DataCentric: Inside as-a-service & Dell Technologies APEX, with Dell's Adam Glick
Steve McDowell: Welcome to another data centric podcast. I'm Steve McDowell. I'm here with Matt Kimball. Hi, Matt. Stevie. Hey doing buddy. I'm doing well how things done last Matt Kimball: Not as hot as usual. So some interesting factoid it is the coldest July on record. Down here in Austin.
Steve McDowell: All things relative, right operating, but which means you broken 90 but not 100. Matt Kimball: Yeah, we haven't broken 90 much. So July still? Steve McDowell: Yes. Anyway, that's not why we joined the
podcast today. Oh, no, now we're in the podcast. So let me tee this up a little bit. So Matt, you know, we talked about we talked about this space a lot. But you know, five years ago, when we talked about IT infrastructure, we talked about on prem, self managed, and we talked about public cloud, and maybe there was some private cloud stuff happening. But those
were largely your options, right? But you look around today, Flash forward to 2021. And there's this whole spectrum, private cloud consumption based as a service, public cloud, but on prem. It's kind of dizzying, and maybe a little confusing for a lot of it, folks, right? About six weeks ago, six or eight weeks ago, at Dell tech world, when we talked about this at the time, Dell, the world's largest server and storage vendor, jumped into the fray with its apex offerings. And that's big, that was big, and we talked about at the time, and we speculated, but we thought it might be time to bring in somebody that knew what they were talking about. So to help us understand what's happening. In the space, we've invited Adam Glick, who's Senior Director for Dell Technologies Apex to talk to us and he's with us today. Thanks for joining us, Adam.
Adam Glick: Oh, it's my pleasure. Thanks for having me on the podcast. Tell us tell us what you do at Dell Technologies. My responsibilities are for the apex portfolio. So Apex is a portfolio of offerings that we have. And my job is, you know, if you put in a pot and boil it down, is to help people understand what it is to understand our customers needs, and to help bring those two pieces together to help make sure that our teams are building the right things to meet our customers needs. And that our
customers can understand what scenarios in what ways they can better do the jobs that they need to do and prepare themselves for the future with the effects portfolio. Steve McDowell: So we'll jump into what Apex is in a second. But something you said that you talked about, you know, meeting your customers requirements, meeting your customers demands. And every question I really have revolves around kind of this one core question, is there a fundamental shift happening? And it are the world of IT infrastructure that that has Dell Technologies, and then seemingly everyone else, kind of jumping into the space of both feet? What what are? What is it the customers are demanding? Adam Glick: For those of us that have been around it for a long time, like yourself, you know that there is always a fundamental shift happening? It's not so much that it's, you know, there's there's one tectonic shift that's happening as much as there is a constant shifting, and that is, that is the life and world of of it. I mean, think about the different
certifications that you've probably had over the years that, you know, when did you get your CCNA? And when did you get your, you know, cloud certification and your Windows Server certification, or Linux, like there's always change that's happening. But the big shift that we're seeing right now, is a shift from traditional infrastructure pieces into Cloud like infrastructure pieces. And I think cloud like because cloud is really kind of more of a an operating model, we can talk about that. But the idea of just setting up boxes and running boxes, you know, or say boxes, but really, you know, we're talking about blades and things in your racks, those days are still very much here for a lot of folks. But if you as you look in the future, there's less and less of that, as people move more and more to layers of abstraction in that. And that's the big change that we're seeing that whether it's on premises clouds, whether it's edge clouds, or it's, you know, multi cloud, public cloud, but this shifts towards more of a cloud based type of model and away from just pure hardware is the main shift that we're seeing.
Steve McDowell: And when you say, so this is a little bit this is not just a technology shift. This is, as you say, kind of a it's a business model shift, I think, for the people that play here, certainly an operational shift for the it, guys. But when you talk about cloud like experience, right, let's double click there just a little bit. Adam Glick: You know, three years ago, five years ago, cloud cloud was a location. And today, it's more a way of thinking
about infrastructure. So we use a cloud like experience. Yeah, I have some assumptions in my head around, you know, kind of invisible infrastructure and things like that. But what do you mean, when you say cloud like experience? I think of it a little sometimes you'll hear people start to say your cloud is an operating model. But when I think about is there certain things about cloud that people start to expect and they really sit around, like speed and agility that when you think about doing cloud based things, and you can think of this when you think.. Like VMware, what VMware did with with cloud and virtualized infrastructure of how fast could you stand up a virtual machine? If you went into vSphere? versus how long did it take you to actually get a physical box set up for whichever department or workload that you're supporting? That was a pretty big shift in terms of the timeline that it that it took them between those two? And then if you extrapolate that out and say, What if it's not just setting up VMs? But what if it's everything that you think about in terms of infrastructure? And what does that look like? And so that ability to kind of point click and have the things you need, that that has worked in that time, I mean, you know, we've seen virtualized infrastructure for you know, what, 15 years now, like, that's not super new. But the part that always kind of gated that is you
could spin up those things only within the kind of the concepts of, or the barriers of what you actually had as physical hardware. So once you start to see things like public cloud, it wasn't your hardware anymore. So like, you know, Microsoft, or Google or Amazon, were the ones worrying about that, you click on it, and they had to worry about that capacity planning piece, that wasn't your job anymore. But what did that mean, kind of, as you spread that out, as you know, VMware has expanded what it does, and Citrix expand what it does Microsoft and all these different players get into Kubernetes, if you want, but all these pieces of what does it mean to have that easy access to your resources to be able to do what you want to do when you need to do much faster than you have before. And that means
basically not worrying about things below a certain layer. And that's hard to do in a data center. You know, like, what happens when a network cable breaks, you know, I like the cables break, someone's got to go find like, it's not like there's a big red light that says, hey, this cable is broken, like, you know, you're gonna go out there with your cable tester, and you're going to test these things. And then, you
know, I always joke that the difference between the experienced it pro and the inexperienced, one is the inexperienced one throws in the trash can where invariably someone else comes by and goes, Oh, my network cable, pulls it up, puts it in the stack network cables to go back into a rack, where it's not working, the experienced one takes the cable and cuts it in half so that nobody does that. So with that transition of like moving towards, you know, more of those management pieces, taking that away making things more instance, and speeding up the processes, that's that shifts, as we look at it, Steve McDowell: A 50,000 foot view, right? When what is Apex what is what is Dell delivering, Adam Glick: go within your scope in your rocket ship, and you look down over Apex Apex is really what Dell is doing to bring all of Dells kind of IT folks products and our goal to try and bring all of our, our technology to our customers available as a service.So that's I mean, that's the kind of fast and simple way. And obviously we can we can drill down if you'd like to from there. But that's at the heart of it. It's you know, what does it mean, instead of going and buying a San, or NAS or a bunch a rack full of servers, network switches, if you just wanted to buy it as a surface? You know how you do that in other places? What does that mean, actually do that with hardware, that's what we're doing. Steve McDowell: So that requires me to deliver these cloud like outcome based experiences you're talking about requires a heavy investment in software and services and support.That, I
mean, to be blunt, right? We haven't seen that sort of thing from bill in the past, or we've seen it in bits and pieces, Dells making the investments to make all of this reality to make cloud invisible, if you will. Adam Glick: Absolutely. I mean, that I, I will, you know, kind of preface that a little bit by saying that Dell has actually done some of these things for years managed services wing that has done done managed services, engagements, and with some of the offerings that we've had, for a lot of our very high end customers, we did have things we call data center utility, where you could actually buy on a consumption based model. And
that would be you could literally outsource part or all of your data center to Dells operation and paper by the month. So we've had that. But it's been in fairly limited scenarios. And we're really looking at how do you bring that scale to customers. And you're right, it's a huge transition
for Dell, that everything is changing inside of Dell, from how we do our engineering to our product management, sales, marketing, HR, like there's a huge shift that's happening. Because to support that for our customers to make that real. It has to be all of those things like there has to be an organizational shift. Steve McDowell: In what I said wasn't a ding on Dell, I think we're Dell has always been extremely strong is a very closely track what's happening in the IP world. You don't jump
into early, you don't often jump in too late. You kind of like land right at the right time, which is, I think what's generated a lot of this interest in Apex because you're landing right now. And let's go on Apex a little bit. So you're looking at your announcement deck for Apex is broken into kind of five buckets, although I think some of these might overlap a little bit you have console, you have infrastructure services, cloud services, custom solutions, and then kind of partnerships. That's a lot to launch at one time. What what among these has you most excited? Adam Glick: Wow, that's like the last thing if like, you know what your favorite child which Yeah, which is your favorite child? The answer is you love them all equally. And in many cases, there are different things that stand out with each of those. So I'll, I'll kind of give a quick version of this, if
you think about, like, what each of those pieces are. So right now, Apex has, you have a console, basically just think of that. That's the, it's a web UI. And it's one place to go to manage and interact with all these pieces. Like that's, you know, I'll try and kind of do the high level pieces. And if you want to know more we can we can drill into them, but like Council is one place to be able to manage all this. So as you're
thinking about all the different things that you have dealt, like, how do you get to one console versus having you know, 20 different consoles that you've got, you know, you one for each year deployments, and you've got one for your servers, then you've got one for VMware that you're doing on top of the servers, and then you've got your one for your file system, you know, in your NAS work that you're gonna be doing in your final stores, and you got another one you're doing for your block storage. And like all of a sudden, you've got all these different things, they're adding complexity, this is our goal of trying to bring that together and reduce that complexity to make it easier for folks. And so you have one console to do everything from procurement, to do your monitoring, looking at at one place to go for that. And then within that, you've got these different areas that are available. So our cloud based offering cloud offerings, as they're called, put kind of cloud quotes, that's really VMware offerings. So if you want VMware, and it's VMware coupled
with our hardware, so basically, if you were thinking buying, like a VMware appliance, basically, and buying that, as a service, that you have it, you just say how many cores you want. And like, what's the ratio you want? Like, you know, Ram, two CPUs day, for instance, do you need GPUs if you're doing things like machine learning, or AI model training, and you just specify those things, and, bam, that's what you get. And it'll, you know, mean to have it, you know, plugged in set up and running in your site for two weeks, you know, so get that you can expand within five days, and there's some flexibility that we build in as well within that, but like, that's the cloud pieces. And then we also have a storage piece. So this is the one that this was the newest thing that we put out there in terms of Apex data storage services. And this is really a cool set of basically, you can have file a block storage, and you just pay for your baseline, and you can flex up and down, we do a bit, we basically what happens is we do a bunch of overprovisioning that you don't pay for, and then you only pay for what you use. And so if you flex up, you only pay for that initial amount, it's the same rate, you can, you know, increase the amount you want, and you pay for it by the month.
And so when you think about what do you need for storage, like the big shift that will happen for customers is one, you don't have to do as much capacity planning. Because we basically, we take the risk out of being wrong about that, because we've over provisioned and you're only paying for what you actually deal with. So that's, you know, Dell takes on that cost and that risk for the customer, you can always increase it, you can always grow with you, you want to shrink it back down, and you pay by the month. So rather than doing a really large payment upfront to be able to buy it and like hope that you're right or wrong about that, that measurement you pay each month.
So it's much more of it's the kind of the term of art is designed for operating expense, but basically means that you're paying each month rather than one big cost upfront. So if you've got a project, start that project, have it running, and then if it turns out that projects going gangbusters, great, you can expand that. And it turns out that it didn't really do it needed to you didn't buy a ton of storage that's just sitting there. And so like giving people that flexibility to scale up and down really important being able to pay us Joe. And then the custom solutions pieces are really like bringing Dells entire portfolio, like if you want something custom, whatever you want, we can structure it in a flexible as a service way. When I say flexible mean, like you can, you
can pay on consumption of what you're using, and flex up and down. And you can do that for all of our portfolio, our networking portfolio, our storage portfolio, our compute portfolio, you can structure, whatever it is you want within that. And the last one you mentioned was partnerships. And we mentioned that we've done a deal with equinix. Because we realized that we'll deliver this wherever people want it to be. So we'll bring it to your data center, because put it to your edge location. But a lot of folks are looking at, hey, as I go to as a service, you know, what do I do about data centers, because my data center turns out is not as a service like that as a physical facility that I actually need to you know, have and maintain. And the reality is that edge locations are not edge
locations, but um, colocation facilities, one, they're located everywhere, which makes it really great for things like geo expansion, to be able to get that, you know, get set up quickly, and to a lot of folks are looking at, hey, what if I wanted to have my data center that same way of like my facility, and pay for it, as you know, every month, and so being able to plug in with equinix, in this case, which is the largest colo provider in the world, and people have Apex services available there, give customers that same kind of flexibility of, hey, if you really want to move to that kind of an accounting model, and you want to have that flexibility, you want to expand quickly, like I know, within Dell, we were looking at doing some expansion into Asia. And we wanted to set up we had manufacturing there already. We wanted to get some of our services set up to do some edge processing there. But
getting the rights to build a data center like getting all the permitting done doing the construct. Like that's a multi year project. And the question is, did we want to wait three years to be able to have that set up and built so that we could start taking advantage of edge processing? In our own particular manufacturing? We're like, No, that doesn't make any sense. You know, what we looked at, hey, is color the right way to do it? Is it is there a closet, we can go into the facility and just set up an edge location where we can set up a few servers that we need to do what to do the things we need to do, and then port that back to one of our larger data centers to be able to do the processing. Like it's enabling those kinds of things. Remember, I talked about speed and agility, that's
what these are all really about empowering our customers to be able to do that, in what ways makes sense for them. Steve McDowell: So Equinix is a great partner in that respect, especially I think, for IT shops that are doing a lot of hybrid cloud and mixed with public cloud, because equinix tends to locate their data centers very close to the public cloud data centers, right, which just takes all the latency out. As you traverse the tip. Hey, someone asked us and this is more clarifying questions. And this confused me at launch. And I'm
still a little bit confused now. So Apex infrastructure, you talk about data storage services, you don't really talk about compute. So where does where does compute fit in here? Is it implicit in the cloud services piece? Or? You tell me right, how should I think about? How should I think about compute when I'm looking at a list of infrastructure services? Adam Glick: Sure, so I'm gonna, I'm gonna peel back the curtain a little bit here for you. Because in general, in apex, we like to talk about outcomes, what's the outcome you want, and spend less time talking about what's the actual technology that sits underneath, like, if you if you want to do that, I know as a as a techie in my own life, like, I care about those things, but like, we tried to make that as simple as possible. But we let's talk about that for a second, because that will probably help give you some clarity on that. So data storage
services, yeah, that's the storage pieces, when you look at what are called the cloud offers, what really sits underneath that is our HCI offerings. So there, there's VX rail that sits underneath that. And so you've got hyper converged infrastructure that's there that really brings the compute pieces. And that's coupled together with the VMware pieces. And so you Yes, you can go in and you can, you know, use
NSX. For instance, if you want to do virtual networking, and there's, you know, virtual storage there. But all of that is really in support of the compute pieces that lie there and what the VMs that you're doing. And so you may just need, like, you know, the attached drives in the block storage that are there, that may be good enough. But depending on the workload that you're doing, you may actually need really high performance, large scale storage, and that's where like ADSs and then connecting that into your VMware infrastructure that's running on top of, you know, what's called hybrid or private cloud offerings is a great way to kind of make those fits. When you think about compute. We think about that in terms of, Hey, what are you going to do with that compute? And so that's a close partnership between us and VMware to give a connected solution there. Got it? Got it?
Matt Kimball: I have a question for you real quickly, if you don't mind. You talked a lot about speed and agility, speed and agility, time to value right, which is what that all kind of. And as you look at, you know, the marketplace and the different types of industries and verticals that Dell plays in, have you found? I know, it's fairly early, but if you found, you know, quicker adoption in one specific vertical versus another, or one company type versus another? Adam Glick: So the short answer to that is no, that we've seen really broad based interests just across the spectrum, what I'll see is, it's interesting, the different questions that you get from different industries. And so we've seen a lot of interest from people in banking and healthcare, that like highly regulated industries that need to be able to maintain control of data, and really care about where that data is located. So like people who have a lot of interests or curiosity around data sovereignty, or they have corporate role like, this becomes really important to them, because their reasons why they can't put their data or use other places for it. But like having that and making sure that they maintain full control of that is super important to them. So although I mean, I could give
you examples from startups all the way up to you know, media and entertainment, there are like when you start to talk about highly regulated organizations and people in industries like that people tend to get really interested in what this can mean for them, because it gives them the set of cloud experiences that people want, that they've normally had to like, say, that's great, but I can't do that. Because someone else's can be touching my data, managing it, you know, for I was talking to a sheriff's department the other day, and they're like, Hey, we can't lose chain of chain of custody, like that data needs to be within our facility. And our people are the only ones who touch it, we manage control of that, or it can't be used in court. Steve McDowell: So Cloud is, you know, when I think about it, infrastructure, we have core, we have cloud, you know, this whole model kind of sits in between that a little bit given kind of the best of both worlds to turn it consumer. But they also have edge, right, and we're pushing compute and storage and things like that increasingly, out of our data center and even out of our clouds data center, where we can operate on data, you know, where where it's where it's generated, essentially, does Apex fit into that model, though, is there an edge story with Apex? Adam Glick: There is. It's one of the real big things that we hear people ask about. And it just, I mean, the short kind of
view of it is, like I said, we'll deliver these anywhere. And so for customers, what that means is, hey, you want to go stand up an edge location. And remember that, that we do management on this as well. And so we don't, we don't touch
their their data layer above, but we manage the hardware layer. So we have organizations, we work with a railway company. And they're actually they do their inspection, their rail cars, tracks, they actually set up their these boxes, and the boxes are racks of servers, and storage. And they're sitting there, they're monitoring the cars. And it's rather than the staff to go and actually inspect physically each of the machines, and they have basically a vision system that looks and monitors the car, while it drives past and analyzes the entire train looking for maintenance issues or other things they need to think about. Like that is the kind of definition of an edge scenario. And going and doing that building. On top of that,
as one example, edge is one of the ones I always like get excited about because people talk about it, like it's this monolithic thing like the edge. And that's a little bit like saying, you know, oh, it's an application or workload, like there are 1000s of them. And the same is true, when we talk to our customers, it is amazing how many different scenarios, edge is unlocking for them. Because it just before it wasn't practical, like if you just think about the you know, if you were gonna set up that it environment safer train instruction, like, here's the challenge you have, once you get once you have all the technology, let's say that you've you know, solved the ML monitor models and the vision system you've built, you've built those environments. Now, what happens when you need to patch it? What happens when a drive dies, like and depending on the size of your organization, you may not make sense to staff a team that's near that. And so all of a sudden, you know, if you've got
something that happens, what's your SLA for being able to respond to something that happens? And that's one of the benefits that Dell brings is the ability to do management of those pieces. So customers don't have to because if you're dealing at software layer, then guess what, you can probably do that over the network. And you don't need to send a team to wherever it is you're at, you know, when you say wherever it is, I mean, every oil platform that's out there, has server racks on it. But their network connection isn't so good. You know that they need to think about how does that work autonomously? Because in most cases, they work over satellite, their satellite connection, what about, you know, cruise ships that are out there, and I say this, and you're like, Oh, these, each of these is a one off case. And that's the whole
point is that there's all they're all one off cases the same way that like every IP environment, like everyone comes to you and says, Oh, well, that's just you know, this 3g architecture, oh, you just install this isn't in isolation. That's true. But in reality, they're all unique and different, because they're, they're all things that people need to work on. And so if you look at like, every airplane, that's out there as effectively a flying data center, at this point, even, you know, people's cars, Tesla's have enough computers and drives in them, that what you're seeing is edge locations are essentially everywhere. And they're only growing because the power of it, and the benefit that it can bring in all these scenarios becomes greater and greater. And
so at a certain point, the challenge becomes not do we have the technology to do it. But do we have the staff and the ability to scale to make it happen. And one of the benefits with Apex is that we take care of that hardware layer, because Dell is a very large organization with people spread around the world. And so if Dell can worry about, Hey, is the network cables running? Did drivers die? Do we need to swap it out? Is there a firmware upgrade that needs to happen, and you can focus on the software layer, then you can just say, Hey, I don't need to staff a team, because we just you know, we did a new deployment over in Ketchikan, like my centralized IT team can manage that and use the tools that they're used to in the scripts that they're used to. And so we can scale organizations and operations, instead of being a bottleneck, we can empower our organization to move faster in those pieces while still being able to maintain the control that we want. So that that somewhat of a
long answer like that I get really excited about edge, because there's just so many cool things that are happening at the edge right now. But that's one of the benefits is like you don't need to deploy I mean, you know, when someone does a deployment, you're talking about things that are relatively small, in terms of like the number of cores that you're going to do or the amount of storage, like you'll say you do a storage deployment of say, you know, 50 terabytes. So yeah, it's not exactly a laptop. But you know, if you if all you need the laptop to go to a laptop, or go to an IoT implementation, but when you take a look at, like, you know, what you can do with that the example I gave with the trains, like, it is shocking the amount of data that generates that literally creates something like, you know, a terabit a minute or something that they're they're running through in processing, they're just, there's no way that you would have a network connection that would be big enough to be able to do that as an IoT scenario. You need to actually have an edge location and process that they are and that's the kind of stuff that you can unlock in terms of value for organizations, that makes a lot of sense.
Steve McDowell: And sometimes you know, as IT people, we get very tunnel vision into the things we know. And thinking about edge as an extension of my data center and another location, you know, it's a rack, it's just a little further away, than it makes sense that Apex is going to play here, and you're going to deliver the product there. And you know, what I like about what you're offering with apex. And the edge in particular, is, you know, HCI is such a nice operating model for those environments, because that really gives me the ability to manage all of those resources to kind of a single pane of glass.
For everything, everything you said there resonatestremendous tremendously. Let's circle back to Equinix, you launched Apex with one partner equinix is that the first of many, how key our partnerships and data center partnerships, but MSP channel relationships to your success moving forward? Adam Glick: Sure, so at the risk of sounding like a business guy for a second hand, and ostracizing myself with every possible tech person out there. Like that, I used to spend a lot of time being an engineer. So I well, familiar. partners are key to to the, just the industry in general and to our strategy in specific, that if you take a look at, you know, equinix is is a super important one, because they are so big. And so we really wanted to go out with that. But they're the first of many. And they're, they're big in terms of like, what it means
for people to be able to look at where they want to do their deployments. But the reality is a Dell has an ecosystem that encompasses partners across the spectrum, from you know, people that are building software that works with our systems, to people that are helping implement the software and the systems, the people who are managing and running services on top of them everywhere from people's own service providers, we have a lot of people who are in of themselves service providers, they provide SAS, and they're looking for this as a way to, to manage that. So incredibly important to that I, for the sake of my friends over on the partner channel team, who worked really hard like they, they spent a ton of time working on how do they make this as lucrative as possible for the partner channel. And so we have a very lucrative program for our partners to make sure that they're along with this because it's a platform play. When you think about it, when you think about what a platform is, I will think I'll steal this from I think it's Bill Gates that said it but like, you know, when you build a platform, 90% of the value is created for the ecosystem built on top of it. But by building that incredible
platform that so much of the ecosystem can build on that 10% is far more valuable than owning that whole piece. Like it's not an application. It's about building that platform that people can build everything that they need. And we've done partnerships, we've announced another one with PTC, so they do smart manufacturing technology. So if you look at like edge manufacturing, and how do people do smart manufacturing technology, we did a partnership with PTC, we have a reference design for that. So people can roll that out on top of the apex hardware, and really looking at how do we continue to grow that of like, we're not going to be the experts on that PTC is the experts on that. And how do we partner with those folks to make
sure that we create solutions that work for customers and help them be able to achieve their needs. Because at the end of the day, that's the goal that we're trying to achieve. It's customer success is our success. You know, what I'd say internally to folks is like, this is an incredible opportunity for us to Dell because never have we been more connected to our customer success. People pay us by the month. And so if they're not succeeding, we're not succeeding. And I'm really excited about that, because it will help us be focused on our customers. And it aligns our incentives with their incentives.
Steve McDowell: AI, machine learning, deep learning, becoming a core part of IT infrastructure and IP services. And it's also something that we're starting to see delivered as a service just because of the you know, startup costs are so high sometimes. How do you how do you look at that space from from through an apex lens. Adam Glick: AI and ml is such an exciting area. And I had some
friends that I used to work with. And I used to say to them, they worked directly in the AI and ml space. And I say you realize when you talk about that, you can just replace every time you say AI NML with magic, because people stay at so broadly these days that nobody knows what it means. And so and the reality is because there's so many different ways that it can be used. So I'm excited for what Apex does, because it serves basically the two directions that I think about with AI/ML. The first is what are the AI and ml applications that people are going to build that they are going to use. So
you know whether they're using pytorch or they're using TensorFlow or they're using, you know, an already built application like people will build incredible things using AI NML. And what is the hardware that needs to sit underneath that in many cases, those things usually when people are training their models, they want to use GPUs. And so you can get Apex compute infrastructure that comes with GPU and you can specify how many GPUs you want to be able to do that and to process as fast as possible because training your models is a really challenging task takes a ton of data, if you've, if you've never done it, I'd encourage you to try it out. Like it's a fun kind of hello world experiment to go and kind of build one of those. And you'll learn a lot about what it really means to do it, I can chat with you a little bit about my experience, when I went, I went through and did that. But
first, as you have the hardware to be able to do those pieces to both, you know, do the model training, and then actually, like run the models and running the models actually isn't that hard. But it's the training that that really requires that computing, we actually have built specific parts of the apex compute stack, that those cloud ones that are defined for use in cases like AI and ml. So I said, that's the first one, like the applications that people are building, and people will build all sorts of unique things for their particular needs, their development teams will build them. The second part is what we're doing with it. And we do a bunch of AI and ml stuff,
because remember, we're managing it. And so it's in our own best interests to be able to identify when will the drive fail? For instance, when will you see hot spotting? What is the you know, what is the maintenance log look like for particular hardware for particular organizations for particular network types of flow? And how do we use that data to be smarter about providing the best available services to our customers, to give them the highest availability, the highest uptime highest performance, because we have an incentive to do that, because we make an SLA with our customers. And we want to make sure that we can exceed that SLA, and that we can always be getting better. Because again, our son, our customer success is our success. So there's those two pieces of what we do within internally, we use Cloud IQ is our our kind of monitoring suite. And that's built into what Apex is. And when you go to
the apex console, there's a link for you to be able to see those pieces. But then there's also as people will build AI and ml workloads on top of that, making sure that there's Apex specific hardware that can help for those particular kinds of applications. And we do both. Steve McDowell: Alright, so let's kind of wrap this up. Let's look forward. You know, as you look out over three to five years now, that's a long horizon. But you know, my magic
and put on your fortune teller hat, right? Do you have a vision for what the space looks like in five years, there's the there's the own your own on-prem model ever go away? Adam Glick: Well, I've been around technology long enough to to be wary of any times the ones making Yo, yo, here's what the future will look like. But when we think about it, when I think about like what it could be, does the on your own to answer your first question or your last question there? Does the own your own model go away? I think the answer is no. But it becomes less of the norm for what people are doing more and more people are moving to doing pieces as a service. And for more and more customers that make sense. Will there always be exceptions to that? Yes, there probably will be so much like, you know, the example I'll give you pardon me, I love color. Colorful analogy
is fun. Like, you know, I bet you that there are less people making horseshoes today, and doing horseshoe fitting than there were 150 years ago. But there are still people who make horseshoes and do our shoe fitting. So there will still be unique and specific scenarios where people want to do that. I mean, if some of the examples I can think of are things like national security, the NSA within their particular facilities probably is very particular about who has physical access to the location and who can go and set up an exercise. And so they might be an organization who would be interested in Hey, I still just want to buy this thing. Like I
want it and I don't want anyone to touch it, you know, anyone doing those kinds of pieces? They might have a need for it. But I think more and more businesses will look at and say is this the core thing that I mean, you know, do I care about racking and stacking servers? Do I care about swapping network cables, you know, fixing swapping out drives that are broken, doing firmware upgrades like is this, we spend so much of our time doing it? Is this the best use of our team's time? Or can we spend our time doing things that really matter to our customers? Can we get out of that Yak shaving sort of experience? So as I look forward, I think you'll see more and more people shifting to do more and more of their workloads. In as a service. I think that shift will be slow over time, like most technology, you know, transitions, you know, if you're an IT and you've been around for a while, you probably don't want to jump in whole hog on anything, because you've, you've seen how that can backfire. And so what you're seeing is you're seeing people pretty clearly move towards a cloud based model. And what's interesting is you see that it's
not particularly clear, are they moving towards the public cloud or private cloud, what you actually see is the percentage of workloads that are moving around to each of them is about equal, that people are figuring out what are the right places for them, to put their particular workloads and to be able to do what they want to do. But the one thing that is clear, is the kind of traditional model of doing things is decreasing year over year. And so as we look at it, it's how do we continue to enable those customers to do what they want to do to help them meet the needs that they have in the future? And I think that you'll see that moving forward, you'll see more and more as a service, you'll see more and more focus less on the technology, you know, what is the particular product and much more on what is the outcome Come, hey, what is the SLA that I'm getting? What is the guarantee that I'm going to get? And who are the vendors who can deliver that to me? So how can I invest in those particular people to make sure that we've got a tight partnership going forward for joint success? Yeah. And I think I agree with that view. And certainly as I mean, what we're all enabling here is, you know, as Greenfield developments happen, and as upgrade cycles roll through, right, you're forcing a very deliberate conversation within it already. Minority IP architects, and certainly the CIO staff, right, is what's the model I want to move to? Over and above? What's the technology I want to deploy? Steve McDowell: You say it all of your decks, right, it's all outcome based. And that's a great way to think about it. And
this model, I think, enables IT folks to think about outcomes, much less than technology. And that's a good place to go. Adam Glick: I think a little bit about what it means for some of the folks in terms of how does that transition happen to like, work in it, you there's no shortage of technologies that you need to know when you have to manage and the complexity that comes with that? And how do you make that transition? But you know, we have, I've had this conversation with a number of folks about like, Hey, does it still exist in the future, for instance? Or is it all just, you know, software defined and development folks, and everyone becomes a developer and it goes away? And yeah, look at that. And I say, well, that's, you know, sounds to me a little bit like the discussions they used to have at the turn of the century were like, you know, hey, the internet is here. You know, there's the there's, you know, search engines, do librarians exist in the future, or the librarians basically disappear? Well, you know, that newsflash, librarians are still there, they're still super important. But jobs change. And you know, what the value that they do and how they present it shift? And what I look to is, how can we make sure that we make that transition, as easy as possible for our customers that, hey, what if you can get all the benefits of moving to this more agile, fast moving model, without having to learn a ton of new technologies, throw out all of your procedures and your policy or go spin up a completely different department that's going to go focus? Hey, these are the folks that all new all this new technology. And then here's a separate department of people that know all the old technology, by the way, who wants to work in that, you know, who wants to work in which department? Because each one sees kind of where it goes? versus how do you bring people along to make it as easy as possible to make that transition? Steve McDowell: I think that's a great place to wrap up. I'm
excited about the story. I'm excited about whole as a service consumption based model right across the industry and what everybody's delivering. Matt and I were both excited to see how that plays out. Is there anything we didn't bring up? Do
you want to talk about? Give me kind of the last question. Adam Glick: Oh, boy, that's a that's a blue sky question. There's lots of things that run mine, really just like the what the as a service transition means, like, if you think about it for for your audience, and for the people that listen, and understanding that moving to as a service is not the same thing. It's just Hey, you know, okay, I'm going to pay by the month versus like, paying up front. And now I'm just gonna have whatever it is the workload or the you know, the infrastructure that I need, but really taking the time as they make this transition within their organizations to think about what does it mean for them? What are the things that they want to hand off? What are things that they're not ready to hand off, and being planful about those, just in general, like a good kind of way to think about it is like, pick something and try it. And you know, the kind of methodology I used to use is pick something, you know, that's disconnected and non critical path to start with. Because if it turns out that like, you
didn't plan for things, you're you're figuring out and broke something like your critical systems didn't go down, and you can figure it out and kind of use that as a transitional way to then say, Okay, now I'm gonna do something that's connected, but critical path, or non critical path, and then like, work your way into it. But this transition is definitely happening. Like, if you look across the industry, in any one of your, your listeners, I'm sure it's already somewhere along this path of doing it. And so figuring out what it means and finding out what's right for them, is really the thing that's most important, because it's not just the technology piece, it will impact your people, it will impact your process. And the
sooner that you figure out what that means and how it works for you and your organization, the more successful you'll be, because what you don't want is you don't want to jump into something and be unsuccessful. So you know, whether it's what we're doing with as a service, or really just what's happening in the industry as a whole taking that planful approach would be the thing that I would kind of leave people with. Steve McDowell: Okay, let's, let's leave them there. Thanks
for thanks for joining us today on data centric, this was a great conversation. And we're gonna ask you back in May a couple of quarters or as your rollout updates and announcements and evolve this forward, because, again, it's a space where I think our listeners are actively interested in and we certainly are. Adam Glick: So thanks. Thank you for joining us. It was my pleasure. Thank you so much.
Steve McDowell: All right. And Matt, I think this wraps up another, another podcast. Matt Kimball: Another week, another podcast.
2021-08-08 03:03
