And they've got these little clusters of scale similar to what you guys had played with.

That little one you've got sitting next to. I've got it sitting on the table. And they put those in each of the robotic production cells. So the robots can just basically create a mini network that's like six feet long. The robots dump their data to it. It gets processed or whatever, you know, Johnson Controls or whatever software is running on there. Then only the metadata, what they need, goes outside that mini network. That's cool.

I was talking to the director, and I'm like, okay, what do you run on there? He's like, well, for this, we only have two applications running on each of those clusters. One is this Johnson Control software, which is processing the data. And the other, we've got a Samba server, right? We spun up a Windows instance run the Samba server to dump the data. And I thought, well, I'm like, you know, again, the tech guy in me, I'm like, I could come up with a better solution than probably an old Windows Samba server, right, to hold the data.

And he's like, oh, he's like, well, the robots, not all the robots, but some of the robots, the operating system that controls the robots is Windows CE, which you probably have not heard in 20 years, right? It only supports Samba 1, which isn't even secured, right? He's like, the thing is, there's no upgrade, right? The upgrade is another $10 million robot, right? And this is that blue-collar reality of IT.

He's got a deal with Windows CE circa 1994 or something, right? Because it's a $10 million upgrade, right? And so he makes it work, right? And this is what you do, right? And this is the cool part of IT, I think. The engineering part of IT is like, okay, I got parts, I got tech, I got to make it work, right? Well, so the other thing that I think KVM gives you a great base, gives you a lot of flexibility. A lot of your customers probably appreciate the open source nature of what you guys do versus others. Helps you be presumably a little more cost effective, especially in some of the markets that we're talking about. Industrial Edge, where those guys are, they, I mean, those guys

are still struggling with the data. So you're right. You talk about that, that cluster to get the data in from, from the robots, but now I need an application that can do something with that to give me a business outcome of some sort, either higher efficiency or better parts, or, or, you know, knowing when, when one needs to be thrown away before it goes to the next step, whatever. So that's what you enable, I think, when you start looking at where is this data coming from? How do I get it in? And then what do I do with it next? So parse the data, get it to your data center, and then run your additional AI, analytics, whatever on that data to have better outcomes. So that's all cool. And you

guys can help on the affordability of some of these solutions, I think. But KVM, how defensible is that? Because I'm trying to think back in the heyday of software defined. So seven, eight years ago, there were dozens of companies, especially in the hybrid storage era, like all these guys were coming up. Some of them, you know, gosh, Maxta, Springpath, Nutanix at the time, Simplivity, like all of these guys that were coming up trying to solve some of the same problems you were, how is it that you guys were able to weather that storm on a primarily open source model when at the time and still today most of it is proprietary? Did you have KVM competitors that I just can't recall at first? And then what was different in your mentality that you're still here cranking along and Nutanix went public, but most of the other guys have sold for, did not return money

to their investors? Right. Yeah. Yeah. I mean, Nutanix has obviously done well, right? But, you know, the others that you mentioned, that's almost the graveyard of, you know, the guys that came and went. I mean the the thing is i i think that most of those companies that you mentioned that didn't make it misread what the market was actually looking for right like so you know there's an element and new tanks are at a first mover advantage and this was again pre hci being a term just the idea oh you should you don't need to use a san anymore like sans the old technology now you can use it but you know a software-defined san right i mean that that's effectively what what they had and and you know

have made hay with that and that that's great scale was always different right like i have always said my number one competitor is not Nutanix. My number one competitor is VMware, right? Because when you go into a data center and you ask the question, what, you know, what's the main, what's the operating system? What's the main stuff that you guys run here, right? Most certainly at the time, 98 percent of people would have said vmware we run vmware that runs our data center and my objective was when people would answer that for us is they would say scale we run scale in the data center um that's that's a different approach all right i'm simplifying that's a different approach than we're a storage alternative, right? Because if I was, you know, SimpliVity that did,

you know, depending on when the investors came in, some did okay and some didn't in that particular deal, right? But in the end, the answer to that question, if somebody ran SimpliVity, the answer was we run VMware, right? And SimpliVity was the storage component of that VMware infrastructure. And what I knew, right? SimpliVity was the storage component of that VMware infrastructure. And what I knew, right, again, this kind of, and this is why we had such success with the mid-market, even though, I the, you know, I don't want to disparage VMware because obviously it's a fine product,

but the bloatware that became VMware, right? And this happens a lot, right? So, you know, as a company, as somebody like VMware starts to get more and more, you know, very, you know, global 2,000, Fortune 500 customers, and they build a feature set that goes to those customers if you're that thousand person cincinnati construction company right there's a whole bunch of stuff in there that you don't use you don't need and in fact it may even complicate the pieces that you are looking for right and so you had you had that and we just wanted to simplify that right and so that simplicity was was always our our thing and you know i will freely admit that there was an era of scale where it really felt like we were

swimming upstream with my like the whole world's going vmware nobody's ever getting rid of vmware and i'm out there saying we're a vmware right? Well, yeah, I want to ask you about that specifically because you're right. I mean, if you're a VMware shop, then you're a VMware shop, right? And so for you guys to come in there and say, well, we've got some open source stuff, we've got our own sort of stuff mixed in, and we assure you that the combined thing is really cool and is performant and cost effective and easy to use and whatever else. Man, that's a tough sell into these entrenched accounts. I mean, how were you able to get people that were more open-minded or was it a cost issue? How did that go? Yeah, look, I mean, you know, it was a slog for a long time, right? Like you had to find those customers where, you know,

the administrative burden that they felt dealing with VMware was enough to open them up to the idea of there being an alternative, right? And so we, and the market where we had success there tended to be these more mid-market accounts. And so we became known as the HCI guys for the mid-market, which was fine, right?

I mean, there was nothing wrong with that. But then probably three years ago, maybe if I'm being generous, four years ago, we started to see the change, which was the edge computing thing, right? And so as customers were looking for infrastructure out at the edge, as it turns out, most edge deployments look a lot more like the Cincinnati Construction Company than they look like the Goldman Sachs data center, right? You probably have very few human resources, if any, from an IT standpoint at that location. The reliability of, I mean, think, it's the opposite of a data center, right? Power, maybe not reliable. Internet, certainly not reliable.

Heat, probably a problem, right? I mean, you know, the data centers are hot, noisy, consume a lot of electricity kind of places. You know, you can't have a server that sounds like a jet engine in the back of a coffee shop, right? You have to be able to deal with the fact that you know this retailer in the mall has you know the whole mall shares of one you know 100 megabit internet connection that goes down six times a day right like this is

the again the blue cut this is the reality right and so you know those environments need something else. And, you know, and I think this is true in any era that we've seen, like mainframes, a client server to data center and now to the edge is that the incumbent operating system that dominated that prior era is naturally what people are going to try, right? So, you know, IBM had the mainframe and then Microsoft's main competitor was IBM, right? And VMware's main competitor was Microsoft. And then my main competitor is VMware, because you're going to go out there and say, okay, I'm going to deploy infrastructure at the, you know, this grocery store, right? And I've got 10 applications, I'm going to stand up a couple of servers, I'm going to put VMware on it, and I'm going to virtualize and off and, you know, maybe some kind of storage and off it goes. Right. And it'll work. Right. Of course,

it'll work. But then something goes wrong. Inevitably, it's IT. Something goes wrong. And I'm not physically near it. I'm on the phone with the manager of the grocery store trying to get them to push the right button or turn something on. And it's, it's just that that operating system was not made for the challenges of the new era, right? There's nothing, again, there's nothing wrong with the operating system, it's just things have changed. And with edge computing, that, that change has now sort of opened up the floodgates for us at scale in terms of people's willingness to look for that operating system alternative, right? And so all of that sort of, you know, swimming upstream that we did to get thousands of customers converted from VMware to scale when we were

still talking mostly data centers. Well, now that's just like, oh, it's not just at scale as an operating system for the edge, it's that we have tens of thousands of systems out there running it, right? I mean, this is like a known thing at this point. And so it's made a big difference, right? And you know, it's just very interesting, right? I've seen customers deploy again,

what we would now call edge environments for, edge environments for long before we had that term. And, you know, one of the catchphrases that I say is that you have to remember the edge is not the data center. Right. It's just it's not it. You know, does it have servers and storage and networking and apps? Yes. Right. But but that that environment is not the same thing right and and having a system which is sort of built for that environment and the challenges that it brings is you know is effectively what we're what we're up to so so now it's become much easier right and it's the the cracks that were in the

wall regarding VMware ten years ago right are now much bigger and obviously got the Broadcom thing now, and everybody's much more open to this stuff. Well, yeah, talk about that. Because I was thinking about that too and wanted to go there next. I mean, Broadcom has made it very clear in their business model that they want to do something that you alluded to with VMware, like service our top 2,000, 1,800, whatever the number is, customers, right?

And derive tremendous benefit for those users, which is great. The further down you go in organization size or data footprint or whatever, I think it's going to be a little harder when those licenses come up. People are just going to think about it a little more, I think. And that's all you're hoping for, right? Is just think about it a little more and give us a chance to get that wedge in there. And maybe you're not going to win all the deals, of course, but maybe your percentage goes up a little bit because of the acquisition. What do you think, what's your feeling? I mean, you don't have any inside access to VMware, obviously.

But what are your customers telling you? What are you hearing for the mid-market specifically about this new ownership change for VMware? Well, look, I mean, you know, in some ways I can say we've been here before from a general IT standpoint. I mean, first and foremost, Broadcom didn't buy VMware to invest in R&D. Like that wasn't, I mean, nothing they've said is that's the motivation. It's what you just said, right? They're going to, you know, Broadcom focuses on customers. And, you know, I remember many years ago, I used to use this actually. There was an interview with the then CEO of, I was probably just HP then, I think it was Meg Whitman. And it was in Fortune Magazine, if I remember right. And she had said that, and it was sort of a,

this wasn't meant to be a statement someone could use against them, right? But you'll see why I did, right? Made a statement that was something like, I won't get it right, but directionally correct. 80% of their revenue comes from 50 customers. Right. And so I would use that and say, okay, well, what happens when your customer, 14,822, right? Like they don't care.

Right. And it's not, again, there's nothing wrong with their business. They just can't care. Right. It's impossible. Right. There's nothing wrong with their business. They just can't care. It's impossible. And so when you look at this, I'm going to bring this all the way back around to what I was talking about earlier with my dad and IBM. He worked at a construction company.

It was an industrial construction company. Went from the mainframes to the AS400. And the AS400, as I said, was adored by its by its customers i mean they're still in use today in many cases widely used in these mid-market type accounts and then what did ibm do with it they just stopped working on it right like i mean they just went away right like and that and in comes microsoft and client server and the whole deal right like and the pc and dell and all this right so so here we are again right i mean if i had you know if history is indicative of the future right that the vmware for the mid-market is just going to go away right like and it'll be there and you can always of course you can always you know force yourself to use you know a tool that tool that wasn't built for you.

But that's what really opens, you know, opens things up. And people are rightfully nervous, right? I mean, this is what many people think is going to happen, right? So as people, you know, customers who, you know, maybe look at scale before and decided, nah, we'll stick with VMware for another cycle, right? Well, when that cycle comes up now, it's like, well, maybe not, right? And they have to think, right? I mean, one of the things that will scare folks is if they haven't, you know, if they don't know scale, right? If they only know, well, there's VMware and there's Red Hat, right? You know, the learning curve and the idea of like, I'm just going to dive into, again, open, you know the learning curve and the the idea of like i'm just going to dive into again

open you know what's rooted in open source even if i'm using you know enterprise version or whatever like it's command line and all that it's not what i'm thinking about well you know again scales there to make all that easy right so you can you have the ease of use you can embrace that open source um technology and get the benefits from it but have it all packaged up with the tech support and everything that that you need right the other part of that that you know i love ibm equation right you know we're we're right there and you know i think that yeah it's it's been it makes me it brings a smile to my face anytime I look, right? Because when I see online reviews of scale, which there's many at this point, right? One of the things that often gets cited is ease of use, as well as tech support, right? How good the tech support is. And, you know, it's not, it's more than just saying, oh, we have, you know, I got better training or something for tech support. Like this is also rooted in the product itself. Like if the

product's easy for you to manage, it's also easier for my support people to manage. Therefore we can solve problems much faster and much more efficiently. But then I very intentionally, like, again, one of my swimming upstream battles I had for many years was that scale spent a much higher percentage of our revenue on support personnel than my competitors. And that was again, I had this this IBM, my dad's words ran through my head. Right. I had that in my head, head like that's what i need to do and i'm not going to do that by just outsourcing support in the standard you know at that point year you know circa 2010 sort of way right outsource it overseas or what have you now we had you know tech support you know when when 95 of my customers were in north america 95 of my

headcount in for support was in North America. Right. And, you know, and we now have customers all over the world, but it's, it's intended to have support that gives you that peace of mind. Right. When we, when I started the business and you know, my co-founders and I, when I would, when I would explain this to my investors, like, why are you spending? Cause this is the kind of thing investors look at, right? Like, oh, you're spending more than your competitors on support. And I said, well, look, you know, these guys are professional investors,

they invest in other companies. You know, if you have a company that you've invested in, and that company is a, you know, software as a service, probably, right? Some kind of software company that provides a point solution of some kind, right? If things go wrong, right, with that product, the damage is likely limited to whatever it is that product does, right? So if it's a, you know, if you had a Salesforce alternative CRM system, right?

Well, the damage is the CRM system, and that might be super painful, but it's known, right? On the other hand, customers who use scale, like we're the entire infrastructure. Like if things go wrong with us, everything is down, right? Like the whole shoot and match, all applications, the business is non-functional. And so, you know, one way that we could provide some comfort for the customers was having this super high-end support, right? I mean, we still, you know, our support number is answered by a human being within two minutes, like 99% of the time, and within 20 seconds, like 90% of the time. You'd prefer to you'd prefer to have someone answer the phone than have some crappy chat bot that makes your oh a chatbot or the phone tree that's the thing right it's like you know they're already on edge right you know

it's like we've all been there they're calling you not they're not calling to say hi everything's working great no no seldom does that once in a while you do get that one but seldom does that call come in right then something goes wrong and you know. Once in a while you do get that one, but seldom does that call come in, right? Then something goes wrong and, you know, you want to help them, you know, get going. And that's, you know, that's been our approach the whole time. And, you know, now we have a reputation for that. So with that reputation, it's a little bit easier. But, you know, it was a long time to build it up right i mean i'm i'm you know here in indianapolis i started a company i decided to take on you know the biggest incumbents in the industry um with a different approach and and then that's what's worked for us so back to your original question right that isn't what when you mentioned the other those other hci guys that's not what they were trying to do at all right and so you know it just became like well it becomes a a technology battle right and so it's like okay well this guy's got hci and that guy's got hci and where do i fit like scale is scale just feels different right scale feels different to the customer than that um and we have a you

know i'm pleased right but i always say you know we have a customer base that would make apple jealous right like i mean our customers are are you know very very active in in promoting us we get lots of referrals from customers and you know we're as big as we are now we're still a you know a little bit of an unknown player in many circles right i mean it's not you know i i sat 50 people down in the room and presented to them you know half the room didn't really know who we were at least, right, coming in. And so, you know, but I might have in those 50 people, there might be like five customers, right? And then that makes all the difference. Because now, you know, I walk out of the room and those customers, you know, talk about their experience with scale and so forth. And, you know, also, we've been able to demonstrate that we've navigated um many many changes in it right

again 14 years in the business is a lot of changes and it's not a flash in the pan kind of thing you know i you know i uh when i started the company i would often get the the feedback like okay well this all looks great right the product looks awesome i love you guys but you know what happens when you get acquired, right? Or you go out of business, you know, whatever. And at the time, right, I used to joke, okay, well, you know, when I started scale, sun was a company, right? I mean, they got acquired EMC was a thing, right? Like these guys, it's not just the little companies that get acquired. And well, now we got VMware. I'm like, well, I guess I outlive them, right? Yeah, so here we go. Right, so, you know, it's,

and I understand the fear, right? But I built the business with the intent that we had to build it with this kind of long-term view. And I don't, you know, I don't know what the business plans were of all the different guys that you mentioned, but I suspect most of them were not thinking of it that way, right?

Because it's like, well, we can build this and get acquired and so forth. Right, you look at it like that, and I don't know exactly what you guys have done differently because you do have venture capital in, so there's gotta be an incentive to repay investors, right, fundamentally. But the other guys, I think, were laser focused on

maybe not even IPO so much as acquisition by, because there's a history of it, by Dell, HP, Cisco, IBM, whoever, right? Marvell's buying a lot now. Because they had points in their portfolio that were missing some sort of coverage and storage was right to go develop those things i mean dell and emc combined made so many acquisitions extreme io equal logic compelling you know dell buying emc eventually data domain going into dmc i mean that's six right there um and the the market got spoiled on those activities. And then it kind of stopped, right? Like there were still some big acquisitions, but there were not a lot of them that were of high grade.

And when we saw some of these, like even SimpliVity, when HPE bought them, it was a relatively low number. It was not something, as we said, that was exciting. It wasn't what they were going for, yeah. But yeah, when your model is to run fast and hard and you're okay to lose money every quarter so long as you can continue to fill that funding funnel, that's fine until it's not. And then you're screwed, right right and so i think when you're

talking about what we're customers concerned about i'm sure a lot of them were concerned about is this a sustainable business and if they fold where are we going to be um but you've gotten through probably i mean it's 14 years you know some of those those uncertain times and and here you sit and i mean you're talking a lot about Edge. We've got the little cluster sitting here on these Intel Enterprise Edge devices. These are NUC-11s, I think. It's a unique play. When we set this thing up, you went through your whole service model.

We did it all remote. You guys sent us a box. I took the things out, got the power cables. Kevin plugged them in, plugged in networking. And then I think maybe we gave you guys an IP address. And that was it. Like we didn't do anything. We didn't prep for it.

We didn't do anything else. And then got on the phone with one of your white glove guys that walked through. Okay, connect here and do this and this and this. And we detailed all the steps. And we actually have a video coming out on YouTube this week. So I'll put that in the description for this for anyone that wants to check that out. But we were, I think we were fully operational in under half an hour and we did nothing. We prepped nothing other than maybe one IP address. And that was it. And we did it all remote. I mean, it used to be, it was a services contract. And I know this is a

little cluster. You guys sell rack scale stuff too, but it's the same thing, right? The software is identical. The software stack is identical. If you had a big server with 48 drives in it, it'd be exactly what's on those little boxes. Same thing, same deployment. So this was great. I mean, your guys did really well. And we are always suspicious of the, oh, it's on a three and a half by five inch note card. You'll be operational in 15 minutes

because it's almost never accurate. And it's almost always, as you said at the very beginning, networking. It's almost always networking. But this thing was really easy. The guys were flexible. I think we might have had to reschedule once for something. Cool, no problem. And really good communication via email. The other thing that you didn't mention, and I want you to spend a little bit of time on this. You talked about the passion of your customers. But you guys have one of the more active communities that you manage that I've seen.

I've seen a lot of it show up for others in Reddit. How do I do X, Y, or Z? And it'll be like some subreddit, right? With success sometimes, sometimes not. But you guys have really built this community that I think the very first thing I got was an email with a login to the community before the shipping notification, before anything else, to start to go explore that. Why is that so important to you? Well, you know, it gets back to this concept of, you know, we're there, you know, we're standing behind it every step of the way, right? Because the reality is, if you're successful at all, right, a community of users will emerge, right? So do you let that

just, yeah, I mean, you just don't let that roll on Reddit, right? Or are you going to like take an act, which you can, right? But regardless of where the community exists, are you going to take an active part, right, in that community? And, you know, and we wanted to do that such that, you know, we'll, you know, my team will sit there and answer tech support questions in the community just like we would if you got online chat or called. I mean, we're going to sit there and do that because we want the right answer for one thing. And other users will do it. of being in tech is that you very often find that customers come up with unique, it's not even like problem solving. It's like they come up with unique ways to use your product that you never thought of, right?

And so then when somebody posts in a community like that, oh, well, here's my use case, all of a sudden out of the blue another customer says oh we're using it like this and that they're even a dawn to us right that that would that would be the case and then you see you know being able to monitor that and see how people are using the product and where you know where the the bottlenecks are because again when you know it's one thing to have product management we do this of course, product management, call customers and say, okay, you know, what would you like to see changed? Right. And so forth. What you miss though,

if that's the only kind of feedback loop that you get is that there are often things which you might say, Oh, this would be really easy for us to change and be more efficient. Doesn't even occur to the customers to ask that. Right. Cause you just, but once you see them using it a certain way, you're like, oh, we could do X, Y, and Z. I mean, you talk about how easy that was to set up, and that's been our setup process for many years. And it's a huge advantage when you're talking about edge computing. And if I have to deploy 150 stores or something, having 30 minutes instead of three weeks is a big deal.

Right. On the other hand, man, if you could get that 30 minutes down to 10 minutes, that would be a big deal. Right. And so, you know, sort of the goalposts move a little bit. And, you know, the stuff that were, you know, forthcoming is, you know, you'll be able to preconfigure those kind of systems such that all you actually do is turn them on, right? And they'll reach out to the system and sort of install themselves based on how you defined it, right, ahead of time. But you can see it, again, in my

phrase, the edge is not the data center. You know, if you think about typical VMware environment, it's a Monday morning or Tuesday morning, the team meets, something needs to be done, that's going to take two minutes, right? Nobody even writes it down, right? It probably in the meeting, one of the one of the IT folks just takes care of it, right? If I have to perform that two minute task, 8000 times across 2000 sites, like you're

putting on RFP, right? Like it's a whole different animal. And so, you know, that's where a lot of our focus is now. Okay. As customers start to deploy edge environments at a greater scale, no pun intended, like how do we evolve the software to address those, those use cases? And it doesn't, you know, these don't have to be big enterprise deployment. They can be, right? But, you know, in the, I'm talking about the little manufacturing company with the robotic production cells, right?

They were a longtime scale customer, very classic scale customer. They have a data center on site, server room, server room on site at the factory. And they've got, you know, the factory floor. Well, when they started doing edge computing, all of a sudden now that one customer

has a dozen clusters deployed, right? And needs to manage a dozen clusters without walking around and getting in the way of the robots and all that. And, you know, if you take, you know, if they open up a second factory, now it doubles, right? And that's what we see.

So, you know, it's interesting. And I'm glad you had that experience of the setup because it's, like you said, this is one of those things that, you know, my sales team can say till they're blue in the face, but nobody believes it, right? No one believes it until they actually get their hands on it and try it, right?

You know, my favorite party trick at a trade show or whatever is to spin up an application on a cluster like you've got sitting there and like unplug one of them while the app is running and then watch the app move over to one of the other um servers and then plug it back in and move it back because it's sort of seeing is believing right is you know trust but verify kind of a kind of a thing um you know the other thing and you guys have had hands-on experiences that you know you've you've you have in your mind right as an it professional what a server looks like and then one of the you know these things show up and it doesn't look like a

server right it looks almost like a toy right and and the um but and i'll do this again i'll hide it right under the classic like cardboard box like or a sheet covering it up because, you know, it's got NVMe drives. It's got, you know, good networking. I mean, I can run an app that's pushing 50,000 IOPS on that thing, right. It'll solve, you know, that 90% of what customers are going to run would run on that. Right. But they, they never realized, especially when you start clustering start clustering or pooling all those resources together, right? It's not just one,

it's you've got three or five or 10 or what have you. So, you know, and again, that's always been the case with us, but Edge, the new Edge use case kind of resets the playing field in terms of, you know, opening the door for somebody like us. Yeah, absolutely. And it's been fun. We're actually messing around with a little VDI on

there now just to further push probe at the edges and see what else they're capable of and where else it can go. I do think though, especially for your larger units, you look at VDI, I mean, that might be a perfect workload where the cost to run that on a premium hypervisor and then another license for Horizon or whatever just may not be appropriate depending on the types of desktops you need to support, especially those that are more ephemeral for call center or whatever, right? Where they're just, they don't need GPUs. They're not doing CAD modeling in these things. There may be some more opportunity there. It occurs to me that we spent so much time talking about beer and history lesson and your mentality and a little bit about

investing that I don't know that we spent a lot of time on the product itself, which is fine. I'll put links to the paper we did on the the little edge solution and the video that we went through the setup video. It's like 11 minutes. It's a good view for for those that want to see how this comes together. That's that we did. Where else can people go? Check out the website to learn more. Where can we send them? Yeah, the best place to go is scalecomputing.com, right? And I can simply say this. Before you renew that VMware license or before you buy that next Dell server,

like, take a look. I mean, this stuff is out there, and it's, you know, I think when you see the stuff you guys are putting out there, it's a very, very different feel to managing your infrastructure, right? And it can give you, you know, it can solve a lot of problems. So yeah, scalecompeting.com. We have pricing right on the website. It's another thing I'd like to be transparent about.

You can create a quote and look at the price right there. There's live chat if you don't like picking up the phone, or there's 24 by 7 phone if you like talking. So we got you covered either way. All right, well, this has been great. I will definitely look you up the next time I'm in Indianapolis. I don't know if the Bengals get up there this year or not. We'll have to see if we've got a football showdown.

All right, we make it work one way or the other. All right, thanks for checking in. Appreciate it, Jeff. Very good. Thank you. Thank you. Thank you.