X as a service (XaaS) has become a popular consumption model for organizations. There are a lot of benefits including variable costs, flexibility, and freeing up resources to create and innovate.
On this episode, Mitch Prust joins Account Executive, Dan Frank to discuss XaaS. They break down the considerations and dive in to HPE’s Greenlake Cloud Services.
This podcast content was created prior to our rebrand and may contain references to our previous name (OST) and brand elements. Although our brand has changed, the information shared continues to be relevant and valuable.
Andrew Powell: Hey, everybody. This episode of “Ten Thousand Feet” brings a new voice to the table, Mitch Prust, our newly hired VP of enterprise solutions. Mitch and veteran OST Account Executive, Dan Frank, talk about XaaS as a service and approach growing in popularity with enterprise IT. They break down the benefits and considerations, including taking a deep dive into HPE’s GreenLake cloud services. Enjoy.
Mitch Prust: Well, Dan, today, we’re going to talk about everything as a service. I know it’s a hot area. What sort of things are you seeing?
View Full Transcript
Dan Frank: Yeah, it is a hot area. And particularly in the last six or seven months when the entire world is changed here. Companies have been struggling to keep up with demand and move quickly, and they’ve been exploring things like everything as a service. We’ve seen over the last couple of years an introduction and an acceptance of everything as a service for several reasons, some of them digital transformation, for example, that businesses are trying to adapt to how customers want to interact with their organization, putting tremendous demands on their infrastructure, changing financial climate is another one. The lease structures changed in 2019, which altered the way organizations can either capitalize or operationally expense leases, so that has opened up the whole OPEX versus CAPEX side of things for acquiring compute and storage has changed quite a bit. Application proliferation. So there’s just—there used to be the same what there’s an app for that, now there’s like 10,000 apps for that, so as organizations are either creating more applications or acquiring applications, cloud-based applications like Salesforce or Microsoft, and they’re having to deal with the application proliferation within the organization, how do they manage the infrastructure that goes behind those, not the cloud-based ones like Azure and Salesforce, those are all cloud based already, but if they have a particular custom application that they’ve created for their customers, how did they manage those sorts of applications? And then data growth, as we all know, you know, over the last 20 years or so, the data growth has just been phenomenal. If you’re like me, I don’t throw anything out in my email inbox, in fact, right now, I think I have 31,000—nope, I’m sorry—28,972 messages in my inbox, so I’m just the worst example of a person who never throws anything away. So that’s just as a consumer, but organizations are gathering data from multiple different data points now, and data is the new oil, so the ability to be able to collect and retain and analyze and gain insights from that data is extremely important for companies to be competitive and to be relevant with their customers and acquiring storage and acquiring the compute capacities to analyze that data in the historical way of simple capital purchase can take an awfully long time, sometimes, you know, three to six months within an organization before they go through an RFP and they figure out and what exactly they do they need technically, they understand how much they actually need, then the budgeting comes into play. That all takes time, that this internet based economy, and instant on society doesn’t have any longer, so everything is moving much, much faster with the—as a service model, and there’s a couple of different ways that it works, but if you look at a model where they’ll have hardware that’s actually onsite at an organization, you can expand or flex into the capacity or the compute as you need it, so it’s almost literally like pressing a button and you get additional computer, additional storage. It’s really as easy as that versus having to go through a three to six month long RFP and budget cycle, et cetera, et cetera. So those are some of the reasons, one of the other big reasons are as organizations have adopted cloud computing, traditional hardware manufacturers have struggled to figure out how are they going to be offering cloud-like services, which companies have really come to enjoy, and still remain—yet still provide the infrastructure and that’s onsite that can handle workloads that just aren’t capable of being utilized in the cloud. So that is a real big reason why many manufacturers are now offering, quote unquote, everything as a service or infrastructure as a service or infrastructure as code. You’ll hear all these different sorts of terms.
Mitch Prust: Yeah, they sure seem like they’re kind of flipping their model of selling infrastructure, but also really wrapping together hardware, software, and services. And maybe one of the best examples of that is, well, the more prominent one is HPE’s GreeLake solution, you know, where they’ve really helped be a leader in kind of representing that shift, and being able to provide that sort of the service, like I said, where they can flip a switch, they flipped their model, now it’s kind of more on a consumption base sale.
Dan Frank: Exactly. It becomes the service, right? You’re no longer acquiring hardware. I don’t—I don’t really—it’s the old phrase I used to use: nobody wants braces, but everybody wants straight teeth, right? If I—I don’t want a server, I don’t want a storage array, I want the compute power that goes behind that. I want the ability to be able to store my data and sort it and analyze it. I don’t really want the hardware, it’s just a necessary evil, in this case, they provide this as a service. So you look at, you know, you mentioned HPE GreenLake, I was at the HPE discover event in June of 2019 back when we were able to get together as big conferences, and HPE’s CEO came on stage for the opening keynote address and made a pretty bold announcement that everything HPE offers—hardware and services—will be available as a service by 2022. Now that was in 2019, that’s only a three-year, I’m quick on math, three-year timeline to put everything as a service. Now, why would he do that? And that was a—that’s a big deal, but it does come down to the remaining, relevant, moving to where the puck is going to be—we’re both hockey players, the Wayne Gretzky saying, right? So you need to be able to set up how—where are our customers going, where are the end user customers going, where are the organizations looking to expand and gain flexibility and agility, and by offering it as a service of not just compute and storage, but also as you mentioned, the services that go behind it, it allows the organizations to focus more on what it is that they’re actually providing for their customers or creating for their customers versus trying to keep, you know, the blinking lights on and making sure that services are up and available for their customers. They can focus on the real value that drives their business.
Mitch Prust: So Dan, I was at I think the previous announcement for HPE GreenLake that happened back in 2017, and I could see kind of what they’re doing, and they’re really starting to make the shift towards some of these accounting models that, you know, the accountants were driving, actually, some of maybe this early transition, so I saw the announcement of GreenLake, just trying to kind of keep up with some of that, but meanwhile, trying to compete also with the cloud providers and pay as you go kind of service, so, you know, since 2017, what I’ve seen is, you know, GreenLake kind of went through this phase where everyone was trying to sort out the actual model and the solutions, so there’s this period of rarely understood niche that it fit into primarily as a direct sale by HPE into some of the larger accounts, but more recently, what I’ve seen is this emerging use cases where GreenLake is a better fit, and with the announcement of GreenLake Central and this Ezmeral kind of wrapper, you know, brings us to today where we’re kind of here and there’s these reference architectures that are available, there’s GreenLake Central for kind of managing it, there’s a wrapper, so I’m a lot more excited about where GreenLake is now, and our ability as OST to kinda sell it becomes a whole lot more clear.
Dan Frank: Yeah. Agree to hundred percent. It has been around for a while, so as a service has been around for a while, it’s not necessarily new, but how it’s being refined within organizations, particularly you mentioned HPE with GreenLake, so other manufacturers offer in as a service model to, but they really seem to be all in on this space. From personal experience, you look at the time to pull together what the offerings are for the customers is reduced quite a bit here in the last even 12 months, so when your CEO comes on stage and says, we’re going to make everything available as a service, everyone gets in line, and I’ve really seen that come through from the field and come through to delivering solutions to a customer, in fact, I’ve got a presentation tomorrow with regards to that.
When you look at what the advantages are for a customer, when they look at cloud offerings, for example, AWS, Google, Microsoft, they provide—they really started to change people’s focus to the actual service that I’m able to provide to my end-user customer, whether as an internal employee or an end-user customer, and that started to change people’s view of, ah yes, how am I providing a benefit to my customer rather than, hey, I just—I can’t get my email or I’m just keeping the lights on or keeping my systems going is how am I becoming—enabling my organization to provide a competitive differentiator versus my competition or to develop a new service that’s deployed faster and can change quicker and can expand more rapidly, whether it’s either locally and then to expand globally.
But there’s always certain things that a true traditional 100% cloud can’t provide, and one of those challenges is in latency, so just the length of time it takes for light to travel that’s one of those, you know, physics laws that we just can’t seem to get past. There is a tendency for—there is a limit on the speed of light, so certain workloads cannot accept latent reactions from their applications, I just don’t work that way. So as much as somebody might have a desire to put a hundred percent of everything in their organization in the cloud, they may have applications that just have to have local access, whether it’s a database or an end-user customer interaction, and that’s where the beauty of these as a service models come in, and again, we talk about GreenLake, it has the ability to have the hardware onsite, and then it can also flex into and leverage the cloud, so if you wanted to leverage AWS is glacier, for example, for offloading old data files, for example, that you don’t frequently use, you can still do that as well. You can still have a mixture of hybrid, really, between leveraging a traditional cloud providers or the hyperscalers, as a lot of people call them, and then as well keeping certain applications onsite, whether it’s governance, whether it’s security, whether it’s speed of light restrictions, because of latency, these sorts of as a service models allow you the flexibility to be able to do those sorts of things.
Mitch Prust: It seems like these 17 kind of configurations that can ride on top of a GreenLake environment are starting to make a lot of sense, right? So the 17, I’ve never been able to count all 17, but they’re kind of broken down into a few different areas where you can have these—you can get a GreenLake solution, and as soon as like 14 days, that will provide you kind of a platform for housing your containers, your VMs, your storage, and also your compute, so—and then have kind of it managed in a hybrid environment. Sounds like a great, great solution. I’m really glad that HPE came up with these reference architectures of these 17 different configurations.
Dan Frank: Yeah. Originally, you know, when GreenLake rolled out, it was a lot of it was aimed at the large enterprise customers, and that’s kind of where they cut their teeth in those programs. There was a kind of a fertile ground for those organizations, but as they’ve added these offerings, they’ve added these reference architectures, they’ve developed out the program to become applicable to more organizations, and it’s just an opportunity to reevaluate how your organization will view the infrastructure, and through these different sorts of reference architectures that makes it easy to do a quick analysis and understand, yep, I think this makes sense for our company, or, no, we have a particular application or geography or requirement that wouldn’t allow it. That’s okay. It’s not an all or nothing. You can customize what you need, and then, you know, customize it as you like kind of tailor fit for your organization.
Mitch Prust: Let’s walk through some of that reference architectures that are kind of out there, you know, so GreenLake can provide general purpose VMs, enterprise ready VMs, and those can be based on, you know, Primera or Nimble kind of storage. There’s also hyperconverged VMs that can provide, can use either like SimpliVity or also even Nutanix software defined platforms within it. There’s private cloud VMs, there’s containers, there’s machine learning libraries that can be deployed, there’s data protection, general purpose storage, business critical storage, mission critical storage, composable compute, right? Another hot area for HPE. There’s general compute, so I think—and more—I think these are like little fires that HPE has started to kind of plant in people’s brain and how you can actually make use of it, ‘cause like, as you said, when it first came out, it was like, could kind of do everything, but very custom, but here it’s almost like they have these, you know, the reference architectures can kind of get you to actually deploying a solution that you can use quickly.
Dan Frank: And in one of the additional benefits of that is it allows your IT staff to focus more of their efforts, again, on those value added projects that the business lines are demanding, right? So I’m no longer spending all this time going through the—you still want to do enterprise architecture, obviously, but they have a good starting point to work from. There’s no longer that long life cycle of going through the RFP process and evaluation process. It allows the staff to focus on higher value things. You hire people as an organization, your IT staff, because they’ve got, you know, a good brain, between their ears. You don’t want them necessarily doing some of these things that simply either keep the lights on or don’t provide true value to the customer, to the business customer within, and that’s one of the things that we found from these benefits, they’ve done several studies on what are some of the benefits with a 65% reduction in time to deploy global IT projects, and that’s exactly where something like these HPE reference architectures can really help get you started off quickly, and your time to market is much greatly reduced. And then I’ve referenced earlier about just spending time keeping the lights on 44% spent less time, quote unquote, keeping the lights on, and that was relevant, particularly when a managed service was part of the solution. That’s something we haven’t talked about too much, is, some of these as a service offerings offer managed services as well, so not just the storage and not just the hardware or the compute, but also the services that go behind those. So you can, you know, literally have the organization do your monitoring and management of your infrastructure, and that provides, again, that band frees up that bandwidth for your IT staff to continue to focus and work on projects that are of higher value to the organization and your customers.
Mitch Prust: Let’s talk about, you know, one of the latest announcements that HPE discover was around their Ezmeral container platform, so I think this is a significant addition to the GreenLake solution, because Ezmeral, as I understand it, really is a container platform play that helps a business kind of take a non-cloud native application. Put the Ezmeral wrapper container around it and allow it to reside on the GreenLake environment, and be quite a bit more manageable, because you can manage things wherever they are without, you know, much difference in how you kind of manage that, whether it’s, you know, IOT or on the intelligence edge or on-prem or in the cloud, so, you know, I think if I was a CIO looking for a way to kind of deal with legacy apps without having the refactor them to make them a little bit more portable, I’d leverage the Ezmeral container platform, wrap it around one of my legacy applications, and put it on top of GreenLake, and make it fairly portable. That sounds like a great solution to me.
Dan Frank: Yeah, absolutely. And it’s something that back in the day we were familiar with on the Unix side when we would do emulators and you would emulate an old application from the Unix space and put it on new hardware, but it would think that it was running on old hardware, so HPE is truly it was a research and development company to begin with, so they got a long history in this, so this marriage between the Ezmeral container platform and HPE I think is a really good fit. They’ve got the technologists behind this to really leverage this in the marketplace, so it is a pretty cool taking some kind of old technology in terms of an old theory and applying it to, you know, cutting edge, cloud-based systems today, so that’s pretty cool.
Mitch Prust: Dan, let’s go back to something that you talked about a little earlier, you know, this kind of battle, or maybe it’s not a battle, but the differences in how customers might want to leverage CAPEX versus OPEX.
Dan Frank: Yeah, that’s a great question. So times are changing, financial regulations are changing, organizations are viewing their capital differently—how do I best leverage and invest my dollars today for return tomorrow? That’s not anything new, but, traditionally, in the world of infrastructure, there was tremendous capital outlays for an asset that would last three, four, five years, and you would go through your depreciation cycle, but it was a tremendous amount of capital to layout and year one for these projects. As cloud-based, traditional cloud-based hyperscalers, began to become enterprise ready, and they were offering services that organizations viewed as beneficial, they could—they realized pretty quickly of, hey, I’m only paying on a monthly basis for this service and I’m able to free up capital to go invest in whatever it is, additional inventory or real estate or supplies, for example, that I believe that I can get a better return on, so organizations started to view that beneficial. Now, when you look at these consumption-based business models, whether it’s HPE GreenLake or others, organizations are starting to view their infrastructure the same way, can I leverage my capital differently, and maybe invest it better than putting it into infrastructure, so it moves on your books. I’m not—you can look at my report cards from college, I’m not great at accounting, but, you know, when you do a capital purchase, it goes on your balance sheet. It’s an asset and you depreciate it and amortize it over a number of years, and that’s becoming shorter and shorter, but the lifespan of some of the products are actually getting to be longer. If you look at storage with their flash drives or solid state drives, they’re starting to last five, six, seven years. There’s no longer moving parts in them. So how can I potentially look to keep those devices perhaps a little bit longer? That’s one thing that to look at. The other one is it’s kind of like when you look at the electric bill that you get, I don’t—I need to—I have a business, I didn’t buy a power plant to power up my office, I paid the utility each month for the amount of gigawatts that I use, and that’s the way it works. So why would I buy all of this compute and storage capacity and grow into it, you know, pay up front for it and then grow into it over a period of time? It makes a whole lot more sense if I can make it work financially, and if it looks—if it works for me as an organization, financially, to just pay as I grow. So some organizations are growing at tremendous rates that they’re—they will outgrow their arrays or they will outgrow their servers. We have one customer that was every 18 months, because of this very rigid electronic medical record application, was outstripping the performance capable of their largest server. 18 months. So if you look at it as a company is like, like I bought this asset, I thought I was going to have to rise this over, depreciate it over five years, but it’s already used its life cycle in 18 months. That’s probably not the greatest use of my funds, so now you’ll ask yourself, well, why wouldn’t they just lease it? Well, there’s a little bit of a difference between consumption-based business models, we talked about GreenLake earlier and some of these sorts of programs, and how it difference differs from a lease, and that a lease is for a fixed period of time, and a fixed monthly rate. A consumption-based business model, for example, again, GreenLake, you pay for—you have a variable cost per month based on what you actually consume. There is a fixed timeline, right? You establish whether they want a three, four, five year, initial timeline, and then you can add capacity as you go via change orders, but your payment will be different each month. It’s based on exactly what it is that you used. We didn’t go into a real—just a very basic understanding of how most of these programs work. Essentially you work with the organization, understand what your needs are, what you think your needs will be today, and then what you think there’ll be three, four, five years down the road, based upon your growth, whether it’s growth in compute capacity, you know, number employees you’re going to add, storage capacity, for example, and that becomes kind of your end—kind of where you think you will wind up, and then they’ll work on providing you a—I kind of think of it as bumpers between the lanes or gutters on either side of the bowling alley, so you’ve got a minimum capacity that you’d utilize during those—that timeframe three, four, five years, and that’s going to be your minimum payment, but then they also have a buffer that you can flex into as you need it. Some organizations are very seasonal, so they’ll be very, very busy during the summer months, and then tail back off, and they can dial things back down, so not only can you flex up into additional compute storage capacity requirements, you can also dial it back down, so again, you’d only pay for what you need. A traditional lease isn’t like that, it’s like, I’m leasing this hardware for a period of time, my payment is X number of dollars per month, and it’s flat, and that’s where the way it works. Really think of this—these are—this is as a service, I am not acquiring the hardware.
The example I like to use is Cintas is an organization that provides carpets—if and when you walk into an office building and there’s a little rug on the floor, they also do like lab coats for hospitals or coveralls for mechanics, for example, or manufacturing. The company acquires their service, they don’t acquire the carpets, they don’t acquire the coveralls or the lab coats, they pay X number of dollars per month for that service, so every time you walk in on Tuesday and the Cintas guys they’re rolling up the old carpets and rolling out the new carpets, the owner of the company doesn’t think, “I own those carpets.” He says, “Oh great. I got fresh new carpets. That’s part of my service.” And that’s kind of the basis of something like an HPE GreenLake is it is a service, it isn’t a lease, and it isn’t a you don’t own the hardware.
Mitch Prust: Let’s kind of talk about and focus on the role of the director of IT, and maybe why they might be interested in something like GreenLake. So, really, from what I’ve experienced is, you know, that IT director might need to be finding a home for a workload, right? So they can, you know, they could—let’s say they don’t have GreenLake at first, but they could kind of figure out what that workload’s going to be, maybe it matches up with one of the reference architectures that they have. They kind of choose the service for the workload, they can kind of do this online, right? They don’t necessarily have to do anything special. They can kind of order, procure the hardware. If they don’t have it, it would arrive. HPE or OST is it’s service delivery partner could install that equipment. They could let the system kind of build, boots up kind of HPE GreenLake Central, and they can kind of point and click their way through this thing, all while leveraging like OST’s managed services for, you know, taking care of a lot of the things that are kind of leveraged on top of this, so it sounds like that’s a great workload or work workflow for the director of IT.
Dan Frank: Yeah, absolutely.
Mitch Prust: So Dan, as you position GreenLake to a customer, I mean, you’re kind of the customer’s advisor, right? So you’re helping to keep all kinds of manufacturers, honest and, you know, really have shared expectations with them. So what would you, you know, what sort of conversations would you have with that customer prior to them pulling the trigger on something like GreenLake?
Dan Frank: We usually start off, I mean, very fundamentally, is making sure the solution that’s being provided technically meets the requirements of the customer. That’s the very first foundation. So it isn’t financial discussion, it isn’t a whether it’s a service discussion or an infrastructure discussion is technologically technically will this solution that we’re architecting meet the needs of the customer and the business? That’s step number one. Step number two, when you walk into or start exploring a consumption-based business models are, there are differences between manufacturers, so it’s important to really understand prior to getting too far down the road, you know, what are some of the requirements, what are some of the minimums, what are some of the factors that might require a change order, so do you think of things like length of term. Some organizations can do them as short as 12 months, others have a minimum term of three months, and just understand, you know, what does that mean and what happens at the end of the term? So for example, on HPE GreenLake, I’m working on one right now, it’s a five-year term. At the end of that term, the organization can opt to continue on with GreenLake and just do a tech refresh at the end of that, and start a new contract, or they could actually purchase the hardware as well. HPE will make them an offer. I’m not sure if it’s a dollar buyout or fair market value, how that—exactly how that works, I think that actually can be discussed prior to the start of it.
The other thing that’s important to understand is—I’m sorry, there’s a train in the background, hopefully that’s not picking up on the mic here, but that’s one of the beauties of working in Downtown—are the—what happens when, or if you move outside of your—the flex capacity that you’ve negotiated, say, hey, we acquired another company, and now, rather than just having an additional hundred terabytes on site, now we’re going to really, you know, grow. If I have to issue a change order, does that mean I have to extend the length of my contract or can I just increase the amount of my monthly payment, again, based on what I pay, but keep co-termed into—and I had an originally a five-year contract, can I do that? That’s important to understand going into it with eyes wide open, right?
The other thing to understand is what are the initial payments like? Is there anything else that I need to pay up front? I’ve heard a couple of horror stories of organizations that have gotten into the ninth inning, they’re ready to adopt as a service model. This wasn’t an HPE example, I’ll leave the manufacturer unnamed, but pretty much at the last minute they came back and said, “Oh, by the way, we need this large chunk of money before we start this flex capacity sort of engagement,” and that caught them really off guard. So it’s important going into it early on, what will the payments look like, when are payments due, and then how am I measured and monitored? HPE has got some fantastic applications that do capacity planning and monthly report outs that you can see at any time, how you’re using, whether it’s your computer, your storage, that gives you just a complete view, a 360 degree view of your environment, and then also future planning as well, you know, based on your current consumption, it appears that you’ll use X amount in years to come. That’s also really great from a budget planning perspective, a workload perspective, and then if you do show backs or charge backs to your organizations, you can give them that sort of information pretty easily.
So those are some of the things just to keep in mind as you enter into this as a service, and then, you know, how easy is it, what are my options to cancel, can I get out of this contract if I need to, what are the implications? Again, I get bought or I have to change geographies, et cetera, if I need to move my data center from Michigan to Canada, can I do that? How does that work? Those are all some things to ask if you’re entering into a model like this.
Mitch Prust: You touched on, you know, a lot of things that would really intrigue a CFO, right? So they kind of want to know how much they’re going to be spending. IT operations wants to know when they’re going to need more capacity, and, you know, GreenLake does that, but also, you know, we touched on Ezmeral. How, I mean, the developers will love that, too, right? So all of a sudden they have a way to kind of get their legacy applications off their hand by kind of putting the container around the legacy applications. The data scientists within the company, they’re going to be intrigued by reference architectures to support their machine learning areas. The CIO probably wants more things as a service, so that they can focus on other things, right? And the managed services from an OST along with GreenLake can provide that.
Dan Frank: Absolutely. And I think from an organization’s perspective, as you’re evaluating, whether to look into the cloud, whether to keep things on prem, whether to have a hybrid between the two, the consumption-based business models give them that option. HPE gives them an option for GreenLake and some other manufacturers as well, but this is providing agility, the flexibility to respond to, you know, today’s needs quickly in a way that your business customers are really are demanding, so there’s a lot more flexibility. There’s a lot more questions and answers to be had, but I think, again, if you start with how technically can I solve my problems first, and then we can look at the flexibility for how to acquire the solutions that go with it, we can help you with that as well.
Mitch Prust: Dan, thanks for your time. This has been a great conversation about consumption, GreenLake, Ezmeral, OST services that we can provide, so thank you.
Dan Frank: Absolutely. Thank you very much, Mitch. I appreciate it.
Lizzie Williams: OST, changing how the world connects together. For more information, go to ostusa.com/podcast.