 Live from Miami Beach, Florida, extracting the signal from the noise, it's theCUBE, covering.next conference, brought to you by Nutanix. Welcome back to steamy Miami, everybody. Stu and I are really pleased to have Dan Gorman on. He's an enterprise infrastructure architect, IT expert. Dan, great to see you. Thanks for coming to theCUBE. Thank you for having me. Your brother is a CUBE alum as well, so that's awesome. Yes. First of all, your thoughts on this conference. Next, we were talking offline. You go to several of these, you know, a year, which is, you know, I think you said five a year, which is a lot for an IT practitioner. So where do you put next? Where do you think about the Nutanix event? Yeah, so first off, I think that this event is absolutely extraordinary. I've never seen this amount of attendees in a first conference in my career. That should tell you a little about the infrastructure that Nutanix has and the software that they're doing. I think it's absolutely revolutionary in the market. I think they're challenging a lot of the norms, and I think they're bringing an approach to the industry that is really transformational. So you know, obviously, VMware very well. You're experienced expert in that space. And VMware, phenomenal, the enormity of that trend. And, but essentially, we were talking earlier, Stu and I, about, but VMware essentially did is take a massive infrastructure that was just incredibly inefficient and make it more efficient. And relatively less complicated, but still very complicated. Correct. What is, where does Nutanix fit in terms of relative to that sort of state of infrastructure today? What do you see as the Nutanix offering? So I think VMware has done a lot of great things and a lot of great products. Bringing the hypervisor to the market was truly transformational. I think the industry has moved past that and that's sort of the norm, the commoditization of that. And starting to move the infrastructure sort of an abstraction model, where you can move the apps in and out of environments. You can move your infrastructure in and out of hypervisors. But VMware really led that way, and now we're ready to do the sort of the second coming of infrastructure. So I get to ask you then, is it from a practitioner's perspective? VMware will talk about making infrastructure invisible. They'll talk about simplifying everything, orchestration, management, they'll talk about, they'll even talk about multi-clouds and talk about OpenStack and Bracen, Docker, et cetera. So what's different? So Nutanix is bringing web scale and invisible infrastructure to the enterprise. Historically it was very brittle to manage, very brittle to use. You had to have SMEs in those areas. You had to build sort of the business around that when the business might not actually be in the business of doing that. Nutanix is bringing the simple point and click enterprise scale out, a web scale methodology that's been around forever, and they're the pioneers of that. Okay, so they're early, but won't the big oligopoly and the monsters, the whales of the industry just sort of do their own version of that? Will they catch up? Or are they doing that on infrastructure and software that is too difficult to transform? So I think each manufacturer is going to want to make sure that people are in their own ecosystem and that makes them somewhat biased. Nutanix is allowing you to say, hey, I don't care what ecosystem you're in, we're bringing that web scale. You can live within whatever ecosystem you'd like. And competitors will have similar platforms. Evo Rail is one of them. But you're locked into the VMware V-Tax or VCN-6 and now you're in that swim lane. And some customers may not want to have that and have the choice to be able to move around. Well, Vinod Kosla this morning was saying, hey, most customers are risk averse, so they'll keep buying what doesn't get them fired. What do you mean by the V-Tax? Can you explain that to our audience? So the enterprises generally in most cases have a five act charge back model essentially resell that the infrastructure back to their internal units. And all those costs, support costs from the hypervisor to the support of that, to the infrastructure, that all adds into that. And most enterprises will want to collapse that as much as possible to stay competitive with other public cloud offerings. It's important that in some cases where requirements require the users to stay on-prem, that that is a cost point that's attractive. I remember going back to the first year we did at VMworld was in 2010 and cloud was a big topic at the time. Everybody was like, what's cloud? Let's try to define cloud. Amazon was doing its thing and all the enterprise guys are saying, oh, we're cloud too. And I remember the time talking to IT practitioners like yourself saying, how are you going to keep pace with Amazon? How are you going to fund that? How are you going to afford to? And the answer was, well, we're not going to do it alone. We're going to look for partners to do it. And Nutanix is a company that's executing on that. Correct. Is that the right way to think about it? Are you essentially building as an IT practitioner today? Think about your colleagues. Are you guys building essentially an internal cloud that is Amazon like? Correct. And the core components that is having a partner that will work with us as an org where we can say we want to build a cloud that is scalable, it's elastic. It will ebb and flow as our business needs change. And Nutanix has been a perfect partner for that. So, in thinking about that sort of theme, do you feel like when you see you and your colleagues talking to your business constituents, do you get pressure to do the public cloud? Whether it's Amazon, Azure, Google, you get pressure to do public cloud and are you able to credibly say, look, we can deliver the same types of services? Or do you say, you know what? For certain workloads, you should go to the public cloud. For others, here's why you should stay here. Can you help us sort of parse that? Yeah, so I think it's a little of both, honestly. The biggest challenge is sort of the crud environment, the create, reuse, update, delete, where the developers want to move very quickly. Historically, operations has a longer pull in terms of being able to deploy that infrastructure. Building on a Nutanix infrastructure with something like orchestration automation, allows you to sort of catch up to that pace of development that the developers would like. In some cases, public cloud makes great sense. In some case, you have to have it on-prem for certain business requirements. But you must be able to meet that agility that public cloud can provide with all the enterprise security and components around that. So Dan, I'm curious. When you hear Nutanix talk about just making infrastructure invisible, doesn't that obviate the need for an architect? Well, hopefully not. But you still have to build the environment such that it can meet the capacity needs. Eventually, you want to be able to working on higher order tasks. So an architect, traditionally, shouldn't be saying you need this amount of spindles, you need this amount of CPU, compute. We want to be working on the business challenges. Things like we have a feature or product X that needs to be deployed. Here's the infrastructure, we need to do that. It can scale as little or as much as we need. So I think architect role changes a little bit. Yeah, I'd love to get your viewpoint because one of the big advantages we see to this new style of architect, Nutanix, kind of all the Kuipert converge is before, you say, okay, this year, go buy some servers. Next year, maybe it's time for storage. And predicting those individual silos was impossible. And when you need to buy more that migration upgrades, I mean, the whole thing is what I called, it was like the hamster wheel of pain. Every step along the way, there's just pain and everything like that. Does hyperconverge really make it easier, Nutanix and these technologies specifically, to expand and are there enough options out there that you can kind of build this for any application or is there still work to do to mature that? So absolutely, and I think the biggest challenges enterprise face from my perspective today is those big cap expense, right? You come in the beginning of the year, throw down a couple of million dollars, you try and anticipate what the year's growth is going to be. You may or may not hit that. The beauty with the Nutanix model is you can start with a minimum viable product. You say, I need a three or five node cluster. You might be fine, you might not, but if you have to expand, you can do it in chunks that are bite-sized. You're not doing these large capex outlays every time. So I think the model from the enterprise web scale has changed that so that people's budgeting processes even change where we can say, hey, we can stop doing these big bang approaches and start scaling out our compute and storage in ways that are more appropriate for our business or product needs. All right, so you throw out the word scale. I'm curious, what's the largest cluster you've deployed with Nutanix? So we've deployed probably about a eight to 16 node cluster, roughly 100 nodes more or less. So we have a lot of experience around the Nutanix infrastructure. Yeah, there's some people that say, oh no, nobody goes more than eight nodes and so you're here saying that yes, you've done it, works fine, not limitations, anything like that. Correct, and you start to get in the classical debate of separate clusters versus one homogenous cluster. It's classic debate. What's the trade-off, help take us through that debate? So it's interesting because when you start putting everything in one bucket, you start diluting cash, you start creating contention, you start getting the most unoptimal usage workload, other environments can affect workloads and you start losing control of that. On the flip side, if you segregate everything out, yes, you have 100% guarantee of that IO or CPO compute, but now you're at a cost point that's cost prohibitive. So typically what I try to do from an enterprise perspective is weigh those scale up, scale out cluster approach and say, what makes sense for the specific BUs? Are we talking a financial application? Are we talking a logging application? Put all those types of applications on the single cluster. Don't mix a corporate resource with a logging resource, right? So lots of those trade-offs and we work on those every day. Okay, so does Nutanix need more quality of service to be able to expand that into a larger pool or is there something architecturally limiting it or are we always going to have these trade-offs in your opinion? I think you're always going to have these trade-offs. Nutanix does a great job in terms of tiering and IO prioritization. You can use things like VMware's priority groups to do IO limitations, but typically we let the apps try and determine themselves and then make sure that the apps are coalesced in a way that makes sense per cluster. So I wonder if we could come back to this notion of when you're as an IT enterprise architect, how you go about things these days. You got, we were in the keynotes this morning, right? You heard the Gartner guy talking about mode one, mode two, IDCs, platform two, platform three, kind of superficial concepts, but they're good ones because they sort of describe what's happening. It's bimodal notion. So what do you do as an architect? You've got 90% of your revenue in most companies comes from the mode one. You know, everybody knows the grow the business. I'm running the business. That's where all my resources go. I want to grow the business, but I can't fund it. So how do you approach that? How do you architect for the future? So I think the biggest thing that we're trying to do these days is show the cost of IT and the simplification of that. Historically, most enterprises have approached via bolt-on. You'll see a lot of the software companies coming out trying to fix problems that are band-aids on more core issues. So the way we approach that is figure out what we're offsetting, figuring out what infrastructure we're removing, figure out how we're actually saving the company money even though there's a spend there. That's your question. If the CEO comes to you, says Dan, I want to completely transform the company. You know, we know we have to compete. We have all these other forces. We want to digitize our business. We got Uber coming after us, right? I want to transform the company. And I don't want to take the band off slowly. I want to go hard, I want to go fast. We're going to make IT the main spring of that transformation. How much money do you need? If he gave you all the money you needed, could you actually transform the company more rapidly? Is budget a constraint to that transformation? Or is it just inertia of existing processes and skill sets? So I think there's a couple of things in that. Budget's always an issue, right? That's the reality of the world. But if somebody approached me and said, hey, we need to turn this thing on its head. We just can't keep executing IT the way we've been doing it for the last X years. What do we do? Virtualize everything. There are some exceptions, but virtualization first approach. Work with the developers, work with the manufacturers on how do you work with that scale out? Oracle's a great example. SQL Server's a great example. Traditional IT is a scale up approach. It requires operations and architects to work more closely with those product groups. It's absolutely doable. And I think there's a lot of politics and perhaps animosity from changing from what is known to something that's unknown and relatively new. And I think breaking through those emotional barriers generally is most of the fight in trying to get people to understand that. So VMware is obviously the dominant virtualization platform. There are clearly others, Microsoft, Hyper-V gaining, Momentum, KBM, pretty popular. You mentioned Oracle. Oracle early on, would fight tooth and nail, customers wanting to virtualize. Wikibon, we told customers, damn, the torpedo's virtualized. Oracle, what's your experience been with virtualizing Oracle? So outside of OVM, I should say, but happy to take OVM and know that. So it really comes down to working with the DVAs and understanding how to virtualize the database. Traditionally people come out and say, hey, we have a single database, we have a single data file, we have a single way of doing this. And understanding what the data actually is in the database. Can you shard it? Can we start doing partitioning? What does that look like? Coming at it from a smaller node, but more scale out. But that causes DVAs to have to rethink how they're doing data management. And that's scary because you're saying, hey, you've been doing data management this way. We're gonna completely redo that and start all over again. And most people aren't wanting to do that work, right wrong or indifferent. So it can be scary and somewhat threatening. So Dan, Dave did an interview about a year or so ago with CIO of Royal Philips. And he said, I want to consume all of my IT by the drink. It doesn't mean I'm pushing it all to public cloud, but I want it all to be elastic. I don't want to do that huge capex outlay, as they say. And they're doing some interesting to do application, they do Oracle and SAP. What's your take on that? And I guess that's specifically, and also how does public cloud fit into your architecture? So the pay by the drink model is interesting in that when companies are doing charge back, how do you actually capture those costs? In my past, I've worked for public cloud companies and that's a core business. If you don't do it well, the business fails. So really understanding the pay by the drink model as it relates to charging back to the business and what do those clusters look like? From the second component is the management of that. How do you enable the business to do automation and orchestration of the pay by the drink model? In some cases, on the flip side, you can say, hey, business unit, I'm just going to give you your own cluster, allow you to deploy your own VMs and infrastructure. We're just going to make sure the infrastructure's running and healthy. So I think there's trade-offs depending on the business and what the requirements are. Yeah, so do you see IT pulling in that public cloud discussion or the business unit still kind of doing a little bit of the rogue stealth IT mode? I think I see both, right? So there's a case for public cloud, for example, ADP, right? We're never going to do, I would never recommend a company doing payroll in-house. However, there are certain applications that there's their core competency, which should be in-house. And so I think there's actually a mix of both in sort of the new world. So can we talk about data protection? When the world moved to server virtualization, everybody realized storage doesn't like this, right? It broke storage and it broke backup. So a lot of customers re-architected their backup. Does hyper-converged or whatever you want to call it have a similar effect on the way in which you protect data or was that more of a virtualization event? So I think when you talk data protection in a hyper-converged world, you're actually up-leveling the conversation in that historically, when you talk data protection, you're talking about replicating a storage frame in some way and the quiesce in the app. And in the Nutanix world, you're actually talking about the OS or the app. You're talking about application consistency. You're not thinking about the underlying fundamentals. And I think Nutanix has taken that with say, hey, I want to protect this VM. We're not talking about alums. We're not talking about volume groups and it up-levels the conversation that's more relevant to the business. Okay, so that sounds like a yes. You would look at, so is your advice to fellow practitioners to rethink backup and well, obviously you're a fan of the whole hyper-converged thing generally in Nutanix specifically, but is part of the domino effect of rethinking backup? And I think not just backup, but a lot of the way architecture and infrastructure is done and backup is a component of that. And I think that can be scary for enterprises because they have to rethink backups. They have to rethink how they're doing deployments, how they're doing clusters, how they're doing a scale out. And it can be a little bit scary. And I think most enterprises are wanting to understand more before they do it. So carry that through a little bit more. What, I'm going to push the advice button. What would you advise in terms of all those factors from an enterprise architect standpoint you're coming in, Dan the consultant, we're going to pay you a bunch of dough to just tell us what we should be doing. Where do you start? Understand what your requirements are, understand what your assumptions are, understand what your business is. Moving beyond that, what is the ability to virtualize? Can you virtualize your infrastructure? If so, what does your growth rates look like? If you're flat, how does that affect your infrastructure? So really understand the business before you get down into the bits and bytes of it and build the requirements and assumptions out of that. All right, so Dan, you've mentioned virtualization a lot. I'm curious what your take is on the whole containerization discussion. Docker. Yeah, Docker, CoreOS is also a partner in that space. So where are you with containers? Where do you see that fitting? Bare metal, virtualization, cloud, everything. So I think Docker and containerization has been fantastic in the industry. From the perspective of a developer being able to do a crowd environment very fast without having to build an entire heavy VM out of that has been hugely transformational. I think there's some concerns around the management, the deployment, the security, just like OpenStack, right? It's getting there, a big fan of OpenStack, but they still have a lot of hurdles to get through. Things like PCI come to mind, right? When you've got compliance and regulatory, whether TEPA or other, how does that work in? And so I think Docker and containers are sort of their infancy where compliance regulations may be a little more lax, where they're more heavy. I think it requires enterprises to do a little bit more homework and take it more. I mean, today we're mostly stateless apps, so getting some of the heavy-duty GRC compliance, PCI, everything, so a little bit early, so. Agreed. According to Mark Hurd last year at Oracle OpenWorld, might have been two years ago. The shows are flying by here in theCUBE. But he said the average age of the enterprise app is 19 years. Wow. Now, sounds about right. That could be one of those numbers, like 70% of our spend goes to keeping the lights on, who knows, but it feels about right. That's not a surprising statistic. And maybe that needle has moved a little bit, but the average enterprise app is pretty old. Yeah. Claims apps, payroll apps, et cetera. What do you see the world looking like seven years down the road in terms of IT infrastructure? How would you describe it? I think you're going to see a huge shift where sort of the SMEs fade off into the distance. You're going to have a generic platform that's more focused on services and service availability than infrastructure. Hardware's going to come in and out, but at the end of the day, IT's going to be more worried about service delivery. And you're starting to see some of that today with even public cloud email offerings and whatnot. All right, so Dan, I want to take out your magic wand. You talked about a lot of things you like. What still is difficult in your job that you really need the vendors out there to make invisible, make it simple, that would make your life and your peers' life easier? That's a good question. So I think a lot of the adoption is based around other people's experiences. So having other people who are running Oracle Rack and virtualization on Nutanix or other hardware, being able to get that broad industry adoption, I think helps shift those opinions and conversations. Hi, Tim. Well, listen, we have to leave it there. Thanks very much for coming on theCUBE. Great insights from enterprise architect and IT expert. Stu and I are very pleased to have you on. So thanks again. All right, keep it right there, everybody. We'll be back from Miami. This is theCUBE. We're live at dot next, hashtag next conf. Got it? Is that right, Stu? You got it, Dave. Finally, after four hours of broadcasting, I get the conference name correct. Keep right there, buddy. We'll be right back.