Greg Irwin 0:18

Good afternoon. Thank you all for taking some time to join our discussion group. We are talking bare metal as a service. We're here with my friend Jack Hogan over at Pure Storage, Rich Grandelli Also at Pure in, we're talking about use cases. And he kind of real world feedback on bare metal as a service. So I think many of you know how we run these forums. I'm pushing for interactivity. So I love cameras, because it's, you know, again, drives that engagement. So if you're in a spot where you can turn it on, great, if not, that's fair, as well. As we go, please use the chat window. It's very helpful when one person's talking, you can just layer it in lob it in there. And it'll definitely provide fuel to the fire in terms of the conversation. So as we go, let's use that chat window throughout add questions, add your own stories, if you have a different viewpoint, those are even better. So let's, let's use the chat. And then as always on these you know, I'm really trying to drive on the relationships. So whether irrespective of whether it's somebody at Pure or BWG, I want everybody to kind of aspire to connect with one person across the Hollywood Squares. So someone says something smart, someone you think might be, you know, a good addition to your personal network. Take advantage of that. Alright, let's get right into it. I'm going to introduce Jack Hogan, over at Pure Storage. Jack, do us a favor, please give a little intro to yourself. And I'm gonna ask you to kick us off here with that intro on bare metal service.

Jack Hogan 2:14

Yeah, sure, Greg, appreciate it. Thanks for having me. Again, it's always great to see you. I really do appreciate this forum, too. I think we've had some really great conversations over the last couple of years that we've been doing that so I'm really looking forward to getting everyone engaged. You know, as Greg mentioned, you know, I'm Jack Hogan, I am the Vice President of Technology Strategy for Pure Storage, my primary role is to work with our cloud partners. So those are going to be everyone from, you know, those that are public, hyper scalars, through software, service providers, managed service providers, systems integrators, as well as data center providers to drive where pure is taking our technology and solving problems for the world. Specifically, I kind of come from a background of being a CTO on the customer side for 20 years in the mid market. You know, work in a consumer cloud and SaaS company, have a very good understanding of what it's like to run systems, you know, that need to be fully resilient and up all the time. And so we have been working very closely with my colleague, Rich Grandelli who's on the call today, who runs our global managed service provider business, in developing new partnerships. And, you know, specifically the partnership we'll talk about today is going to be around what we're doing in the bare metal as a service space, with our first partnership in this area of being with Equinix. You know, what, you know, many people in this call probably have an idea of what bare metal is, but let me just make sure I set the stage, you know, in the broader kind of call it x as a service framework, or you know, everything as a service, you know, the traditional kind of standard things, which is it's going to be your infrastructure, your platform, your software as a service, you know, those being kind of the four, the three main pillars, what we're seeing emerge between enterprise IT, and the infrastructure as a service providers is this space of bare metal, and bare metal being the ability to have full root-level administrative control over the underlying hardware that's dedicated in a single-tenant format. So this varies dramatically from what you get out of the public cloud where you're, you know, being put on to very stringent multi-tenanted platforms that ultimately you know, have their own operating models and own operating platforms that they you know, force customers to comply with. In the bare metal as a service sense. You're able to bring your own software, bring your own licensing, bring your own automation layers, and ultimately deliver the what what I like to refer to as kind of the past, present and future challenges that anyone in IT is facing. How do you handle your legacy, technical technical debt, those applications that don't forget that don't refactor well into the Cloud, as well as then those that are in the cloud that are being developed in containers or cloud-native services, how do you leverage that and bring that into this kind of well called Goldilocks zone of being able to have the full control the full power, but also have the benefits of as a service and being consumption-oriented. So what Pure is doing very specifically, with our lead partnership in the space, which is with Equinix, Equinix acquired a service provider by the name of Packet about a year and a half ago. And in that they provide what if you go out in the market, you'll find bare metal servers are starting to become, you know, something available from a number of different primarily Colocation Data Center providers. But nobody has ultimately cracked the nut that the public clouds have, which is creating a disaggregated storage layer. And that's what Pure is doing now with our partnership with Equinix. And what they're now calling Equinix. Metal that's come out of their acquisition of packet where they're bringing the compute the network and the orchestration layer. And Pure is providing a full data platform. That data platform is really in the form of block file and object storage services, all consumable just like you would see in any sort of public hyper scalar in a pay as you go model, all fully capable of full administration, administrative control. So we're introducing something to this bare metal space. That's, that's unique and novel, but we're also seeing that it's fitting the 95% of the world's enterprise workloads that currently don't sit in public, hyper scalars. And it's for the reasons of regulatory compliance, security, control, all of and performance, all of those are reasons why this is solving that I call Goldilocks solution of the 95% of where all the operators that are on the call here, and I'm really interested to hear, you know, different points of view on this. But you know, where everybody is trying to have that control, but be able to consume with the same flexibility you get out of, you know, what the promise of the cloud is. And so we're solving a lot of really novel and unique but very, very important problems, we believe in today's technology, and technology landscape. And, you know, so again, really looking forward to conversation. Hopefully, that was a good enough setup, Greg, kind of for what the topic is and where we're going and, you know, happy to happy to answer questions, or just kind of get into the conversation.

Greg Irwin 7:31

Good setup. I'm gonna let's let's get Rich with an intro. And then I'm going to ask you about a real world deployment of Rich, let's get into it.

Rich Grandelli 7:42

Thanks, Greg. So, everyone, thanks for joining us. So Rich Grandelli I'm responsible for the managed service provider business and strategy globally for Pure so as, as Pure looked at the marketplace, and as and how it's evolving and how folks like yourself and how folks are looking to consume technology, clearly has been moving over the last few years down that subscription model to buy as a service. And beyond just the hyper scalars. Yes, there in the world, as Jack just talked about, but alternative ways of how people were going to buy and the service provider space. And that's a very big word. You know, it goes across from what you think of traditional telcos, that as they move into the edge, and they provide their solutions based on IT and 5g, oh, that's evolving. When you look at traditional managed service providers, something like the Rackspace is of the world, the systems integrators. And now you'll see in the evolving, what were data center operators are colo providers evolving, what they are going to be offering as a service. And then you're seeing the SaaS providers, and so on. So it really became an area that the company saw as critical to our business to become a part of the platform that these folks are building. So we really put a focus globally on what we call those managed service providers. And then over the past, like I said, several, several months working with Jack really, over the past year, really starting to look at what are the evolving spaces, and one of those being this bare metal space as its evolving, as Jack had described. So putting a focus on that as an area, but my team sits globally. And we work on both the strategy across all routes to market within managed services, but then specifically covering a set of managed services on a global basis.

Greg Irwin 9:24

Got it? All right. So guys, I want to go to a one story of an organization that's deployed. And what the benefit is, in terms of just running a managed hosting. That's probably the direct comparison is go to an MSP and run a managed hosted by a subject tell us yeah, thanks for that.

Jack Hogan 9:50

So So I'll start by categorizing those type of customers that were saying and you know, and I'll give you a specific example. Um, you know, what, what we're seeing in the uptake of this in the early stages As a note that we launched this, this partnership back in March of this year. So we're still in the very early stages of this. But what we found in the early stages is that those that are in the services business, whether it's they are enterprise services for their own corporate IT, whether they're in a managed service provider space delivery, delivering kind of higher-order services, or they're delivering a full Software as a Service, those are kind of the primary that we had as our primary hypothesis. And ultimately, it's proving out to be true. The story I'll tell you is a little is about a software service provider. I can't mention the name just yet. Look, person things coming from us soon on this. But, um, they they ran their own data centers, they had a data center in Seattle, they had a data center, and in in Denver, they had staff dedicated to running their physical environment. And they were coming up for the renewal of their data center footprint. And so they started to look at different options. And they looked at different options from public cloud where they had a presence, but realize that a lot of their licensing and software that they had advantageous, advantageous pricing on would not port well into the licensing models of, of the public clouds, they also looked at other providers of kind of full stack, delivery of a service. And, you know, ultimately, you know, kind of the, they recognize what we were doing with our Pure as a service model. So I'll pause here real quick and go through what Pure is done by bringing the whole ability to bring a consumption model to what has traditionally been a capex asset based purchasing model, Pure introduced this model to the market, you know, little over three and a half years ago, and now we're seeing a lot of our competition, you know, in the form of, you know, certain colored lakes and different peaks of mountains and different names of our competitive products. You know, we feel like that, you know, that we kicked off this revolution of consuming physical hardware by, by a consumption model. And so what, when we started to talk with this particular SaaS provider, they were very interested in a consumption based purchasing model, but they were more interested in focusing their internal resources, their human resources, on revenue driving projects, rather than just keep the lights on. And they recognize that by by having a data center by nature, they had to have resources dedicated to this. And so by now, providing them the ability to have a fully virtual data center, where they can define the specs on the hardware, they can define how much memory and how much compute sits in each node, they can define exactly how much block storage are different levels of service that Pure can provide, that they ultimately found the massive benefit to be able to exit their data centers. So they've now fully exited, they're their primary data center, they're now working on their secondary data center, and will be by the end of this year, basically, fully virtualized within, you know, what was once a year prior a physical data center, they can operate at the same efficiencies by having this virtual data center model. And so you know, the benefits they've been able to achieve is now faster, time to market more flexibility. And to get back to the core element that their VP of infrastructure pointed to us in the very early stages is they want their resources, their human capital, driving revenue, and now they're able to do that, because everything is automated and orchestrated, you know, in the platform, they don't need to go physically turn a wrench or stand up servers, it's been phenomenal, you know, great customer for us to watch. Now, I'd say we've got a couple of others in the managed service provider and, and, and other type of service provider space. But you know, that's what we've seen. Now, we're also seeing enterprise customers start to recognize this model, as something very powerful as either an augmentation or complete replacement of their data center. But that story probably tells the best, which is, let's put the people in technology from being a cost center to being a romanism.

Greg Irwin 14:16

Jack, what's the range of apartment that somebody can deploy? Maybe I'm my apps are configured, you know, qualified to hp. Or maybe, you know, I'm comfortable with Lenovo. Does any of that matter in the wealth? What's the range of hardware because we're not just talking Pure Storage here, we're talking compute network and the whole, you know, all the whole infrastructure.

Jack Hogan 14:44

Yeah. So from a hardware point of view that they they offer compute resources from a couple of different vendors. supermind probing kind of the the primary Dell has offerings in terms of some of their poweredge servers. But at the end of the day, customers are going in and sizing more tos. The requirements, the compute the memory, the performance, they do have available options for GPUs, they do have options for arm chips. So they've got a fairly broad range of compute options. From a networking point of view, it's going to be your regular, you know, succeed expect, you've got, you know, your junipers, you're in your Cisco's and, you know, they're they, but they run that in such a way that they provide a management layer for the hardware that looks very much like any cloud control plan that you would see out of a hyper scalar, where the actual physical administration of the the network layer or the ability to segregate out at, say, layer two, or layer three level, all of that is is fully included in the the management plane. And so the actual hardware manufacturers are less relevant when it comes to computers, you know, relatively dumb when it comes to features and functionality. But it's then what you kind of place on top of that, and that ability to now operate at, you know, physical, I mean, you go from a pixie boot level, all the way up to fully orchestrated container model, you can go through bringing in VMware, and, and, you know, different levels of virtualization. But that's all a customer's decision, you know, again, back at the hardware level, which is what the service that Equinix is providing, you're getting the raw, the raw management of that hardware. And so, you know, there's limited, limited challenges, and you're capable of really kind of deploying above the stack. And this think about the hardware being a utility now and your ability to deliver your service or your, your outcome on top of that.

Greg Irwin 16:53

Jack, one last quick one. How broadly available, is this across the Equinix? footprint? Is it just selecting centers? activity is key, right?

Jack Hogan 17:02

Yeah, proximity is key. So we for those that know Equinix, you know, maybe for those that don't, I'll get a little bit of a an overview. They are, you know, one of the largest data center providers on a global basis. 230 or so data centers globally. Those, you know, range from I think it's 26 different countries and 50, in multiple different metros in some of those countries. And so that, that availability of full Equinix is full footprint of 220 data centers is the promise of the platform, what they've launched, and where we're following with them. 100% is in 18, core locations 18 core locations being you know, I think we're looking at four in the US. You know, we've got multiple in the APG region multiple in the EMEA region. And we've got, you know, in Canada and Brazil, and you know, growing into the rest of Equinix, his footprint, as the as the uptake of the Equinix metal product increases, the the plan is that that will eventually get into the bulk, if not all of those, those available data centers from aquidneck

Greg Irwin 18:18

are very good. Hey, let's, let's start stirring the pot here a little bit. I'm going to go around the group. I'd like to talk a little bit how, how people are considering hybrid environments, whether its own data center, or whether it's crumble, it could be, you know, Microsoft Azure, Azure Stack, it could be anything that's appropriate for you. And I'd like to hear the range. So I have a question for folks in the chat. What's your mix of hybrid? I'm going to assume everybody has Azure and or AWS? What's, what's on the local land? So do us a favor, drop, drop in the chat. What's your mix of of hybrid? And I'm gonna, if it's okay, I'm going to let's bring some people into this. So I'm going to invite Michael Mike, you've been kind enough to turn on your camera. Do us a favor, give a quick intro.

Mike 19:18

Hi, my name is Mike. I work for. I'm a technologist. We're looking at moving into the cloud. And off site. Off Prem hardware utilizations were a little unique because we're a defense contractor. So we have to deal with lots of compliance issues, both in Canada and the US. Yep. overseas. So you start dealing with stuff like you know, unclassified documents and international trade art and stuff, that kind of kinds. It's a little tricky to what we're doing. So we're kind of approaching the whole idea of AWS is or carefully at this point, to see where we can leverage it.

And on the US and Canada, they just got Control, good certification. So we're still quite a ways away from actually implementing anything to cloud, you know, we're looking at or office 365, we're looking at going into the cloud for that kind of services. But for the most part, it's still primarily investigative investigatory trying to figure out what we can do where we can do it. Rarely gate, I got all sorts of ideas popped in my head about what we do experimental for. So as we build that model, it looks like it'll still kind of kind of fit the pieces together. So I've got some ideas to go,

Greg Irwin 20:33

I'd love to go there with this. Mike, where are you using their mouth?

Mike 20:37

Right now it's all internal. We have internal clouds. And one of the challenges too, is we're we do a lot of arm based, arm based development. So having arm based hardware available will be kind of interesting and fun and exciting to us. So kind of keyed off on now and actually wrote that down. So

Greg Irwin 20:54

cool. What's the advantage of bare metal? Is it. I mean, I think I know the basics, but I'd love to hear it.

Mike 21:02

So for us, if we're dealing with, you know, we're doing cross compiling, which is a bit of a pain in the butt. So if I can get my guys to work on an arm based framework, reduces the cross compilation time, your moves, cross compilation all together. Yeah, I, like an Old Dominion, one mac book came out, I got all excited, I wanted to go buy a bazillion of them and give our developers one because it's our base rate. So and then just bare metal, is bill right to improperly gives me a much, much faster response time over virtualization, particularly for the developers of the build side of things in our testing frameworks, right? So when we started, start talking about virtualizing, environments, that kind of stuff. bare metal just gives me that much more speed. One of the challenges we have right now, just even from a testing perspective and a virtualization perspective, we want to test on real hardware as much as we can. But you can't because we have a limited number of boxes that we can play with. Right. So having the ability to virtualize but then getting the speed that we would get out of a normal box would be phenomenal. I you know, because you have those different translations layers, we're trying to try to avoid that as much as possible. So, right.

Greg Irwin 22:11

Hey, Greg, Greg . Thank thanks for jumping in. And stay involved questions, governments, however, maybe eight Hey, Guy let's bring in, let's bring into the mix. What's it? What's the, what's your deployment of bare metal across your environment?

Guy 22:32

Surprisingly, we're 95% virtualized, but just for pure disclosure purposes, I am a pure customer. And we are evergreen customer. So we are a large enterprise in the southeast. And what I will say is one of the things that bare metal gives me is, is the service that they're offering is exactly that this problem we're trying to solve right now with candidates that actually the I have four generations of technicians. And the older generations love to mess with hardware. The younger generation could care less. They just want they just want it to work. And I have younger generations that I'm having to teach about hardware because we have hardware, where are we at? So we currently run across three different types of data centers, we have a local data center on our campus, we have a Colocation Data center that is, is currently our disaster recovery, but is going to become our main data center. And then we also have a virtualization in Azure. So we kind of mix it across the board. But one of the things that I've been trying to do for the last 12 months is exactly what Pure in Equinox is saying is that we need to get away from having to deal with the patching the firmware, the switching out of Nicks the switching out of, of just pieces. I spent, just this week alone, I can tell you, I had two high end technicians in a data center, standing watch it another person that we had from a third party come and replace a motherboard replaced all this stuff. And it's like, why do we have this? It doesn't make sense anymore. It 20 years ago, it made sense. Now it doesn't. So in the midst of it, we are currently I'd say out of the 556 servers we have in our virtualization environment. We are currently managing about 500 of them. And we're trying to get as a consumption model. As we reflect lifecycle, we'll replace them with consumption. And that's it's just the way it has to be with as Jack said earlier, it's a it's about getting revenue generating and not a cost center and it has to work. The business on that? I didn't answer your question, Greg, but I just thought I know

Greg Irwin 25:09

better than then my question what so I hear the value of it. I also see your virtualized like we we spent we 20 years virtualizing our environments? You know, there's tremendous value in virtualization, of course. But there are a lot of workloads that don't necessarily aren't. Mike described, there are a lot of workloads that basically get get tied up with, with virtualization. So, you know, Guy, what percentage of your workloads do you think are suited for bare metal? Is this an odd percent case?

Guy 25:42

Oh, no, I'd say it's probably closer to 15 to 20% case. Right now, I'd say we're currently 5% virtualized, because that is, you know, I am with the enterprise. And that's what we manage. But I will tell you that just this week alone, there's people that I've quoted hardware for because they're like, we need hardware for this to run, we need hardware for this to run it. And I'm pushing virtualization because I'm enterprise managed, but it is. their mindset is, it's whether it's rent an R or Python, or whatever, you know, database are trying to do analytics with those, or they think that those should be bare metal. And if I could get them to not own the hardware, because they don't patch it, they don't firmware, they don't do anything to it. They just let it sit there for years. And getting them to a consumption model, which I'm, I'm going to say we're looking at high performance computing, as a consumption model here in the next couple months. That is, how we're going to do it is get them away from it, but give them the option. We have it for you. We just don't we're just not magnitude. So

Guy 26:54

yeah, from for exit from the enterprise. We're 95%. But I know that there's use cases outside that we don't manage because it doesn't fit into our model. Sir.

Rich Grandelli 27:05

Greg this is Rich, if I could just jump in with a comment. As I'm listening to this. Just shortly after we made the announcement, I had a conversation with a customer that I had known that was interested in, call me up and just wanted to talk about it. And his perspective, as he listened to it. This was a company that was looking at a significant portion move into Amazon and was deciding what workload what he saw was they were coming at their infrastructure and saying, Okay, how do I go, I'm looking for that alternative. I'm going to the cloud. But is that the right answer, I may need an alternative to the cloud, which he meant by that was the hyper scalars, what he listened to what was described with what Equinix was doing with us, it became he was really looking for an alternative to on prem, not that there was going to be a role for the hyper scalars, there were going to be workloads that they can characterize they could determine fit there. And it made sense. But but they also had this drive, to move from their data centers, whether it be driven by cost, but the flexibility they need, you just went through Guy all that it goes through when you run more treadwear environments, what made sense, there's this, you know, challenge with talent, right? And retaining talent and getting talent, and that that's not going to go away, you just go through all of that, too, it became but yet they had needs to could to run an environment like on prem. So it became you started to change his thinking to, I'm looking for that alternative to on prem. And it may not just be the hyperscalers, what's a different approach. So it became a different way of looking at it, which was a different perspective.

Guy 28:43

Exactly Rich in common, what Rich has one, I need to connect with you after this meeting. But to that's exactly the point is it's we're trying to not be having on prem because we were using a third-party data center now. And we're looking to move off our on prem data center to another third party data center. Because we really don't care about the heart rate. We don't care. We don't we don't want it. We know that it's a, you know, somebody else can do that better. We can have our people do better, more efficient things with their time than going out and patching things. Let somebody else do that. And that's exactly the problem we're facing right now. It's, it's not that we had to get, I'm not saying I have to get past but I'm having to socialize and get people to understand that we're not getting rid of the data center. We're just moving the data center. Go. It's really there. It's just it's not going to Azure, it's not going to be a segment. It's someplace else. And we're able to we're able to use it just as if it was ours. And that is a concept. The younger generation understands the older generation has a mindset that we have to be able to touch it. We have to be able to see it we have to be able to see the blinky lights, and the blinky lights need to go away.

Greg Irwin 29:56

That Guy Thank you Rich. How can I How can I Both afterwards, I'm going to circulate here a little bit more. Let's bring Eric. Eric, I know your your camera's on a bit earlier. I don't know if you're, if you're able to comment. Are you with us?

Eric 30:14

Yeah, I'm here, Greg, how are you?

Greg Irwin 30:17

Excellent. Thanks. So thanks for joining, give me give a quick intro here to the group.

Eric 30:21

Yeah, my name is Eric. I'm one of the directors of infrastructure here at. My area has been sort of carved out where I own about 700 plus retail and Doctor locations nationwide.

Greg Irwin 30:38

Why you have 700 locations? Are those owned by the franchise?

Eric 30:44

There are a combination of of owned locations or a combination of franchise locations and others, depending on what the regulations are, you know, in each given state.

Greg Irwin 30:58

Where does the it how do you support those locations from it? Is there a positive nature? You? You do hogs? How does it a great?

Eric 31:10

Yeah, great question. You know, we're actually running on a very old sort of antiquated point of sale system. It's really old, right? 20 plus years old. But we're in the process of rolling out a cloud solution to all those locations, we rolled out our beta sites, which were six and April, on another platform of one of our sister companies called. We're rolling out another 79 stores in October, October, November timeframe. And then another 650 stores in one fell swoop in early 2022.

Greg Irwin 31:46

Yeah, all right, what's, where's bare metal in your environment.

Eric 31:51

Um, you know, I mean, one of our big one of our big moves later this year, early next year will be to extend up to Equifax. And because we want to get out of the DC business, know where we, we have our own data centers, one of the challenges that we do have, as far as footprint goes, is that we have two large manufacturing facilities, which are our primary and backup data centers here in the San Antonio region.

And we manufacture eyeglasses, 24 by seven. So we'll always have to have a small footprint, you know, in those facilities to support some of, you know, the manufacturing machines that that make up that process. But we're trying to move anything and everything that we can out of those facilities to lower our our footprint there. As far as hardware perspective.

Greg Irwin 32:41

Right now, are you in terms of the actual rack virtualized virtualized or not? You know, is on your on a bare metal call or a bare metal discussion here? Is that a particular interest? Or is that just is this just is just, you know, just exploration?

Eric 33:00

I would say it's exploration more than anything else right now?

Greg Irwin 33:06

Yep. Makes sense. Makes sense. And one last one, we've been talking a lot about data centers in Texas, given the the power range last year. Has that changed any of your thinking? on where you want to put your compute?

Eric 33:21

Yeah, I mean, it absolutely does, you know, and for us, you know, Equinix is sort of a perfect fit. Because if you know anything about South Texas, all of the communications comes through the Dallas area down south Texas. So if you can get north of, of all of that real estate where something could happen, that's sort of our objective, you know, and, and also, you know, given the fact that all of our stores use, you know, our own Belo SD win, and we have, you know, different providers that, you know, locations all across the United States, and tying back, you know, to anything we might need, you know, up in an Equinix data center is just a just a smart move.

Greg Irwin 34:03

Yeah. I was telling us, we did a data center specific call last week where that was a big conversation was moving Dr. Vegas, Reno, of Phoenix, just out of the Texas power grid. So you have some diversification beyond that, for sure. Our Eric, thank you very much. Appreciate you joining in. And let's go over to Doug. Thanks for joining

Doug 34:27

me.

Doug 34:29

Give a little intro. Yeah. So I don't know. I'm a director in what we call our infrastructure solutions practice at. For those of you guys that don't know we're the biggest company that nobody's ever heard of. We're the fifth largest tax audit and consulting firm in the country. So KPMG, Deloitte, those guys were under five. Um, over the last three to four years I have focused on carve out some integrations in the private Equity space, also large scale technology transformations for under invested companies that, you know, decided they were going to squeeze it till it broke. So that's the world that I live in is that basically building companies from scratch or integrating them into into platforms and parents? Got it? Um, what's your interest in bare metal? So, I love all the stuff that I've heard so far today, Guy, I feel your pain, like worse than anybody in the entire world. And Eric, absolutely, I understand the whole, you know, manufacturing is always going to have to have stuff on premises. Um, what I find is my, my clients fall into two different categories. There are the ones that are just general commercial businesses with very light regulatory, or compliance obligations. And those guys are 1,000%. in somebody else's data center, we call it Bob, I define the cloud as not my data center. So it could be an Equinix, it could be, you know, Bob's computer stores, colo, it could be whatever, right? Where they're not owning anything, right? They're leasing or AWS, Azure, whatever the case might be. Now, for those clients that have compliance or regulatory obligations, almost exclusively, they are in a colo type of environment, right. So think Equinix, or a managed infrastructure as a service type of solution.

Doug 36:34

One of the major gaps is that, or one of the major areas that I hear a lot about is companies would like to focus in on consolidating down or moving their their workloads into one of the bigger cloud providers. Um, and what we run into is that if you have those compliance obligations, and you have the old school, IT folks that guy referred to, when he was talking about stuff, you know, you need that dedicated hardware to meet a lot of those requirements. And what I think bare metal as a service offers, is in these communications and news, you know, eventually, it'll hit the the larger cloud providers, it provides an opportunity for that, that dedicated hardware in a manner that is cloud scalable, were those those organizations that have compliance and regulatory obligations, so I'm not not much like everybody else here. But you know, I see, you know, we do hundreds of these transactions and carve outs a year, and, you know, community banks, we're taking them here, you know, somebody else, we're going to address somebody else, we're going to AWS, um, you know, I mentioned in the chat and three years, outside of like Microsoft deployment toolkit servers and domain controllers, I haven't deployed a server on premise and in three, four years, right, amazing. Amazing.

Greg Irwin 38:06

What's the so let's take, we can do a separate call and talk as you're an AWS, there's plenty of that happening, we're not going to pretend it's not. It is. It's a big part. But let's put that to the side. What are the challenges here? for the, for the physical world, the physical world for the non?

What did you call it? Not my datacenter? Yeah, let's call it my dataset, but the main data center world that's going into your it closet, as you said, You're not going in there unifying them in a managed service environment, you're deploying in colo?

Doug 38:47

What are the pain points? So, yeah, so And again, I look at things a little bit differently, like from the the business function of it, then then kind of the nuts and bolts A lot of times, Guy had a really big one early on, right? It's, it's capacity and capability of my staff. Right? So if I'm having to implement, you know, a really complex involve, you know, MRP or, or whatever the case might be your P system, whatever, warehouse management, right, and, you know, I, I don't want to have to pay people that, you know, $150,000 a year to keep blinky lights on, right? When I can go grab somebody who can, you know, configure software, do stuff like that for $65,000 a year. Right? So there's, there's a definite, like, cost of running hardware that is well beyond, like, anything that most people think about from their when they're computing what they call, you know what it costs operate. So when you think about your skill sets, plus, all of us old guys are getting old, and we're falling off and so, you know, if you guys remember back in The early 2000s as for hundreds, right? If you guys know anybody who knows as for hundreds, they get charged like $700 an hour, right, and we get it all day long, because there's six of them left in the United States. Right? That's what hardware is gonna turn into. So when you look at bare metal as a service, right, and being able to take that model that bad, man, well, there's the the Pure guy. All right, Jack, and Rich. Yeah, well, that he was talking about before, you want to bring that into kind of those manufacturing facilities and warehouse facilities where we need to have low latency high, you know, close to the business type of services, that's immensely valuable to organization. Because no matter what the big boys want to do, right, they're never going to be less than 30 milliseconds, most likely, probably closer to 50 to 100 milliseconds. Yeah, and a lot of times, that's, that ain't gonna work. So, and then on top of that, right, when you're you,

Doug 41:05

there are, we are always in forever will be subject to what our software providers architect, and they still build applications that scale up, not scale out. So having an application that resides in multiple data centers, and where I can get, you know, I don't have to go across the Rockies or go across the Atlantic, right, that that doesn't really exist these days, and a lot of core, like functions. And so you know, you're having to have those local, you know, data centers, or it closets or whatever you want to call them. And, and, you know, I guess the other the other big complication of this is that, you know, five years ago, eight years ago, we would have capex versus OPEX. conversations, right? And that's kind of what drove consumption and IT budgets anymore, I'm having cashflow conversations, and the last thing anybody wants to have is, hey, we're gonna implement, you know, I want to spend $4 million on this Cisco bomb, right? And then just so you know, in five years, or four years or three years, I'm gonna need probably $5 million to replace $3 million for spending this year. companies do not want to have that conversation. And anybody who's really an IT department does not want to have that conversation, even guy who's running, you know, seven zillion servers, like, that probably keeps them up at night. So to speak for your Guy.

Guy 42:33

You said exactly what I feel. So the payment

Greg Irwin 42:37

Guy is about you. But I also see the flip side of this. And not to go against the grain here. But you know, if you're going to spend the money somehow, and subscriptions, you know, add up? Obviously, the big sticker price of a big cap x is, is clearly not something you want. But you know, how much attention? Are we hitting a pain point yet on subscription?

Doug 43:12

Yeah, I mean, you are right. But that that's also the, you know, conversation shifting from that capex and OPEX conversation to cash flow, right? Where, yeah, man, we want, we want to capitalize as much as we can. But to come up with if I'm a $40 million business told with $3 million in cash, that sucks, right? Like, that's really hard to do. So your conversation turns into a cash flow conversation, I think there's absolute fatigue on subscriptions. And there has been for a couple of years, and that's probably a whole separate five phone calls with you guys to go through that stuff. But the, you know, I, and there's always going to be exceptions to every, you know, one of these, you know, things that I that I throw out there, but for the vast majority of my clients, and that don't have regulatory, or sorry, that do have regulatory or compliance, or they have that warehouse or, you know, manufacturing or, or whatever the case solution might need to be, right, anything that I can do to lower my costs of operating the infrastructure I'm going to take advantage of and that includes, right, the headcount and the and the training and all the stuff that needs to come with that to support that infrastructure. I think just one last thing to add on that I think, I think our our CFOs and controllers in our it directors and CIOs have gotten a lot more sophisticated in articulating the total cost of ownership of a solution and are able to put together those numbers in a much more holistic fashion than we thought when we saw 567 years ago. We Cloud started popping up and everyone is like, I'm just gonna move all my crap to the cloud.

Greg Irwin

So great sky. Alright, done good stuff. Thank you very much. By the way, I'll remind people, Doug's, Doug's, one of the best, I'll encourage people to connect across the grid here. It's, it's worth it, I promise. Let's keep going. Bob, let me get you into the mix. Do us a favor, give a little intro. Tell us a little bit about your products. Hi, thanks. Yeah. So

Bob 45:33

basically, my area of expertise at is in it economics. So when it comes to off Prem, on prem, bare metal, or just, you know, virtual instances, we look at applications case by case, there is no such solution, as you know, you know, there is no panacea and public cloud versus on prem, like everything moves there, or whatever. So, the we are a bank. And basically, data privacy is our core sort of thing, that we just can't compromise our client's information, putting it in public cloud. And our infosec is very adamant about that. So having said that, there are cases where we deal with public data, market data. And the case of capital markets and wealth management, there's like market causes they want to do. And the initial or initial foray into into public cloud was in around bursty applications where, for a very short period of time, maybe two to three hours, sometimes four hours, we needed, you know, several 1000s of cores of compute, and, and, you know, memory to process the data when the market closes, and, and, and then basically term Down, down, go down to a bare minimum. And so, in cases like those, we, you know, the economics was in favor of public cloud. And, and that's, that's when we started investigating the Furthermore, the question of Dr. is very much a, an issue, we do have two massive data centers that we've already paid for, and it's there, it's ours. And the problem is that our main production environment is in one location. And for every application, or every hardware, or bare metal, or whatever you want to call it, we have an equivalent set up hardware sitting in our other data center collecting dust. And every five years, we have to refresh those, you know, once the, once they fall off the books. And so the,

Bob 48:05

you know, osfi, I don't know, if you guys are familiar with the ospiti organization or office of Superintendent of Financial Institutions has, for certain types of applications and programs, they've mandated it, you know, that we would have compute and not just data backup by compute. Dr. And, and, and that is when we, you know, looked at the public cloud and said, Okay, well, you know, I can run the production in one data center and have on standby without paying anything to the service providers, a disaster recovery setup. And so No, no, I no longer have to buy computers, there is just, they're just there. If something happens to either Azure, or AWS or whoever else we're working with something happens to one data center, we can immediately come back up in the other one. And so that's that's where the savings come from. The bare metal thing, Kim's up, I came up recently, we never had bare metal situations till recently. And that was when it was built around the persistency of the applications being up. So so as you know, like, for example, if we just use AWS as an example, they have a an on demand model, they have a reserve model, they have a spot model, and they all have different rates for each one of the instances that you would get, and, and then. So we looked at those applications that were on demand, but they were always on and we just said, you know, why are you always on? If you're always on? Why don't you go into a cheaper solution, which is the reserve instance and run it there. I think it's 60% cheaper

Bob 50:00

And then when we looked at all of these application, we said, okay, well, it looks like these guys want to stay there. And the reserved instance is is is they've got it. Now we looked at the bare metal, we said, okay, why don't we just put these reserved instances internal, our own bare metal, and run it that way and have another bermeo on standby. But no pay for it in our data center for disaster recovery. So, so we're looking at that we're just saying, and then kivas is one one example of a we would create an environment, reserve some instances, and we just told people, if you want to run cables, just go into this environment and just do your thing, right. And then we looked at a sector, well, you know, what, I can do this much cheaper, if I, you know, get a dedicated bare metal, at a fraction of the cost of individual instances of, you know, either a shoe or Amazon's instant prices and said, like, you know, we get a whole box, and it's got so many cores, and so much memory and stuff like that. And then we can divvy it up amongst all the keyboard users. Instead of, you know, buying, you know, equivalent number of vcpus.

Bob 51:19

In in the in there, in the reserved instance model. So, so we're, we're course looking at that, and certainly did er, is basically where it's added, I don't have to go and buy another box and put it in our data center, or disaster recovery data center and just have them sit there till

Bob 51:45

I'm close to remove them and that DRS is it's one of the it's, it's, it's, yeah, we it's costing us a lot of money. But there is no plan or I don't foresee in any, any near future, anything to do how to happen with our data center. Because we're not just dealing with you know, our public data, we have private data privacy, and we have a lot of mainframe and somebody mentioned as 400 we have a series we have hv non stop. That is a dedicated sort of appliances that right now we don't see them in any of the public cloud offerings. Um, I have no idea of mainframe ever going to get there. But we were one of the largest mainframe

Bob 52:34

you know, sort of customers or clients for IBM in Canada? Actually, we are the largest in Canada,

Greg Irwin 52:42

but you're at the scale where you are your own clouding just make sense? I can?

Bob 52:45

Yes, we are. I don't see any any advantages into I mean, our data centers are just as large as any of the AWS data centers, right. So we do get the economies of scale, we do have but but it's the it's the mix. That's the issue. Like, you know, the most of the AWS and I assure our our x86 platforms, majority of it. They do have, you know, GPUs and stuff like that, but, but we have the MCs and we have it everything and and we've got, like I said HP non stop as 400 got mainframe, because of course, they say 6000 of them, and an AI x and Solaris and you You name it, different OSS. And it's not that, that, you know, we just, it's because of, you know, the applications that we bought, or went through acquisitions, we bought a company that that was an AI export, you know, focus company, another one that was, you know, and so just as a hodgepodge as a very, very heterogeneous environment.

Bob 53:51

But But, but having said that, it's we have moved roughly about, say, 10% of our x86 footprint into Cloud, public cloud. And like I said, most of the applications are bursty applications. The ones that, you know, the set a certain period of time during the during the day,

Greg Irwin 54:19

the squat, I mean, I, we're just a couple minutes away from wrapping up here. I'm sure I'm gonna I'm gonna cap cap this thread for a moment. And I want to do two things I want to do one is come back to Jack, and Rich and just ask the question on supply chain, which is something we've we haven't touched on here, which is the basics of some of this, particularly a networking gear can't get some next you just can't, you can't get them. So what's the availability in terms of decentralized purchasing them bare metal as a service to get customers Equipment faster than you might be able to get through your own distributor.

Greg Irwin 55:05

And then can't desponding you've been even joining listening quietly a month just to ask you if you have any questions or comments for the group. But no, no, I think the networking a complimentary perspective, I think the question you're asking is irrelevant. Okay. Perfect. Thank you, Jack, let's get at it.

Jack Hogan 55:32

Yeah, sure. So, you know, let's kind of talk about the supply chain aspect of this. So first of all, you know, without going to an accident, their global footprint, you know, they're very large purchaser in all kind of just say, you know, with the scale that they're at, at a network point of view, that, you know, we haven't seen any sort of supply chain challenges that are coming from their side. Similarly, on the peer side, and the guidance that we provided the street is that, you know, as of now, we're, you know, keeping a very close eye on things. And, you know, we're not seeing anything impacting our ability to deliver in the short term. And so, you know, the, but what, when you start to get into the supply chain conversation, you also start to talk about that ability to spin up on demand capabilities. And so, you know, the the conversation that the bomb just talked about, and we've talked about it a little bit, you know, around around Dr. The, the availability of compute the availability network is, in a DR environment, something that's not necessary only until the point you need it. But the data itself, and where your data sits, has to sit on a data platform that's available in your data in your disaster recovery zone, whether it's in or outside of your blast zone, and ultimately, where that data sets. And so you know, the term being kind of pilot lighting, your Dr. environment, that ability to place your data there, and have that compute available on demand from a scale of betania scalability. And whether it's birth stability, or scalability,

Jack Hogan 56:54

that is a critical capability that a bare metal service, and specifically what aquadex is providing that global availability, that is really, really key. And so you know, that ability to have multiple points of distribution for that, so 18, different data center, availability now. And then ultimately, that ability to, you know, have your data be persistent in those environments is what pure brain so you're not reliant on, you know, hard drives within the servers, you're actually relying on a data plane that pure provides, but the pier platform and so the two of those very, very nicely to solve a lot of the current, you know, supply chain problems where you can optimize for what is the only thing permanent in your IT environment, and it's the data and everything else is going to be swapped out over time. And you know, that example that was given earlier about, you know, you spend the $3 million, the first year, 5 million, the fifth, you know, is very real. And so you know, that it is about optimizing the utilization. And I'll make one final comment, cuz I know we're running up against time. And I want to go back to one of the early things that guy talked about, and Greg, you kind of took off on it. It's not bare metal versus virtualization. As a matter of fact, that example customer, the SaaS customer that I talked about earlier, the software licensing advantages they had was because they had extremely aggressive pricing for Microsoft, for Hyper V, that did not translate into the public cloud. So it also is about optimizing how you're deploying your software onto the physical layer as well. Because even if you're running up in a public cloud, there's still compute behind it somewhere. But if you can't put your licensing or you can't put your application, that's where this becomes a much better and more highly available model, you know, and again, supply chain kind of gets solved in

Greg Irwin 58:44

that to the big point that we haven't really touched on, but there are meaningful supply chain issues and certain skews. And it's nice to have that turn that into an SLA. Yeah, for some that somebody else can deliver against. We're gonna have to wrap here. So let me thank Rich. And Jack, let me ask everybody to take advantage of the group that we've got here. We'll do some follow up. And obviously, we'd love for people to dig in directly on the bare metal as a service offer from Equinix and Pure. Jack and Rich any closing comments here for

Rich Grandelli 59:14

now, just say, my point, thanks for everybody for joining and listening and interactive discussion and yours for hosting it. And I would just say, my closing comment would be just think about that alternative, as you go down your journey as everybody does, right? It's an alternative to on prem, how can I? How can I do it based on what I want to do. And there's other alternatives out there besides the cloud as part of it as just one element, but it's a journey. So anything we can do to help we're here?

Jack Hogan 59:41

Thanks. echo those sentiments, Greg, always a great conversation. You know, it's been phenomenal to hear from Michael and Guy and Doug, and the rest of the call here. This was a great conversation similar to the type of conversations we're having with customers all day long. And I you know, aside from us reaching out and you know, continue

Jack Hogan 1:00:00

During the conversation on a one-on-one format, any interest in this you can visit PureStorage.com/baremetal. You can learn more about what peers doing, you can specifically learn about what we're doing with Equinix as well. And so, you know, I think these are great ways to start conversations about how new and unique problems are being, you know, solved in it. And that's why I love these conversations, Greg, thank you.

Greg Irwin 1:00:27

Thanks, Jack. Thanks for Thanks, everyone, for joining. Thank you everyone for coming. Thanks, Bye, everyone.