Max: 00:09 I just finished a conversation with John Spivey and Dave Nowicki with Expediant. And Expediant is a service provider that’s focused on modern computing for enterprises with a fun twist. So there’s lots of service providers in the market that are offering colocation, data center services, and private cloud, and, you know, this this private to public cloud, you know, hybrid environments. And what Expediant has actually done is they’ve gone to their their market and their customers, and they’ve started buying customer data centers. So large enterprises that are running data centers that no longer wanna be in the running data center business, Expedient will go out and acquire and purchase that data center from them and operate it and help that customer transform and and move down the path of of virtualization and operate operate operating without maintaining their facilities anymore.
Max: 01:00 I can’t speak for some reason. And the other thing, that they’ve done in in line with that is create what now they’re calling the Expediant Edge And this is a, custom hardware, custom r and d, ruggedized platform that their customers are using in situations where there’s either where an application still requires a high degree of, or really low latency with high availability and, and is it has some sort of constriction that prevents them from going into and using an application, traditional data center, or traditional cloud environment. It’s a fascinating conversation, something I was looking forward to doing for a long time and learning more about, and I hope you enjoy it. Hi. I’m Max Clark.
Max: 01:39 This is Tech Deep Dive. Today, I’m talking with Expediant. I have John Spivey, who’s a partner account executive, and Dave Nowicki who is a solutions architect. So AKA sales guy and nerd guy. My favorite combinations because we can talk about business and we can talk about tech.
Max: 01:54 And, this is a conversation that we spent some time to get set up and on the books. Mainly because of what Expediant’s doing around, what we’re just gonna use as edge as a placeholder. And then we’re gonna get into the details on that because edge means everything to anybody in all times. But also, how expedient is taking and helping companies transition their on prem hardware into cloud infrastructure to then back into edge infrastructure and everything that goes in between. So I’ve been looking forward to this guys.
Max: 02:23 Thank you very much for, for getting on the horn and and finding schedule before I, I go out and vacation for a while. And, you know, let’s let’s say, let me let’s just kind of jump right into it. John, let’s start with just kind of can you give us, you know, high level 10,000 foot background on Expedient? What is Expedient? Where did it come from?
Max: 02:44 What’s it doing? What’s it been up to the last couple of years? And, start there.
John: 02:48 Well, sure. And, first of all, thanks a lot for having us, Max. Really do appreciate it. And, yeah, Expediant, we’re a you know, we’re well we’re starting our our 3rd decade. We’re 22, 23 years old at this point, and, we’re a, you know, we’ll say we’re a full stack cloud service provider as you can hold on.
John: 03:10 As you can see behind me, a full stack cloud service provider. And, you know, what that means is is, you know, we’re bringing everything from enterprise cloud to VDI to disaster recovery, backup. We’re, essentially, we’re simplifying IT operations for, for our customers. So and, again, we’ve got 14 data centers and 10 markets throughout the US. So we also have colo offerings as well.
John: 03:36 And, and like you said, we do have something that is is pretty unique to us, which is our our edge offering, which is something I’m really looking forward to talking about. So very much that is that’s 30,000. That’s not 10,000. That is a really high.
Max: 03:49 Expediant’s been on an acquisition, you know, kick recently. So, what is what is that been like? And and, you know, who is it? What capabilities have you added with bolt ons?
John: 04:00 Now when you say acquisition I feel like some.
Max: 04:02 This makes it sound what?
John: 04:04 It makes it sound like we’ve been buying companies out there, and and would I don’t know of any, Dave, unless you don’t have some.
Dave: 04:12 I think it’s the data center acquisition. Right? How how are we growing?
John: 04:16 The the date? Okay. I get you. Yes. Sorry.
John: 04:20 Yeah. Yeah. We we have a very unique way, and that’s true. Yes. We are in a, acquisition mode as far as data centers go.
John: 04:29 And, it it it is a very unique situation as we’re we’re going in and looking for enterprise owned data centers. And these are data centers that 15, 20 years ago when it made sense for an enterprise to own their own data center. They built a data center and it may be 20,000, 40000 square feet. And at one time, it was either filling up or close to being filled up. And then, of course, the proliferation of VMware came along and their footprint started shrinking in their own data center and then AWS and Azure.
John: 05:00 And then now after 20 years of sitting in a corner of their 40,000 square foot data center paying taxes, paying power. They need to buy new diesel generators and then what we do is we see a situation like that and we step in and say, hey, look, we will give you a very healthy offer for your data center here. And by the way, with your own equipment, just keep it keep it where it is. We’ll do a colo deal and, and then that’s how we expand into different markets. So and and by the way, if they wanna get out of the hardware business altogether, we can just say, hey, we can, you know, we’ll take your hardware and build a cloud for you and it can either reside in that data center or wherever they want it to.
John: 05:40 So that’s that’s how we’re expanding. Our our last three markets we’ve expanded into is Milwaukee, Denver, Phoenix. We’re kind of we’ve we’ve had our data centers kinda I’ll just say bunched up in that upper Midwest area, where we’re founded. We’re out of Pittsburgh, Pennsylvania, and we’re looking to expand west and and other markets that we’re not in right now. So it’s pretty exciting.
Max: 06:03 This is a I don’t know what phrases. This has been very fascinating to me to kind of watch because when we think about, traditional data center market and operators, there’s there’s the REITs that went out and did sale leasebacks. So you saw big assets where somebody would come in and build a data center and then, you know, energize the shell and maybe start putting customers in it. And they would immediately do a sale leaseback on that actual, CapEx cost they put into it to somebody else and then and then lease the entire building going forward. And there’s some very large REITs, you know, that we all know that do this sort of operation.
Max: 06:36 You know, you go back, you know, 15, 20 years in every large, you know, you know, Fortune 500 North, they had data centers. Right? They had they had computers. They had, you know, mainframes. They had infrastructure.
Max: 06:46 And, if you’ve never been in the data center business, it’s a very expensive operation to maintain and and and do. And so, you know, you know, really couple pressures happened. Right? The first pressure was just this general proliferation of colocation and the ability to go colo, which, you know, also kind of dictated more major markets. Right?
Max: 07:03 Like, you know, there’s this idea of, like, secondary and tertiary markets not so much action in it until we had the birth of the cloud. And then the cloud became this really, like, hyper like, you needed to have a 100 megawatt plus facilities with lots of land and lots of cheap power, and then all these rural markets started coming in and becoming big big presences. But I haven’t seen really anybody in the market offer this sort of thing of, you know, hey, you know, hey, company, you know, you need to have this data center. You’re in a place where this data center makes sense for you. But it just doesn’t make I mean, it doesn’t make sense for you to keep operating.
Max: 07:33 So, like, how how did you get how did Expediant get started with this? What was the moment that really drove this? How many have you done rough numbers? Like, what’s how does this work from, like, you know, rough, like, ballpark economics? We can you can use whatever size that you want.
Max: 07:48 Well, I’d I’d imagine that most of these things have to be, you know, north of 10,000 square feet of floor space just to make it make sense. And how does this and then and how does this all get work together? I mean, I you know, in some of the scenarios I see where, like, enterprises have campuses where data center is a half a floor, and I’m asking too many questions in one shot. But I’m just I’m going through my thought exercise here because I’m excited about it. You know, it’s like, oh, we’ve got as this half of a floor that we’ve turned into a data center.
Max: 08:10 Does that then become, you know, a multi tenant environment going forward where, like, other people can show up and, you know, take equipment in and out of this facility and and do their own thing? I mean, that’s that’s this this so anyways, I’ll stop there because, otherwise, I’ll ask, like, another 15 questions before you responded to the first one.
John: 08:27 It it’s all good. And and by the way, I’ll just kinda go through my bullet pointed list of some of the the rules of acquisition for for what we’re looking for. And you mentioned, you know, starting off in some of the major the tier one markets out there. And while we don’t completely avoid them, a really fun way of looking at the markets that we are searching for is, a, a market that we’re currently not already in. But, we’re looking roughly for any market, as a rule of thumb, that can support at least 1 Major League sports franchise, but not 2 in the same sport.
John: 09:06 Okay? So New York, Chicago, LA, really San Francisco, not that interested. They’re they’re a glutton market with overpriced, you know, overpriced, properties, high taxes, etcetera. They just don’t make financial sense for us. We are looking for, like I say, a stand alone building.
John: 09:23 We’re looking for something to own. And to your point, we get this all the time. We’re not looking for an office building with an attached data center or, yeah. Our data center is on floor 3 and 4 of the 7 story building. That’s nothing that we really can use.
John: 09:43 It does need to, you know, have its own entrance and exit minimal amount of land. I’m currently while I’m based on the Dallas Fort Worth area and we are looking for a data center in the Dallas Fort Worth area. I’m currently in our Memphis data center right now, which is again a data center that we purchased from a major hotel chain. I’ll just say that. That’s the previous owner of this, data center here.
John: 10:09 And, it’s it’s actually it’s I think it’s the poster child of what we’re looking for in a data center. First of all, the grounds are beautiful, although it’s raining cats and dogs right now. I apologize for the background noise, But, you know, we’re sitting. We’re surrounded by golf courses on one side. There’s this beautiful old southern mansion that is the a sorority’s national headquarters right behind us.
John: 10:34 It’s just this bucolic wonderful scene in the middle of of Memphis. And, if we could do this with every data center we acquired from here on out, that would be amazing. So and again, we’re not looking for the 100 megawatt. We’re looking for, you know, a couple of megawatts. And it’s like you said, we we give a range of 5000 to 50,000.
John: 10:54 5000 is really at the that’s at the low end of things. I would like to think that we’d probably kinda really start getting serious about 10,000 to 50,000. I’ve been told if we go over 50,000 square feet, we’ll still continue to look at it. The economics make sense. And, so I guess those are some of the top of my head, some of the the, you know, major bullet points of what makes a good acquisition target for us.
Max: 11:17 And are these facilities that are, you know, what’s I mean, what’s driving the, you know, customer to wanna get out of this? Is it just operationally driven? Or is this the way maybe they’ve already realigned their infrastructure? Or is this, you know, equipment has depreciated and had maintenance cycle that, you know, is gonna force a lot of upgrades? You know, this is and it’s it’s just not it’s not worth it for them to service replace generators or HVAC units or or decom FM 200 and you guys come in and you do all that work and you end up with, you know, with the or they’re and they’re downsizing.
Max: 11:46 Right? And maybe they had 50,000 square feet they really only needed 5 going forward. Like, what’s what what’s what’s the actual, like, business driver that you see mostly, you know, that when you hear, like, somebody show up and say, hey. We’ve got this scenario. Like, oh, this is a home run.
Max: 11:59 Like, here we go.
John: 12:00 Yeah. To I was just gonna say the answer to what you just asked is yes. I mean, everything that you mentioned is the driver. It’s it’s typically and it’s more I’ll just say more often than not a financial driven we I as I understand it, when we are doing these deals, more often than not, we’re talking to the CFO. You know, I think a lot of times people in the IT and and the the tech department see us as some sort of existential threat.
John: 12:28 Whereas, as I would say with our cloud business, we are, you know, somebody who’s going to end up benefiting the IT department when we take over their facility and when we take over their infrastructure with a cloud environment, we move their jobs from being a, you know, being the break fix guys and spending their time spinning, you know, spinning their wheels in the mud of daily operations to turning them into a strategic part of the team where they’re building apps that are customer facing, revenue generating. So that’s why we typically deal with the CFO because he’s looking at these costs and they’re it’s a capex albatross, you know, around their necks. It’s, again, it’s taxes. It’s the power. They they do know that, hey.
John: 13:12 We’re going to need to, you know, refill our diesel tanks. We’re going to have to get new diesel tanks. Our our cracked units need to be replaced. It’s all of those things that just gonna make for a perfect financial storm for us to step in and say, hey. We know your building needs a little sprucing up, and we do that every time.
John: 13:29 We come in and we we put a really nice coat of paint on things and and, spruce things up operationally for them to give them the best, you know, the best environment for the remaining equipment to live in on top of everything else.
Max: 13:40 So we we talk about a little bit from, like, the MEP, but then going deeper into it, you know, network asset I mean, so how how much of what the what customer currently has stays, or does all of it stay? Or is it, you know, you’re gonna come in and and bring in your own transit links and routers and and patch panels and build out a MDF, IDF structure and and, you know, completely kind of reengineer and redesign how the customer’s existing infrastructure works. I mean, if there there is some level of work. Right? This isn’t like a sale leaseback operation in the pure in a sense of, like, they don’t change anything.
Max: 14:15 They’ve their footprint has to adjust and change and, you know, layouts and hot cold aisles and, like, all these, you know, cages and access control systems, right, like, have to be rethought once you go multitenant.
John: 14:26 And and I’ll lean on Dave a little bit here for this one, but I would say it really varies by acquisition. And I can tell you, for example, in Arizona, Phoenix, we we purchased a data center from a insurance organization. And as I understood it, they really wanted to get out of the hardware business altogether, and that was a transition from hardware to the Expedient Enterprise Cloud. And again, it varies from customer to customer. Dave, any insight you might have here?
Dave: 14:56 Yeah. I mean, you know, preserving what they have is key for us. We’ve gotten very adept at, going in and figuring out how to build what we need, how to freshen up those assets, but keep their workload intact and be minimally disruptive. It also helps because a lot of these engineers, facilities guys, they’re on staff with us. So the amount of, outsourcing that we have to do is pretty minimal.
Dave: 15:20 So we can get kind of novel with some of the things we have to do. But, yeah, most certainly. Right? I mean, the electrical systems, batteries, transfer switching, connectivity, A lot of that, we like to bring in as kind of priority to make sure the foundation is good. And then right that’s been months down the road, and we’re getting more comfortable with that particular client, tenant, whomever, and get to understand their needs a little bit better.
Dave: 15:44 And then we can bring in better integration, options to them than this kind of, you know, ragged tear lift and shift when when we take over.
Max: 15:52 And so, Dave, this is maybe a good good trend you know, technical transition here because part of what John’s also talking about is helping the customer get out of the data center business, but also helping them get out of the hardware business and not maintaining assets. And, you know, your I I my mom always tells me not to assume and I’ll always always kinda think of the acronym that she used, but I won’t say it right now. I think of 2 things. Right? So the first part of that I think about in terms of equipment and life cycle management is who owns the equipment, who’s depreciating the equipment, who’s servicing the equipment, who’s on support contract for the equipment, who, you know, etcetera.
Max: 16:24 Right? And the second phase of that. And and I I feel like this probably lies lays into, like, your enterprise cloud. When I hear most most companies talk about a private cloud environment now, I mean, like, I just assume that they’re running some version of VMware v cloud, and maybe it’s HCI or maybe it’s not. You know?
Dave: 16:40 Yeah. And, you know, recently, within the last couple of months, there’s been some accounting changes, at certain state and federal levels that allow even some degree of shared assets to be capitalized or depreciated by by customers. We’ve got a lot of prospects and clients asking around those questions. So that’s actually, one of the the levers that we use in design work is, you know, how does the customer spend? You know, what benefits are they looking to get that CFO group that John talked about?
Dave: 17:08 Right? What do how do they apply to this this design? You know, it usually starts off pretty mundane. Right? They stay as a Colo customer.
Dave: 17:16 They’ve got life on their their gear. They wanna realize the benefits of a better uptime because we’re gonna recapitalize all of the the infrastructure. Then it may be, hey. We’re dabbling in AWS or Azure, and you guys are a multi cloud strategized company. Help us get some connectivity there so we can leverage our existing backbone, our existing pops, and, you know, there’s nothing really too interesting with the service.
Dave: 17:39 Give them access to these on ramps that they would have had to go build and contract themselves. We’re able to turn that up faster for them. And then, right, maybe the next step is backups. There are backup agreements or their backup storage is starting to end of life or they want better features. We can very easily reach into the colo rack and offer managed backup as a service.
Dave: 17:59 And, oh, by the way, you need that immutable and multiple copies. We just turn the dial, and now you have your backup data immutable in Phoenix. Whereas before, you would have had to go collocate that or deal with those contracts. So it kind of starts that way, real real easy commoditized things that we can extend into their rack. And then ultimately, right, that’s a hardware refresh time, or they’ve got some kind of, you know, big cloud initiative that comes down from on high, and we start talking to them about multi tenant, private cloud, IS, security services, all those things.
Dave: 18:29 And, hopefully, by then, they’ve we’ve both learned how each other works, and those are, you know, very low risk operations because we’ve got the relationship.
Max: 18:37 So, Dave, I wanna I wanna hear your opinion and experiences with 2 things. So the first thing that comes up from your your comment just then of hardware refresh time. I feel like this becomes a big driver for a lot of organizations around this desire to go into cloud or cloud type services because, you know, even there’s this like weird it’s like a mental block is how it feels to me where you purchase equipment, you finance it, you pay for it in 36 months, you depreciate it over 60, you maybe extend the lifetime over the 36 months with 1 or 2 year of additional support agreements. But at some point, you know, around month 46, you know, I’m sorry, month 48, somewhere in that, like, 48, you know, plus range, you start talking about, like, equipment failure, hard drive failure, motherboard failure, you know, you know, batteries on NV RAM going out that have to be replaced. You know, like, really just kinda basic maintenance operations that go on.
Max: 19:28 But this becomes this, like, paralyzing Kryptonite within companies of, like, dealing with this, you know, thought process of, like, oh, we have to now replace all this equipment. Or what do we do with all this equipment? And versus, oh, we could just take this stuff to the cloud. We don’t have to think about this ever again. Now you have to think about it differently.
Max: 19:42 But, like, the this this this, like, a resistance around the hardware life cycle seems to be driving a lot of that decision process around. Do we maintain our own equipment, or do we go into some sort of cloud and consumption model that’s slightly different? I’m curious, you know, what you’re seeing and what your experience has been with enterprises, especially in this idea around, you know, buying buying their data center and then taking over their operations and and then, like, helping them through this this journey. Yeah. I mean, look.
Max: 20:04 I think the days are mostly long gone
Dave: 20:05 where customers, woke up or organizations woke up and either said we’re gonna go virtually forklift all our stuff to the cloud because it just doesn’t work great that way. Right? If you have a 1,000 VMs and you move those to Azure or AWS, the prices are outrageous, and the the spend is is ridiculous, right, because you don’t know variability how that accounts because everything has a taxi meter. We saw a lot of that early and up until, you know, a little bit ago. What we see with the bulk majority of enterprises that we do this data center lease buyback thing with is that they’re trying to replatform their crown jewel applications to use hyperscaler PaaS platform as a service.
Dave: 20:50 Right? They want to use serverless services. It’s a horrible term. But, right, they don’t wanna deal with an OS. The consumption model and type with these platforms in cloud is more in line with what they’re expecting to spend than hosting traditional VMs to put those services on.
Dave: 21:11 The problem is you have to replatform that stuff. So there are a couple of case studies that we provide, one of which is University of Phoenix, and their struggle with this. And it’s it’s a time game, right, because, you can’t replatform your apps fast enough. There’s just not enough manpower, to responsibly replatform in time for that hardware refresh. What we found, kind of our our value there, is that we can bring some of those traditional, you know, VMware based cloud environments to do that lift and shift at responsible pricing and kind of fixed fee billing.
Dave: 21:43 All of a sudden, now they’ve they’ve bought time to replatform their apps responsibly. We fit really well there, but I think on as a whole, that that mid market or the enterprise, they’re really staving off not just the maintenance, the care and feeding, but the hardware refreshes to try to get that that replatform, cloud migration, you know, going. And and I think they wake up a little bit when they first start to look at at that, and it’s it’s very
Max: 22:09 daunting. The the economics I mean, a lot of this stuff becomes relatively simple math. Right? You know, a lot of this breaks down into a spreadsheet. And it just is who’s carrying the cost and what care what what cost they’re carrying.
Max: 22:18 So part of the premium that you pay when you go from any of your own equipment co in any sort of cloud platform. I don’t care what you call it. PaaS, IaaS, private cloud, hybrid cloud, multi cloud, cloud cloud. Right? You know, that hardware refresh just gets baked into the monthly price.
Max: 22:31 And that service provider is now assuming a refresh cycle of whatever equipment that they’re baking into the cost that gets powered for. You know? So it’s you know, you’re not you’re not escaping the cost. You’re just maybe not aware of it, or you’re not thinking about it in the same way. So we we talk I I think about this, and it’s it’s like 2 ends of a of a of a pendulum.
Max: 22:48 You know? And, like, where have people swung from one side to the other? And, you know, your example of, you know, they wanna go into hyperscale and replatform into serverless and where do they go with that and what do they do in this whole, you know, migration and journey. And then meanwhile, I’ve got clients on the other end of this where they’ve gone completely all in on serverless and they turn around, around they say, hey, we’re spending, you know, okay. A relevant example numbers.
Max: 23:10 We’re spending 12% of our gross revenue and 30% 33% of our gross margin now on our, you know, cloud environment. And we need to have that number in order to be a viable business. And this is a conversation I’m having with the CEO and the CFO of a company right now. Right? But, like, you can’t there’s no levers for them to move inside of the cloud platform that they’re on in order to achieve that except going I don’t wanna say backwards, but, like, going back into a different consumption model where they can control their costs and they know what their their margin is.
Max: 23:36 And they can do what every business really wants, which is improving margin with scale. You know? The CEO wants to scale the revenue of the business, but he wants to fix expense. So that way his margins improve. And that’s not available to him on the cloud platform that they’re using.
Max: 23:50 Right? It’s just not the way the cloud works. And and so that becomes this really interesting dynamic where it feels like companies are somewhere, like, going like this now, you know, back and forth between these extremes all the time.
John: 24:00 Hilarious that you bring that up. I have a a slide with a an animation in it that shows the pendulum of of where we are in the cloud adoption process. And it’s it’s like, you know, over here, it was like, hey. You know, back in the day, we’re all, you know, on prem server all the time, and then it’s like, get everything to the hyperscale cloud. And then we finally came to the realization that that’s not a practical solution, and then it’s kind of swung back into the middle where we have a multi cloud, you know, solution to to talk about now that really makes sense for most people.
John: 24:34 And it’s interesting. I think that if a lot of people would spend some time upfront doing it and it’s like you said, it’s it’s basically kind of a spreadsheet, conversation that you have. And we have a thing, you know, working with partners like you where we can walk our customers through this scenario through our cloud transformation calculator. And we’re asking pretty detailed questions that are as high level of, do you have your own data center? If so, how many?
John: 24:59 Or are you have, you know, do you have colo? How much at what facility? How much power are you using? What does your IT staff look like? How many VMs do you have, etcetera?
John: 25:07 You’re walking them through this process and and talking about what their cloud transformation looks like. And and to your point, when when that edict comes from on high, it’s from a a c level person that doesn’t know that, you know, he he’s telling everybody, we’ve got to get to AWS as fast as possible. That’s the trend in our industry. We need to do that. Not knowing that many of the applications that they built that run their companies 20 years ago, that’s going to be a multimillion dollar, predicament that they get in as they try to have to refactor these applications to fit into into, into AWS or Azure or what have you.
John: 25:46 So, you know, that’s something. And once we walk them through this cloud transformation calculator, you know, it does spit out this report that says, hey. Look. If you do things with, you know, with expedient, you know, we can get you get the applications that matter to the cloud in an expedient fashion. And then those that that but I’m those that make sense to be in AWS, we can get those quickly.
John: 26:10 You can get those quickly to the cloud. And here’s what it looks like if you were on your existing path. It’s gonna take you, you know, 5 years to get 38% of your applications to the cloud, or we can effectively get you a 100% into an OPEX model cloud environment almost immediately and save you, you know, a 100 and 19,000, you know, personnel hours and, you know, $1,400,000 blah blah blah. It really spells it out for you. So that’s that’s a cool thing that we offer that helps companies upfront in their in their cloud journey.
Max: 26:42 How much of this is do you think is is this desire to move from a CapEx to OpEx model? And how much of it do you think is this still kind of lingering belief that the cloud is cheaper? Because that’s not the case.
John: 26:53 No. And and, again, most people when they say the cloud, they’re talking, you know, they they just they’re they’re talking AWS, and they are not taking into consideration the fact that there are, you know, additional people with certifications and specialized training that need to take over this role, you know, there’s that expense and then there’s the, you know, oftentimes overlooked expenses of egress fees, for example, of pulling your data out from one instance even into another instance within an AWS environment can start racking up these unknown charges. And then I will say this, occasionally people will say, you know what? Like you said, we had budgeted $15,000 for a month for our AWS spend. We’ve had our 3rd bill in a row over $30,000 Help us.
John: 27:43 And then, you know, occasionally, what that ends up meaning is we’re going to bring in a company that will manage this for you. It’ll bring the cost down a bit, but you’re still gonna be paying about a 30% increase on top of that spend. There’s no way of of getting around that, you know. So but but and unless Dave and I’m sure you see different things all the time.
Max: 28:05 My version of this is is much more extreme. You know, we see any sort of tech related business at scale and sort of, you know, in this and there you see cloud dominantly AWS, maybe 80, 90%, and then, you know, some percentage in GCP. And it really just makes it comes down on the makeup of the company where they started and how they they approach things. You know, if you’re talking from a pure compute and egress model, we’re not talking about a lot of managed services, you know, like the Kinesis and the Athena’s and mix dynamos of the world. Right?
Max: 28:34 We just look at the spend around units of compute, units of data egress. I’ve had clients spend, you know, decide to go colo, spend $10,000,000 in server equipment, put it into a colocation facility, and and, you know, cut their effective cost 87% a month. Now that’s factoring in, you know, the entire weighted load of maintain you know, buying and and capitalizing that equipment. So it’s that’s a number that people don’t believe until they see it and kind of experience it for the first time. And and, you know, the strategies and people will push you into, like, our eyes and all this different crazy stuff.
Max: 29:03 But, you know, it’s it’s I I think it’s shocking when people actually really start to understand. I mean, if you’re running steady state applications and you have a consistent predictable application load, you know, what you what you really, you know, get into. I don’t wanna I don’t wanna, like, harp to death about cloud because I really wanna talk about your other stuff. Yesterday, I had a presentation. And in the presentation, this person defined edge in a really kind of curious way to me because historically edge has been defined by the telcos of, claim around latency.
Max: 29:33 Right? Like there’s this push towards the edge which really means, you know, sub 5 Millisecond, you know, 1, 2, 3 millisecond, like, response times. And this is why we need to have equipment at every cell phone tower and and 5 g, you know, cell phones and all this different stuff. Right? Anyways, this person in the presentation defined Edge effectively as you having equipment on premise in your own facility again.
Max: 29:54 Know? And there was an example of, like, automated teller machine not automated teller machine. Parking lots. You know, these these control systems when you pull into a parking lot and shows you how many spaces are available in each floor, and you need to have your edge computing environment in the parking lot in order to, like, you know, maintain the system of figuring out, like, which spaces were green and and red and and how many spots are available. And I was like, this is really funny.
Max: 30:14 She’s just she’s she’s literally explaining of, like, putting a computer back in the parking garage to run this system, and that was a definition of edge that she used. Anyways, so, Expediant has an interesting example of edge. You’ve, you know, you’ve got your Expediant Edge. And, you know, just if I don’t know who wants to jump in here. John, Dave.
Max: 30:33 Like, let’s talk about it. Like, what is what is the Expediant Edge product, and how does this what what problem are you tackling, you know, for your customers with it?
John: 30:41 Hey. Let me let me throw this out here. Dave, if you wanna take that, I just realized I’ve got an I’ve got an example of what our edge looks like, and I can I can bring it, but I’ll step away for 2 seconds? Okay?
Dave: 30:53 Yeah. So, for us, this was kind of a natural evolution because we saw these migrations that we were doing, and there was always a subset of workload in certain verticals where our clients elected needed that thing to stay behind. Sometimes it was fear. Right? That’s the the crown jewel application.
Dave: 31:12 And it when it breaks, we know how much per hour we lose, so don’t don’t even look at that thing the wrong way. And then others, it was latency. Right? This doesn’t perform well at the end of, you know, even pretty robust connectivity because it needed to talk to printing presses or RF scanner guns or, you know, just in time shipping machinery, right, operational IT. So we had a a vested interest in figuring out a solution that would let us provide this cloud operating model, this operational thing to those those clients and prospects, but still keep or respect that application and let it live in its own ecosystem.
Dave: 31:50 Right? The parking garage example you gave is great. Although, I’d be interested in understanding that because to me, that’s something that could talk back across the Internet. But
Max: 31:58 I’m not I’m not claiming it as a as if it’s correct or not. I’m just saying that was the example that was used, and so I’ll I’ll it it’s it’s stuck out.
Dave: 32:07 But but very similar thing. Right? We we need to keep that workload close to machinery, close to physical things, and they just they don’t cloud well. But the other thing that we were seeing and talking to these companies is that it kept their IT department incredibly busy and and bothered when it came time for refresh or break fix things because, you know, you can imagine you’re a meat packing company, and you’ve got 30 plants across the US, none of which are probably close to direct flight kind of airports. You were staffing up 2 or 3 additional folks or dealing with a vendor who could get out there and do hardware refresh stuff.
Dave: 32:42 By developing a solution that, number 1, let them keep that workload close to the location, close to the gear, operationaling it, operationalizing it, making it as a service, all of a sudden, we took a lot of the worry and the labor and the care and the feeding out of that equation, and they can consume it the way that they want to, which is a cloud model. And I think that’s a good segue for John to
John: 33:03 Yeah. And and by the way, it’s it’s it is unusual and that we have an actual piece of and and this is just the actual size of this thing. I I say, you know, what ends up sitting at the edge of our customers’ locations is a box about the size that you could put 2 soccer balls in, you know, but, and we can talk about, you know, what the actual specs are here. But, yeah, these are for our customers, like like Dave was saying, where, you know, they’ve got applications that don’t cloud well, that need to be close to what, you know, whatever it is they’re running. And it’s interesting, manufacturing is probably our our top target market here.
John: 33:43 And that’s where we’ve had some really good success right off the bat running applications like Wonderware that, we’ve had a lot of success there. But again, it’s companies that, for example, like retail, you know, that’s that’s a good one where they’ve got digital signage and inventory, etcetera. They’re running their business if and POS systems. If if the Internet goes down, you know, they they can’t be out of business for so it’s actually, you know, working while they may be in an area with sketchy Internet connectivity. That also goes, for medical clinics and something here in in Texas where we’re finding that there’s a lot of interest in the energy markets with people that have, monitoring stations for oil and gas drilling and pipeline operations.
John: 34:38 So, you know, the cool thing is and how it fits in there and it works well for manufacturing is it’s a ruggedized piece of equipment. You know, it’s vibration resistant. It’s able to work in temperatures up to 130 degrees. So, you know, a very cool piece of equipment here. But, again, it’s it’s a piece of hardware that we are, you know, turning into an OPEX advantage for our customers.
John: 35:03 They’re they’re paying for this on a monthly basis, and we’re managing it for them.
Max: 35:07 I mean, if you’re if part of the presence here is remote locations, I mean, what kind of connectivity do you need back from this thing in order to manage it and orchestrate it and configure it? I mean, is this are these, like, micro v cloud deployments that you’re doing with custom hardware? I mean, what what’s, you know, give me a give me kind of like a comparison high level just just for
Dave: 35:27 Yeah. So we integrate right into the customer’s domain, right to their LAN, right to their network. So the they access their workload the same way they do today as they did yesterday. One of these pieces of secret sauce here is the management aspect. So, you know, we’re we’re VMware through and through.
Dave: 35:43 So when we deploy that box, whether it’s with a partner or the customer self deploys if they have an IT shop or something along those lines, it phones home to us. So we need a drop from them that allows HTTPS secure protocols out to the Internet, shows up in our dashboard, and then we run through the provisioning, which, again, in our world, we have a a vested interest in automating that. So if they have 5 or 6 of these and we’re deploying them all in the same week, it’s relatively low lift because the framework is there, and these things just check-in once they have, you know, minimal Internet access. If the Internet goes out, it just interrupts our management aspect. The workload continues to run because it’s on the LAN.
Dave: 36:21 So that prompts, you know, tickets, service, items from our service desk to the customer location to figure out what’s happening there.
Max: 36:29 So AWS started this with with, you know, started their Outposts. And and it’s I mean, I was supposed to I mean, it’s it’s in it’s an interesting application. Right? It’s, like, basically, you run EC 2 on or you run AWS on, like, your own, you know, infrastructure inside of your facility. But if you haven’t seen one of these things, they’re massive.
Max: 36:45 Right? It’s a full size rack. It’s a it’s a, you know, it’s an 80 inch tall. It’s a 42 RU cabinet. It’s a it’s a, you know, 80 inch tall, 24 inch wide, 40 inch deep, you know, cabinet that’s preloaded with a lot of equipment in it, and they’re not cheap.
Max: 36:57 I mean, these things are massively expensive to actually deploy. And then Google came out without with, with Anthos, which is basically like their GKE engine, like run Kubernetes on top of their own, like, software stack on your own hardware. And Anthos came into market and was really aggressively priced and then became really aggressively expensive, you know, with a little flip of a switch. And and and I’ve seen a lot of clients that went down these roads to start with these these kind of things, these ideas really quickly kind of realized, like, this was not gonna be advantageous to them in terms of the actual cost, you know, models that were being pitched. You know, it was it was interesting, but it wasn’t you know, at the end of the day, it didn’t actually help them.
Max: 37:33 So, you know, how how much of this is your edge is, you know, complimentary to other services that the customer is consuming from you and how much of this is just like this is a greenfield need that they actually have. And because they’ve, you know, like, I think one example was like me packing or an energy. Right? You’ve got stuff that’s just in the middle of nowhere, for lack of better word, that needs an appropriate footprint with with things still running it.
Dave: 37:56 Yeah. I mean, so we’ve always done and I’ll just give an example because that’s how my my mind works. Right? We’ve always done those cloud pods that we build in our data centers. We deploy those to customer sites when it’s a large workload.
Dave: 38:07 So outpost analog, if you will. Frankly, the problem we had with applying that to these small workloads that support OT was that they were just too big. The spend was too high. There was too much effort and money, so we needed to design for small workloads, which is where we saw the majority of the need in in these engagements that we were in. So a lot of r and d time.
Dave: 38:29 We’ve got a dedicated r and d team internally. That’s that’s all they do is think tank stuff. And they reached out to vendors and OEMs, and we tried to figure out, number 1, in terms of support, what are we picking? How are we building to give these clients the best support and uptime we can? But it’s gotta be small.
Dave: 38:47 It’s gotta fit small workloads. Right? So that was a big part of the decision tree is this platform, that box that John showed. Right? Normally, that thing is hosting, say, 50 or 60 gigs of running RAM for a small workload.
Dave: 39:01 And then on the converse side, right, customers love the support and the experience that we can build those to be very large. Right? Think 2, 3, 400 gigs of running memory, and we can parallel them. Right? So, you know, we’ve got a prospect that we’re working with now that is trying to build in Alaska.
Dave: 39:18 Data centers in Alaska are a little strange. There’s not many, but it’s a small, customer base for them. So they’re looking at dev on one of these smaller edge platforms and production on one of the larger built. And then that phoning home, disaster recovery perhaps, or even managed remote backup for some of those low key applications that, right, they could run elsewhere, just painfully so, if they weren’t in Alaska. So, yeah, size to be to be relatively small, and it bolts right on to both our existing support systems and other solutions that we’ve got.
Max: 39:52 So, so is is that like I mean, you gave an example of, like, 50 gig of RAM. Is that, like, the entry point to this? Like, how small is small, and, like, how big becomes large? Like, where’s you know, what’s what’s that what’s that small?
Dave: 40:03 So there’s a couple levers there, but memory is usually the rate limiter. Right, we can size these to, you know, 20 ish gigs of memory. Below that, we start asking on, you know, maybe that application really could live somewhere else because we still have that that barrier to entry on price. And then, yeah, like I said, it’s, you know, we’re we’re building with pretty high end flash storage systems in there. And we have some some latitude with what we’re picking for processors and density.
Dave: 40:28 So we can run some pretty interesting, high IO kind of workloads on these things. You know, most certainly as we approach a terabyte of memory, then we’re looking at a dedicated pod outpost type solution. But, yeah, like I said, 2, 3, 400 gigs of of memory, not not too difficult for us.
Max: 40:44 That’s a lot of application. So if you haven’t so people if you haven’t been through, like, a VM if you haven’t been through any sort of, like, virtualization sizing. Right? So one of the tricks that you try to figure out is what what the actual, configuration flags are using on the platform. So Amazon does does, you know, we’ll just use Amazon.
Max: 41:01 Amazon defines, you know, instant types based on their definition of vCPU. But there’s no actual clarity on what a vCPU really means. And when you if you’ve ever operated, like, a we’ll just use VMware. If you’ve ever operated VMware v or vCloud, vSphere or vCloud, when you install the sucker, one of the things it asks you is what you want your oversubscribe rate is on the CPU that you have. And it’ll tell you, give you guidance that says, you know, like a 4 to 6 oversubscribe for your vCPU is probably normal.
Max: 41:26 Right? And then you and then you can kinda tweak that up and tweak that down based on your actual real workloads. Just adjusting that number in an environment, is pretty incredible of the difference of the density and the cost to operate that environment. If you think about let me just run them from a basic example. If you are in a platform with a 4 x oversubscribe and CPU and make that an 8 x oversubscribe and VPU, I mean, just, you know, you’re you’ve doubled your technical, like, VM density, you know, provided that the workloads actually support it.
Max: 41:55 So it’s you know, I I think a lot of this for me just comes back to what does the application look like and what are they still running and and what state does it exist in. And I’m a huge fan of containerization and what containers have done because they’re still portable. You can run a container anywhere. You can run on your desktop. You can run it in a private cloud.
Max: 42:13 You can run-in a public cloud. You can run it in any public cloud. Right? It’s a container. It’s a container.
Max: 42:17 It’s a container. But not everything containerizes. How much of this with the edge for you is lift and shift of some like PLC control application or, you know, custom development where they could go into something that’s more containerized and and have that kind of pipeline with a container into this edge appliance. And it’s just about density, proximity, ruggedization, etcetera. Like, I mean, you know, oil and gas, have they moved into containers for their applications?
Max: 42:44 Are they still running, you know, Windows 98 VMs, you know, that need to be ruggedized?
Dave: 42:49 The majority of what we see is that custom built application, where the the person is not around anymore. They’ve left the company, and they’re just dealing with that technical debt. I won’t pretend to know the the intricacies of it, but there’s a really interesting story, interesting story, case study, that Chick Fil A underwent their edge.
Max: 43:06 Oh, I know this one very well. It’s my I love this story. This is one of my favorite examples of all time.
Dave: 43:10 And I won’t do it justice, but they tackled an edge thing with a replatforming effort, and it took a lot of time. You know, we see in our data centers where we’re talking about, you know, cloud, a lot of anchor customers, our enterprise folks, they approached us and said, hey. You know, we love this service. We love the the traditional hosting. We’re looking at replatforming, and we would love to have some kind of Kubernetes Docker Rancher environment that sits next to these things so we can dabble and decide.
Dave: 43:40 So we actually built that in most of our key markets for customers to to have both, and, you know, it’s very easy for us in those platforms to extend it to to the edge. But your first inclination is correct. The majority of what we’re seeing is is legacy apps. I mean Or there’s legacy app and an Oracle blade that we have to contend with because don’t touch that licensing, so we can fit that, right? We’re not beholden to templates or hardware sizes with this.
Dave: 44:07 We look a lot more traditionally in that, you know, if you need a blade in there that’s just for, you know, for CPUs for licensing, we can accommodate that for you. So a lot of that.
Max: 44:17 I I kind of have this thing where it feels like that any company that’s founded, post 2010 is cloud native, and any any company pre 2010 of any size is got all these landmines laying around on them and, you know, I I know plenty of examples of this where it’s you find healthy 9 figure a year enterprise and there’s some ETL process running on the back of FTP that takes and does something that ends up becoming the linchpin of, like, a third of the revenue, you know, of the business. Right? And but this and then everybody’s like, why don’t you change this application? Right? You’re like, well, because there’s like 1500 trading partners that have all integrated their platforms in with this thing for this process.
Max: 45:00 And so it’s not just us changing it. It’s then convincing 1500 people that are paying us money that they have to spend money in order to keep doing business with us and these things just never die. The Chick Fil A one is and this is by the way, this is posted. They did a series on this on their engineering blog and I don’t remember if it was on medium or or what’s what’s the actual host for it. But they were trying to tackle 2 things.
Max: 45:21 So modernization but the but the big edge play and the and the actual business case is fascinating here was they did a combination machine learning, machine vision to take and do predictive, to do a predictive modeling of how much chicken they needed to have ready at a store at any given time. So it’s hyper localized to the stores based on the pattern of the actual store. And, you know, my my kids like, one wants grilled nuggets, the other wants, you know, they just call it Chick Fil A. Doesn’t understand that Chick Fil A has, like, a bigger menu, but it’s the actual, like, fried nugget. And that and what this application does is it is it takes and has a camera that’s focused on the basket that knows what the time of day is and what this what the traditional demand is and how much is there and can prompt, you know, the operators of the store.
Max: 46:08 You need to make chicken or don’t make chicken. It’s, like, basically at the at the core level of what this thing is doing. But the output of it is all about, you know, the efficiency of how many cars they can put through a drive through per hour and how efficient they are with their with their supplies and what their overhead is as a supplies. And and the the initial deployment of this thing was they were using, Intel NUCs with, Kubernetes on top of it. So very, very simplistic deployment in terms of hardware.
Max: 46:35 But, I mean, it’s fascinating. I’m sure in this this this engineering case I mean, this is old. This is not new. This is, like, 4, 5, 6 years ago. We’re talking, like, 2018 vintage, you know, 2019 kind of stuff.
Max: 46:48 So who knows what I don’t think they’ve posted any updates to it, but it was you know, that that kind of stuff is is fascinating to me because, like, you know, you don’t think of Chick Fil A as like a high-tech company but it just shows, like, how much of difference applied technology really makes to every kind of company, you know. And it doesn’t matter if you’re doing oil exploration or manufacturing or c and c or printing or running a drive through, you know, you can you can get a lot of, like, extra oomph out of your business if you do stuff like that.
John: 47:18 You know, to to your point, there is no better restaurant chain that does a better job of managing drive through traffic and inventory control and all of the great things that ends up as a customer service item that’s coming from their technology. A, that’s great. At first, whenever we would talk about the Chick Fil A, you know, situation, I would originally just say there was a famous chicken restaurant that you know? And then, you know, I came to the realization that this was self published. You know, they they actually they they put it out there.
John: 47:52 I’ve started reading through this and, yeah, this is not, NDA type information. And and and while they have had some great outcome, outcome, it does look like there was a lot of pain in getting there that, you know, had we had the Edge product at that time, you know, we could have really helped out that, that pain and rolling it out. 5 years, I think it was, and 4 different teams that had to be created for this. You know, it was a a gigantic undertaking that, you know, now it’s a a managed service. So
Max: 48:23 How, you know, along the way and actually deploying these things and getting this into the field, like, what’s I mean, there’s a set of assumptions of, like, what you would expect. And then there’s then there’s, like, actual use cases and then there’s surprises. So, you know, looking at and thinking about, like, use cases and surprises, like, what has surprised you about rolling this platform out to to customers and people actually using it? Like, what’s you know, because there’s, like, an idea of, like, okay. We’re gonna give you this ruggedized platform that you can put on to fill in the blank, you know, place so that can run your, you know, y application.
Max: 48:52 Right? But I would imagine there’s been some stuff that’s kind of was like, oh, that’s that’s interesting. You know, you’re doing what?
Dave: 48:58 Yeah. I mean, those came out, as our first rollout, which, right, we were clear with the the client. Right? This is we’re cutting our teeth here. We’re gonna we’re gonna learn together.
Dave: 49:07 We think we have it, but 15% of this is unknown. It was the deployment. Some of they were manufacturing. Some of these sites were rough. I mean, they were a broom closet, or it was under Sally, the front desk’s, for the assistance under her desk.
Dave: 49:24 And then there was a couple of times where we needed to get, you know, some additional capacity or, you know, RMA, for instance. And it was our existing support organization that handles, you know, day to day of all of our cloud and colo, solutions. Now they’re working with the edge solution and, you know, talking to Lyman number 1 on what do the blinky lights look like. Now granted, we had back end dashboards. Right?
Dave: 49:48 So we’re it’s more of a validation thing, but I I think that exposure to a different type of support was definitely something interesting. But you have the locations. I mean, I mentioned a no direct flight kind of airport. There’s a lot of that with this ecosystem, for sure.
Max: 50:03 I mean, obviously, Edge, we really talk about, like, computing resources closer to the app to the to the you know, putting an application, putting a workload, putting a computing resource close to the consumer of that. Right? Part of what I’m constantly looking for is what is the latency threshold? So what is the distance threshold? And, like, what is the application that drives then that latency threshold?
Max: 50:25 Now I’m I’m not gonna give you the easy example of you’ve got an oil derrick in the gulf that doesn’t have fiber optic connectivity back, by the way. There’s actually fiber optic networks connecting oil derricks in the gulf. So it’s a bad example. So let me know. So you’re in the perimeter base and there’s nothing out there.
Max: 50:37 Right? You know, whatever it is for oil and gas. Like, there that’s an example. There’s just no connectivity. Okay.
Max: 50:41 That’s an easy example for edge. But in these other markets where you’re talking about no direct flight, you know, secondary, tertiary markets, you know, there’s there’s still connectivity out there. There’s Internet out there. Starlink is available for most people at this point. You know, you can get these sorts of things.
Max: 50:55 So what application are you seeing drive this? What kind of latency, you know, is driving this? You know, what what is the what is the primary thing that is is actually pushing, like, the real edge now for you, you know, with this product?
Dave: 51:10 So lots of examples. We have a customer, who’s with us, production style in one of our clouds in Pittsburgh. Again, example. But they have some tape media that they need to touch frequently, so we’ve loaded that into their colo rack in in Baltimore. That Baltimore environment also caters to some of their app dev work.
Dave: 51:29 And the question was put out there. Right? We have these production assets in Pittsburgh. We’d love to use production firewalls and production gateways in Pittsburgh for our dev stuff in Baltimore. A 100 gig or better protected backbone between those markets.
Dave: 51:42 Latency is about 10 milliseconds round trip, and customer experienced issues with that in their app dev environment, and we had to pivot a little bit. So that’s a good example in my mind where 10 milliseconds on a backbone, should have been fine and it wasn’t. But I think what drives more of the edge is in that lost revenue if the workload gets isolated. Right? So, you know, just in time shipping, where if they miss a shipping window, that product doesn’t leave, they can’t bill for it, they may have lost $500,000 and shift workers that went home and all the HR associated with that because they lost Internet connectivity to the the workload that was in the cloud.
Dave: 52:21 So that’s generally what what drives these things is the the risk and the loss associated with, isolation. At least as it stands now, it’s a little bit less around the latency.
Max: 52:30 Are you finding this more like a response to pain? I mean, so 10 milliseconds of latency is not a lot of latency. I mean, that’s especially if you’re talking about round trip, you’re talking about 5 millisecond one way. I mean, that’s not but that’s not a lot of latency. That’s metro.
Max: 52:40 I mean, that’s that’s sub metro, you know, speed in most cases. It feels like there was a lot of pain at 10 milliseconds of latency for them to go out and spend money to solve that pain or that, you know, the example of, you know, shipping and and inventory management or where, you know, like these different things. And inevitably, for me, it feels like that’s pain response. Like, something happened that generated pain for us that we went out and had to, like, you know or or am I completely wrong? Are people being proactive with this and just looking at it and saying risk modeling?
Max: 53:06 Like, this is just not worth the risk for us. Because if this event does happen, it’s not I mean, so are are companies getting proactive with this? Or are they just responding to, you know, like, we had this thing, and now we gotta go solve for this thing and never have this thing again?
Dave: 53:18 I mean, the the the door opener, the conversation that we have when we’re engaged with a company we’ve not met before is just that. What would happen if, or what does it look like when the last time this thing happened? And to my, you know, earlier point where that’s the crown jewel and the revenue generator and don’t look at it the wrong way because it can’t go down, they’ve kind of abandoned the thought or thinking or their research that maybe this could be done better. So it’s always been just leave it alone, be as safe around it as we can. With the edge and how we integrate with the client networks, it’s a lot easier and and and less risky.
Dave: 53:54 And then, by the way, we say, hey. This can be operationalized, and you don’t have to do your tour of duty to replace things or to break fix things. And not only have we derisked that, but we’ve given it a better support path.
Max: 54:06 Baltimore and Pittsburgh are not small markets. I mean, these these are not, like, strange rural cities, right, we’re talking about. I mean, it’s it’s it’s I’m surprised to hear that as an example, honestly. Like, I’m still kinda digesting that a little bit in terms of it’s they’re not far from each other. There’s plenty of network assets between them.
Max: 54:21 They’re established metros. There’s plenty network asset within the cities. You know, these are these are very I mean, these are sophisticated markets for network and Internet. And they’re not far from, you know, your traditional, like, hyperscale sized cloud hubs. You know, the latency and the distance between Baltimore and Pittsburgh.
Max: 54:36 Yeah. For I mean, it’s not far. You’re you’re on top of the Internet as far as the cloud is concerned. So it’s it’s, I’m I’m really actually impressed that that was you know, that that that actually drove that because I I wouldn’t have I wouldn’t have come into this conversation
Dave: 54:50 and put that Well, and the converse is also true. I mean, you know, Memphis is a market that I cover where John’s at. And, you know, there are pockets of Memphis that the connectivity is rough, and you wouldn’t expect that. So even getting, you know, a point to point back to your colo rack in the data center, depending on where you are in the town, you can’t get redundant paths. Right?
Dave: 55:09 That kind of thing too.
Max: 55:10 You know, I’m I’m I’m going through this and just kind of running through horror stories and experiences in the past that I blocked out from my memory from PTSD. I think the commonality there becomes high density industrial warehousing locations where there’s massive square footage of facilities, but not very high square density of of users. And so convincing, a cable company, for instance, to do any sort of real build in a market where they’re not gonna have a lot of doors they pass along the way doesn’t hit their economic model, and those things just don’t work out. And then you end up, you know, in places that just, you know, it’s it’s weird. Right?
Max: 55:42 Like and I’m thinking of several examples they blocked out in my mind because they were just so horrible to deal with where, you know, it took us 250 days to get something built into a into a building. And what do you do in the meantime?
John: 55:55 That’s interesting. That probably speaks to why manufacturing is such a fantastic opportunity for us in this you know, the the edge is exactly. It’s it’s a a large facility potentially out in the middle of nowhere that, you know, again, the cable companies aren’t burning a path to to their door because they got a single user out there. You know? So interesting.
John: 56:15 I I didn’t think about that’s that’s probably what’s driving a lot of this demand.
Max: 56:19 Let’s let’s we actually didn’t get into this too deep, but let’s talk about this a little bit. Like, who is your ICP? You know, like, when, you know and I mean, give this to me, like, region or geography or, I mean, you gave the example of, like, small edge deployment in Alaska. Right? Like, that’s probably not, like, the top tier of, like, anybody’s ICP in terms of, like, finding some kind of rural Alaskan deployment to fill in the blank.
Max: 56:37 Right? But, now that I’ve I’m thinking about, like, ice road truckers and, like, diamond mining in Northern Cal you know, Canada, I’m sure that’s probably a phenomenal edge deployment scenario. Right? You know? Sub subfreezing temperatures.
Max: 56:48 I mean, like, what is what is Expedience ICP? And I guess that’s that’s, like, generic in terms of, like, overall, but then also as it relates to, you know, acquiring data centers or, edge deployment. Like, what’s you know, where where industries, what markets, what what sizes, like, who’s who is your customer?
John: 57:05 It’s typically customers that have incurred a lot of of technological debt, and that is, generally speaking, it’s older companies. It’s like you said, anything built past the year 2010 is probably built for the cloud, and we’re we do great with those industrial companies that are, you know, again, they’re into 9 figures, but they’ve got substantial base of operations from an IT standpoint that are on homegrown applications that they’ve built. Those are the perfect customers for us. And but, you know, with regards to the edge, it’s it is a little bit different in the fact that, again, that location thing with health care clinics and retail come into play. I don’t know, Dave, would you agree that’s kind of our, you know, the 30 plus year old company that’s established, but, you know, they’re certainly not built for the cloud, but they need to get to that, you know, that OPEX
Dave: 58:02 model. The vast majority fits that. The other thing I’d say is, you know, service providers themselves or, you know, companies that have built, their own intellectual property, and they’ve evaluated, right, Kubernetes on AWS versus something else. We get along. We understand them as as providers ourselves.
Dave: 58:19 We do really good, work in that in that vertical, at the same time. So not as common. Just there’s not as many of those companies out there, but another good example.
Max: 58:27 Besides spending another 2 hours just talking about, like, crazy examples of deployments and strange locations and people doing weird things, anything that we haven’t touched on that you think, you know, we we should talk about for a bit. I mean, have I missed anything within this Do you think it’s important to I
John: 58:44 was gonna say, but those were the fun things to talk about is the all the
Max: 58:48 What what’s my favorite example? My favorite example is, like, the mainframe that was drywalled in underneath the staircase somewhere. And they had no idea where it was running, and then they were trying to find it. And all they had cables hear it.
Dave: 58:58 Where where’s
John: 58:58 the noise coming from?
Max: 58:59 No. I mean, they had no idea. This thing was, you know I mean, this is one of the classic quintessential stories of, like, you know, early early tech, you know, or I mean, it’s phenomenal. Like, these things I mean, they do exist. Right?
John: 59:10 Yeah. I you know, and and Dave, I mean, one thing that we haven’t talked about, one thing that probably drives the initial conversations that we have so much these days is a, I guess, the fear of of ransomware attacks, you know, has a great you know, we have a great story to tell to help mitigate that threat, through our disaster recovery, products. So, our push button DRAS, especially that, you know, powered by Zerto, we have 3 different sizes, but with almost instantaneous our p o r t o, you know, that’s a great story that that that we can tell and drives a lot of our conversations and usually is our toe in the door to and and, you know, then we expand our our talk, more generically.
Max: 59:54 We kept this pretty narrow. Right? So let’s let’s actually back up. And, John, this is a good good point. Like, you offer a lot more stuff besides I’m gonna buy your data center.
Max: 01:00:02 I’m gonna give you an edge appliance. Right? And so so backups and Doctor, we know we touched on immutable backups briefly, and you just mentioned Zerto. But, like, what’s what’s the rest of the service, portfolio? Let’s
John: 01:00:11 Taking a look at my background there, it really does. You know, we say we’re a full stack cloud service, provider, and I love showing that whenever you think of Expediant, you either think of, you know, just an IaaS provider, we’re a VMware based cloud provider, and it sounds pretty generic, but once we you take a look at our, you know, NASCAR logo slide of all of the services that we offer, you know, from, again, you know, Colo is, you know, where we cut our teeth to disaster recovery. Enterprise cloud, and we have a couple of different flavors of that. You know, we do when people say cloud and they actually say the word public cloud, they usually are talking a hyperscale cloud, but we do offer, you know, really, that’s kind of what we offer is a public cloud version of Expediant Enterprise Cloud, a VMware based cloud that, you know, can be consumed in that, you know, utility model like you are supposed to consume, AWS, but we don’t have the the shocking fees or, you know, the the outages. We we also have obviously a private, you know, purpose built single tenant cloud that we can create as well.
John: 01:01:17 But, we also have offerings around Kubernetes, disaster recovery, backup, and also VDI. So when we’re talking about Edge and purchasing data centers, those are 2 right now, Edge is a relatively new field that we’re in, so that’s a pretty small part of our overall revenue. You know, purchasing a new data center is a relatively rare thing. Those are the 2 exciting things that we’re doing, so it’s fun to talk about.
Max: 01:01:42 Of course. Those are the exciting things. That’s why I wanted to talk about them. Right? You know, like, I mean, VDI, everybody gets out of bed in the morning.
Max: 01:01:47 He goes, like, oh, man. I wanna talk about VDI today.
Dave: 01:01:51 Yeah. I
Max: 01:01:52 I mean and not to I don’t wanna I don’t wanna throw a shade on VDI or or, you know, the different, you know, desktop as a service or whatever we’re calling these things nowadays. But I’m a Microsoft Windows Terminal Server 1 point o. You know, I’m a big fan. I go back years. You know, this is a multi decade.
Max: 01:02:07 Like, if you’re not running it, you should run it. That’s the plug. Okay? You know, these things are really awesome and good for you. And actually, so the ransomware is interesting.
Max: 01:02:14 Let’s just talk about this. We’ll end here. The the trick with ransomware becomes, I mean, there there’s a whole bunch of tricks with ransomware. But part of the issues that companies deal with ransomware is, like, when did it come in? And then how far back do we have to go in order to get back to clean, you know, cleanliness in order to recover data?
Max: 01:02:32 And, like, what are we actually losing? I think this becomes a big question for a lot of people. And there’s a lot of stats related to dwell dwell time and average dwell times. And, you know, whether or not, you know, you have, you know, any sort of the cyber apparatus around EDR, SIM, XDR. Let’s use lots of different acronyms to try to figure out, like, what the heck happened to you.
Max: 01:02:50 But, you know, when you we talk about, like, immutable backups and going back and recovering from ransomware. You know, what’s what’s, like, the standard now? Like, how like, what’s what’s the expectation in terms of, you know, for you with with your customer base, you know, running Azure platform? Like, how far back do they have to be prepared to I mean, are you you know, 31 days? Is that the magic line?
Max: 01:03:09 Is it 61 days? 91 days? Like, what what’s what kind of counseling and guidance are you giving people? You know, what’s what’s your view in the world when when this comes up?
Dave: 01:03:17 Yeah. So this is very, consultative type conversation that we engage with clients in because, like, there’s there’s all kinds of different variables and needs and things like that. It’s multifaceted. So it’s rare that we ever say, Zerto fixes this, or use VMware VCDA, and you’ll be golden. It’s it’s always a multistep approach, especially where compliance is involved.
Dave: 01:03:38 Right? Are they HIPAA? Are they PCI? Do they have some things that we need to be cognizant of? And we find it’s usually a combination of of layers, a disaster recovery, like a replication based thing with the appropriate length of of recovery point, whether that’s a journal system or a snapshot system, and then the backup system.
Dave: 01:03:55 Right? We’re doing backups anyway, so let’s make sure that the retention there fits your business requirements but also fits the risk requirements. But then there’s some neat technical widgets that we’re actively working on that, number 1, look into the backup stream and identify higher than expected change rates, so indicative of an encryption event. And how do we feed that data that we are noticing to our support organization, right, these guys that that answer your tickets in the phone out of our our data centers and get you up in the middle of the night because we think something funny is happening? So there’s some technical stuff that we’re also hard at work at to bring to the to the table.
Dave: 01:04:32 Today, it’s it’s backups, it’s replication, combination of those, and integrating with, you know, the enterprise systems. So if you have an EDR or an XDR, how do we get that information, and how do we use it during a suspected breach? And then parallel, because, you know, we’re consumption based, we’ve done this for clients that have been infected, isolate this environment, sandbox it. Here’s a same day new resource pool that we’re gonna either turn your backups on or you can rebuild or whatever. And now you’ve got a forensics based environment that, you know, you’re gonna litigate and your partners are going to you, mister prospect.
Dave: 01:05:09 Your partners are gonna get in there and and do the investigation. But at the same time, you’ve got a running environment that’s healthy, and you can be more responsible with that forensics. So couple different things there.
Max: 01:05:18 This is probably one of the biggest non discussed use cases around VDI and the DAS desktop, you know, virtualization things. And there’s imagery I remember, and I think it was Boston Hospitals that had a ransomware attack. And they called up the National Guard to help them reinstall the operating Windows on all the desktops in this hospital group because they were completely down. You know, seeing uniform National Guard you know, photos of uniform National Guard in a hospital, you know, with installer discs going through and reinstalling desktops to bring this hospital back online is just it’s just not really something you wanna see. And it and it’s still to this day really kind of bothers me and probably the best stat and this is the this is I mean, it’s the the best worst stat maybe, but the best stat that I think about is average length dwell you know, dwell time length have been plummeting.
Max: 01:06:04 So, you know, what they were 2 years ago versus now. I mean, on average, ransomware attacks seem to launch much quicker against the target. And so you’re not dealing with less exposure time or less potential rollback time because I don’t know if you’ve I mean I mean, going back 25 years, I can remember a manufacturing client that had massive data loss on their ERP system, and they had to manually reimport, 6 months worth of data into their ERP from, you know, paper records. And, you know, it’s they hired an army of temps to sit in a conference room, and it was a very long arduous process in order to recreate that data. But, you know, you think about it from ransomware.
Max: 01:06:37 You know, if you have to recover back 60, 90 days, I mean, that’s there’s a lot of change that happens in 60 to 90 days for a company today. And, God willing, that’s on that’s as far back as you have to go.
John: 01:06:47 Yeah. And I I know of a a, you know, a famous case, and I’ll just say of a health care provider system that, you know, was attacked by ransomware and, immediately hopped into the world of processing things by paper and pen, and they are still months months later, still pretty much on a paper and pen system right now. And, so yeah. And the the the the cost have now soared into the 100,000,000 plus range with this particular group. So, yeah, it’s a,
Max: 01:07:17 I don’t wanna end this on a depressing note. I’m gonna do, like, a whiskey Wednesday ransomware thing at some point, and they’ll just be prepared for it. But I don’t I don’t wanna get too deep into it right now. You know, I mean, the takeaway I think I get from here is is, you know, Expedient really is seems to be tackling, like, modern enterprise needs in a way that, you know, it’s it’s a lot of people on the market offering colocation. There’s a lot of people out in the market offering VM Cloud.
Max: 01:07:41 They don’t call nobody calls it v Cloud. Right? But there’s a lot of people in the market offering v Cloud inside of their data centers wherever they are. And this is, I mean, 2 very unique offerings in market that, you know, has a very specific customer that needs this type of thing. But if if you’re that specific type of customer, like, what what a gift to be able to take this path and be able to, you know, leverage these capabilities to just take care of your business and do what your business needs.
John: 01:08:07 And I’ll I’ll say my my great takeaway from this is that there is a thing called Whiskey Wednesday. So, I’m I’m gonna I’m gonna celebrate it.
Max: 01:08:16 It was it was just invented. We just did a Taco Tuesday to Wednesday Wednesday Whiskey Wednesdays.
John: 01:08:20 Alright.
Max: 01:08:21 John, Dave, thank you so much for your time. This is fascinating. I’m sure we could spend many more hours and dig dig deeper and deeper into this, but, this is this is great. I’m very, very happy to be able to spend time with you and and learn more about, you know, data center and your edge operations.
