?

TEC37 E07: Why SDDC May Be Right for Your Organization

42:05
95
Plays

Software-Defined Data Centers (SDDC) are designed to overcome the drawbacks of traditional IT infrastructure to offer the accelerated delivery of consistent data center services. Despite the performance gains, some organizations have been hesitant to adopt a SDDC environment. Learn about the pros and cons of SDDC and whether this approach may fit your organization.

Please view Transcript below: 

 

Robb Boyd:                   Welcome to the TEC37 Podcast, your source for technology, education, and collaboration from World Wide Technology. Today's show is sponsored by Dell Technologies and Intel. My name is Robb Boyd. Today's topic is a big one, SDDC, that's software-defined data center. It's not a product or technology you can buy, but it is, I would argue, where you should be moving your organization. See, a well-implemented solution combines a software-defined infrastructure with automation and orchestration to deliver services predictably and reliably. And that one statement may sound like the Holy Grail because of course, this is what everyone strives for.

                                    So then why are so many of you hesitant to move this direction? I hear you, it's easy to talk about these things, it's completely different to actually make big changes within a production network. So, today, we're going to ask our experts to explain how this might actually work, and make sure that the challenges are acknowledged along with the benefits. This is a great conversation, once again with a group of really smart people to guide us. 

                                    All right guys, we obviously have a big panel today, appropriate for the subject. So, let's run through some introductions here real quick though. Right next to me, I should say, we've got James Harless from World Wide Technology, solutions architect, thank you so much for joining us. Also, Sean Hicks from World Wide Technology on the show with me again, that's becoming a pattern, but I like it. Shanon Berry, one of our sponsors from Dell Technologies, appreciate you taking the time to join us. And of course, Robert Looney from Intel. Robert, glad to have you. All right guys, so here's the question, we're talking SDDC, so let's start with the basics, and I'll start with you James, what is SDDC, how do you define it?

James Harless:              Yeah so, software-defined data center is kind of an aspirational principle as opposed to a strict definition, it's really not something you can buy, but in principle what it is, is a virtualization of the compute, network, and storage layers, and then building a data center model around that. So, we're moving to a model that is less focused on buying hardware specifically for projects and then having those assets tied into those particular projects, and more to a model that can be consumed on-demand. 

Robb Boyd:                   Okay. Well, let's stay on this definition for a moment, because one of the things I always like to arrive at is what something is not, but in general what I'm also hearing is this is somewhat aspirational in terms of somewhere to go, but I don't know if that applies to everybody in our audience equally. Who can weigh in on software-defined data center and it's appropriate as a goal? How appropriate is it for ... maybe certain industries, or is it something everybody should be going for?

Sean Hicks:                   Well, Robb, you see, software-defined data center as kind of an operating model, if you will, is a distraction above proprietary hardware, and the islands of skillsets that have developed in the data center over time. It's really a model that can exist in various formats. We see a lot of popularity with cloud, with public cloud in particular, and what companies may or may not realize is that when they're adopting things like AWS, and they're adopting things like Azure, and Google Cloud, they are adopting a software-defined data center model, they're adopting that cloud provider's SDDC. 

Robb Boyd:                   Okay, okay, well that brings up an interesting angle on it-

Robert Looney:             [crosstalk]-

Robb Boyd:                   I'm sorry, go ahead, who is that?

Robert Looney:             Yeah, it's Robert. Yeah, from that perspective it's very much in keeping with that, what are the aspirational goals of moving to a software-defined data center or software-defined enterprise is, I know I need compute, I know I have storage resources, I know I have network limitations and needs, I even have security needs. How do I express those all where they're not isolated from each other, they're not necessarily managed as independent stacks all the time as they kind of grew up, how do I allow the kind of closer to unfettered resource of bringing those, constituting those together for an application or a VM, through software. It's not somebody running through the data center pulling wires, right?

Robb Boyd:                   Yeah. 

Robert Looney:             It's clicks and policies driving the utilization of the resources. We often describe software-defined as pools of resources that are put together and somehow orchestratable to serve the application, and at the enterprise level, at full data center level.

Robb Boyd:                   Let me pose this question to Shanon actually. Sorry, I'm going to just jump in right here, because Shanon you can reflect on anything you want, but the starting question is just I want to get a solid feel for what does the data center look like that's not software enabled? I want to make sure we understand how to recognize where we are now in most situations that need maybe a change.

Shanon Berry:               Yeah, no, it's a great question for a compare and contrast standpoint. A traditional data center has historically been made up of three tiers of siloed infrastructure. You've got your compute layer, you've got your networking layer, and you've got your storage layer, and those are usually, or historically have been completely separate components, or elements, or pieces of hardware that have different teams that are responsible inside the organization for managing and orchestrating those, and that's led to a lot of inefficiencies, which is one of the reasons why public cloud has been so quickly adopted, because like the rest of the panel has indicated, you're adopting a software-defined data center when you go to public cloud whether you want to or not. All of those different elements that are the hardware pieces of it really become irrelevant and companies are looking for an outcome. Software-defined data center delivers those types of outcomes on the customer's premise with incredible levels of automation, and also some pretty compelling financial considerations as well.

Robb Boyd:                   But you're not saying that all we got to do to get to SDDC is simply move everything to a public cloud provider?

Shanon Berry:               No, I would never ... I certainly wouldn't say that. There are limitations, I mean, application [crosstalk]-

Robb Boyd:                   Yeah. 

Shanon Berry:               ... cost factors, data gravity. I think that there's clearly a place for public cloud, I mean, it certainly wouldn't have been adopted the way that it has been, but there's also a place for software-defined data center. I mean, one of the kind of key tenets that we have in Dell Technologies is that cloud is not a destination or a location, it's an operating model, and that's really the same kind of conversation about SDDC. It's about the operations model, and what the outcomes are that customers are trying to accomplish. It's kind of irrelevant the place that it happens.

Robb Boyd:                   Gotcha.

Sean Hicks:                   Yeah, so SDDC is definitely the foundational layer for building that cloud operating model. Why are the cloud providers so popular? It's because of the cloud operating model, it's because of the popularity of things like elasticity, and agility, and self-service, the ability to do metering, and all those things rely on some underlay of an SDDC, if you will. So, what we're trying to do is for reasons mentioned, various reasons, there's still a number of applications and the data that still is on-prem, and it's not going to go to the cloud anytime soon, so how do we get all of those things that have made public cloud so popular? How do we extract that from our own private cloud environment, and the idea behind that is we need to start with that same kind of idea of an underlayment.

Robb Boyd:                   Okay. Before we dig into that further, I do want to highlight the fact, so this is going to be a two-part series. So, today we're focused really on the business side, which is a lot of the why, and making sure we define this stuff correctly, and hopefully build some momentum towards the fact that obviously I think we all would agree that a software-defined data center is something we want to move towards, but I just want to call it out that there will be a part two if someone feels like, "God, these guys are just really avoiding the, 'How do I do this?' question." 

                                    But so with that understanding, I also want to highlight the fact that ... and you guys taught me this kind of as we were getting ready for this, but this notion of this is not just about technology in the sense that this is really ... there's reasons why organizations have not already moved to something that ostensibly feels like it's better, because it's a big transformational change from an organizational culture I believe, if I'm wording that correctly. 

James Harless:              Sure. 

Robb Boyd:                   What is meant by that? What are some of the things ... in fact, let's just do it this way, what are some of the things that are holding people back from having done this ... why haven't it already been done?

James Harless:              So, I think that a lot of it just has to do with organizations continuing on with their current direction. I think that's really why public cloud came in and was so easy, because you can basically change directions of the IT organization by outsourcing it to Amazon or Microsoft, or whatnot, and really adopting SDDC means in principle challenging some of those things inside of your own organization that have their own momentum. So, for example, currently, a lot of organizations buy hardware, compute hardware, storage hardware, networking hardware per project and they assign that gear to that individual project, and that is the exact opposite of an SDDC where we're guying buy a large pool of hardware, we're going to think of the data center as a unit, we're going to parse it out individually to projects as they come in, but we're going to use virtual constructs, virtual networking segments, virtual security constructs, virtual storage constructs to parse that out to our business units and to our projects. So, that trickles all the way up into finance, how do we [crosstalk]-

Robb Boyd:                   Wow. 

James Harless:              ... how do we fund it, how do we do money moving back and forth between business units, and so those ideas are very difficult.

Robert Looney:             Yeah, James is dead-on correct on that. If growing a part of your business meant I need to see the whole five year projected growth of that business so I can go sit down with finance and figure out how I'm going to amortize the cost of this new SAN purchase over that five years, you never start the business, right?

Robb Boyd:                   Right, yeah. 

Robert Looney:             Versus if I'm using kind of my common hardware infrastructure building blocks, and I could orchestrate your needs as you grow, not having to buy the five year in advance, I can allow the business to grow on top of my software-defined data center infrastructure's capabilities. 

Robb Boyd:                   I got to think, too, that one-

Sean Hicks:                   Another part of the conversation-

Robb Boyd:                   ... benefit that the cloud organizations it feels like has over, say, a traditional business is the fact that they figured out this model from kind of a greenfield type approach whereas all of our customers are already up and running in a model that may not be perfect, but yet they're familiar with it, and it's operating in 24 hours a day and that kind of thing. But I cut somebody off there-

Sean Hicks:                   Glad you put it out there, Robb, no, glad you put that out there because that transitions nicely into what I was about to add to those comments, which is basically we have a problem with value, with how we assess value in our organization-

Robb Boyd:                   Interesting.

Sean Hicks:                   ... we have generations of people who had it beaten into their skull that their value is in the busywork, the day to day stuff that's really not valuable to the organization, you know?

Robb Boyd:                   Yeah.

Sean Hicks:                   Taking somebody who's a highly compensated IT professional, say a network engineer, and having them crawl around with the dust bunnies to run cable, it's not bringing value to the organization. It's busywork, it's something they can point to and say, "I did that.", but that's not what your company actually hired you for, your company hired you to innovate, to do architecture, to make decisions around security. And all of these things that we would be able to do if we could just let go of some of that busywork, and some of that whole idea of my skillset, and kind of guarding my space, this is my fiefdom, tearing down silo walls, and going towards a model where we're all working together towards the organizational goals, having that communicated from the top down so that we're all fully aware of how our jobs best achieve that. And then that whole 80/20 model, trying to use some of our time to innovate and win back time from the old just stay busy all the time model of work to actual innovation, and actually doing new and interesting things for the organization.

Robb Boyd:                   Well, does an SDDC actually assist with being able to see the value creation happening in that way a bit better, or are you just saying it's an impediment to moving towards it, because there's stuff built in where we have trouble valuing it? But does it actually get easier if you get through there?

Robert Looney:             It's fundamental. Where I would give you ... this is an internal IT example, public knowledge, we embarked upon the SDN, software-defined networking side of our transition about a decade ago, a gig, to 10 gig, to 40 gig, on our way to 100 gig across the network, but adding the software-defined operations on top of that has really changed things to the point where that gets us into the point where we can say, "Listen, we're making our people more productive, because we've torn down some of the boundaries in the software-defined data center construct, we're working more as a team, to the point where IT can express, in our size, $2.8 billion in cost avoidance, right-

Robb Boyd:                   Yeah. 

Robert Looney:             ... versus if we'd had to go get that on the open market from, say, public cloud. We could actually build up our capabilities, build up the infrastructure behind it, build up the practices, so we're just operating in a better environment. We don't have $2.8 billion more to operate IT, but we can show that cost avoidance." [crosstalk]-

Robb Boyd:                   So, the budgets still don't go up no matter what we call it? The budgets aren't going to get-

Robert Looney:             Correct.

Robb Boyd:                   Okay, this isn't that answer.

Shanon Berry:               I think that what everybody's kind of touching on is doing business as usual is a huge obstacle to innovation. When we look at IT organizations spending 70% of their man-hours just doing things that are below the value line, keeping the lights on, doing patches, and reading hardware compatibility lists, I can't tell you how many organizations I've seen where key strategic initiatives or application releases have been delayed and pushed out time and time again simply because IT wasn't able to be able to respond quickly enough. Software-defined data center, through the automation that it introduces, clears the road and allows innovation really to take off and organizations to accomplish what their objectives are.

Robb Boyd:                   Yeah. Hey, I tell you what, guys, here's one thing I want to tackle, and this is based on one place I go to for knowledge, which I encourage everybody watching this to do is wwt.com, which is the platform as I think I've heard it referred to generically behind the scenes, but you can go on there and you can search for more information on these topics as well as some labs and such like that, which we'll talk about in a moment, as well as on the next episode, but in that vein, I came across a couple of really good articles by our own James Harless right here on the call with us. 

                                    James, one of the things that I like that you did was there was an article that you wrote, or helped to write, or something, Six Challenges to Address when Preparing for SDDC. So, if we go under the assumption, and this is for everybody, it's just I'm crediting James with, of course, having published this, but in that, I want to go through each of these and kind of get your feelings, each of you, on this type of thing as needed. But number one was on these six challenges to be aware of, we kind of touched on this, one, breaking the silos. What do we mean by that? Let's make sure we've nailed that one down.

James Harless:              Yeah so, largely in that article what we're talking about when we refer to silos are the silos inside of IT which separate networking, from storage, from compute teams-

Robb Boyd:                   Oh, those silos.

James Harless:              ... and those silos, I think of them as being very vertical, and they're basically structures within IT to support vertical environments, but they don't really overlap very well. So, we don't necessarily make design decisions about our network that would support a better storage infrastructure. So, we're talking about tearing those down, we're talking about supporting a platform that is all of those things as all of our platforms are today, but instead of thinking of them as siloed units, we're going to merge those concepts together, and put teams together to support that foundation, and they're responsible for that platform as opposed to just network availability. 

Robb Boyd:                   Okay, so these are some of the traditional silos we've been breaking down ... Yeah, go ahead, what were you going to say, Shanon?

Shanon Berry:               No, I just was going to say that I think that James brought out really a critical point in that article, when talking about breaking down that, what I call, internal inertia, it really requires string executive leadership, because the people in the trenches that are executing on their day to day tasks-

Robb Boyd:                   Yeah, they're going to be-

Shanon Berry:               ... they need to know that it's okay for them to be able to work together and collaborate and even have some gentle pushing from executives, because they're used to doing what they're used to doing, and they're very comfortable with it, without some sort of a change catalyst, which really is just leadership, it becomes very, very difficult. So, I just thought it was great for him to call that out, because it really is key to beginning able to break down those silos.

Robb Boyd:                   I'm glad you brought that up, because I think ... and that's going to come up again here, I think, in this ... Let's go to number two, and obviously others other than James can comment here, just because he wrote the article, number two is setting appropriate objectives. That kind of blends into what you were saying there, Shanon, but what do we mean by setting appropriate objectives? What would be an example of that perhaps?

Shanon Berry:               Well, I mean, from my perspective I always think about objectives as far as outcomes, what is it that the organization is trying to accomplish? And having clear definition of those really is foundational to what type of software-defined platform you're going to choose, what the software layer is going to be, what the virtualization layer is going to be, what the hardware pieces are that's going to go behind that, because there is no one-trick pony, there's no one solution that's going to be best in every single circumstance. So, it kind of goes back to Stephen Covey, begin with the end in mind, have your objectives clearly defined, because that's going to create the framework for all of the decisions you make as you kind of step through the process.

Robb Boyd:                   Well, it's interesting because you mentioned software and hardware, number three is prioritizing the software platform, and obviously this is called a software-defined data center, so ... well, software doesn't exist without hardware, but at the same time it sounds like we've got a re-prioritization, whereas me from my Cisco background, love the hardware stuff, can't get enough of the hardware stuff, obviously, software's where it's at, applications are where it's at, serving the end users and making sure they get what they need, and we're being flexible with that, but let me ask you this, Sean, prioritizing the ... excuse me, the software platform, what does that mean to you, and what are the pros and cons of doing that appropriately?

Sean Hicks:                   Well, really it means standardizing your platform. So, when we built out these hardware models, and as James pointed out earlier, they were mostly built out of project-based type procurement models where I need to get this particular application rolled out, so I can get some money dedicated to that application, we buy a bunch of stuff, we build it, and then that becomes its own little island of skillsets in our data center. What we're talking about with software-defined data center is moving to a model where resources are pooled, and provisioned on-demand, and even metered so that there can be interdepartmental billing, or even external billing if necessary. By doing that, we're moving to a more, I would say, standardized model of deployment where we could more easily take advantage of things like APIs, and automation, achieve quicker time to market, achieve less fat-fingering, less human error. 

                                    I mean, how much time do we spend troubleshooting because somebody made a change to one firewall rule in one part of the network somewhere out there, and then how long does it take to roll that change back out, and go back to operational? That's the old way of doing things, and through this layer of abstraction what we're trying to do is move towards that model of automation and agility. But automation loves standards, it loves things to look the same, feel the same, and that's not what happens when we have a hardware focused approach to building out our data center.

Robert Looney:             Yeah, and the hardware guy here, we can look at it like, "Well, there are standards for this.", right?

Robb Boyd:                   Yeah. 

Robert Looney:             The software-defined data center essentially runs on x86, you abstract it by virtualizing its resources, its compute, its storage, its network, and even its security, but you have an operating model, and ... when my kids built their first PCs they learned what the components are, and what the operating system does to that PC, to hand out the resources to the appropriate applications. Software-defined data center is somewhat a similar type of a construct it's just your PC's the whole data center, and all of those resources still have to be called upon, you know?

Robb Boyd:                   Yeah.

Robert Looney:             If you want security across all things, fine, make it a capability in software that it's expressed across the compute, the storage, and the network in that system, the software-defined data center. I think the next level where we're headed is extending the software-defined data center, the enterprise, also connecting to clouds in the future [crosstalk]-

Robb Boyd:                   Yeah, because it should be seamless, that's the dream, yeah. It shouldn't matter where the workload's located, it should be simple to put it where the resource is needed, and where it's most cost-effective. To some of what Sean is saying as well, and I don't want to take it too far, but Sean, you're not talking, or maybe you are, because I can't remember if this was ever resolved, but it seems like we went back and forth in the industry for years talking about IT becoming, not an expense department, but a revenue generator. 

                                    And part of doing that is IT's forever complaining, for good reason, that we don't get the investment, and then it gets hard to say, "Well, your investment here does this.", and you could do it on a project basis if you're allocating physical resources to something, but when you're talking about the entire infrastructure being able to be carved up and allocated appropriately, boy, you solve that, if feels like you're now in the executive seat you're at the table able to make decisions for the business, right?

Sean Hicks:                   Yeah. No, we've heard for the last several decades people saying things like, "IT is a necessary evil.", that it's a money pit, or all these various things that have been said over time that have led to this, I would say in the last few years the thing that we hear over, and over, and over again from customers, if data center is not their core business is, "I want out of the data center business. I don't care about racks, I don't care about power and cooling, I don't care about cabling, what I care about is my applications, and getting service to my customers, whether they're internal or external, as quickly as possible." 

                                    What IT has struggled with is, one, doing that in a time-efficient manner, and two, being able to prove back to the business that they do have value, and they are providing value to the business, I would say there's a break in communication somewhere where over the years IT professionals have really landed at a point where I would say 90% of the people I talk to in this industry don't actually know how their job contributes to the company's bottom line, or strategic initiatives, or objectives, and that's a break in communication. We were talking earlier about how it starts at the top with executive sponsorship, and that commitment to doing things in a new direction, and part of that is getting everybody on board-

Robb Boyd:                   That's a great one.

Sean Hicks:                   ... with the idea that IT is yet another part of this business, that they're our business partner, not a necessary evil, a business partner in what we're trying to achieve, and they're here to enable us, and they can prove to us the value that they're bringing to the conversation.

James Harless:              To the point that Sean made earlier, one thing that has to happen in order for that to happen is we have to get some of our smartest guys away from crawling around the data center dragging cables, we need those guys to liaison between the IT department, between the business directly, to help understand what services need to be provided. 

Robb Boyd:                   Now, that's interesting, because-

Sean Hicks:                   [crosstalk] focusing on software instead of hardware.

Robb Boyd:                   Yeah, and you're talking about highest and best use, because I think that's something we want to encourage an organization, but I love the fact you're bringing up stuff that, hopefully, as an individual, it reminds me to always go back, we all need to understand what value we add to whatever we're involved in. On one hand, it's a leadership issue above us, but at the same time, we need to take it upon ourselves to make sure we can define that, because to me, that's the route to, what I call, not job security, but employment security, because if you know where you can add value you're going to be valuable everywhere, because you can go add that value if it's not appreciated where you're at, but that's a different conversation. Sorry, guys. 

                                    Number four, because it bleeds into this a little bit is understanding the cost savings associated with SDDC. So, we talked a little bit about revenue models, and the ability to kind of see as we divide this up when we're in a more cloud-oriented model, what kind of cost savings are possible here when you're kind of on the other side of making this kind of jump?

James Harless:              So, it is a very difficult question to answer for customers, sometimes we get the question, "Hey, can you sum up how much money I'm going to save by moving to an SDDC?", and that's very difficult, because some of that involves understanding what money you're wasting today. How much money is wasted while your business is waiting for IT to move [crosstalk]-

Robb Boyd:                   Just ask them for that list, "Where's that list of stuff you're wasting?"

James Harless:              Yeah, you show me all the stuff you're wasting, and I can-

Shanon Berry:               Projects that aren't getting done.

Robb Boyd:                   Yeah. It's right next to the network map, yeah. 

James Harless:              So, that's the bulk of it is really understanding ... the other part we talked out a little bit is kind of wasting our people, wasting resources on some of our best and brightest people that are sitting in some of the most obscure parts of our organization doing little in the way of value, and then ... A lot of times people will look at that investment that's required around software-defined data center and have a difficult time grasping it, so it is important to get executive sponsorship, and let's dive into this and try to build a model that explains how this works. Obviously, if hardware-defined data center were better, Amazon, and Azure, and Google would look a whole lot different today.

Robb Boyd:                   I would assume that when ... you're also not saying, just but correct me if I'm wrong, but you're not saying that cost savings is what you would use to go sell this at an organization-

James Harless:              No.

Robb Boyd:                   ... because you wouldn't want to build it on, "Oh, think of the money you're going to save.", like we've done with some other technologies, specific technologies [crosstalk]-

Robert Looney:             It's part of it, right? It's a requirement [crosstalk]-

Robb Boyd:                   Yeah. 

Robert Looney:             ... generally for budgets. So, cost efficiency is helpful, but the parts around software-defined data center to really hunt or when you start talking about your versatility that your software-defined data center's going to offer the business, the ability to have a consistency across how you deploy one data center versus the other if you needed to, in the future, express your software-defined needs into other resources, God forbid there be a problem in the market or in the world and you can't get to your data center so to speak, you can't rack and stack quick enough, can I express a policy that takes my software-defined data center enterprise strategy to utilize a resource that's not in your data center? So, there's versatility behind moving to this path, you can't take your Big Iron storage administrator and say, "Hey, can you go help me out in the public cloud with that right now?" It's not fair-

Sean Hicks:                   Yeah, and so we see that a lot in like ... with procurement teams. The big flashy number that's always right in front of them and that they're trying to attack, is what's it going to take to acquire the solution, to acquire the technology to accomplish this outcome? And what James was mentioning, it's harder to measure the cost savings, because the cost savings come operationally, they come ... the same way we want to achieve these pay as you go models, what's funny is, that's exactly how our costs arrive, they arrive as we go. The big costs in operating are things like power and cooling, real estate if you're co-locating, the operational waste of having very, very smart highly compensated employees doing fairly menial tasks. 

                                    All of that combines over the lifespan of equipment, let's say you spend $1 million on some servers on a storage array and it has a five year life expectancy for your organization. Well, stretch that $1 million out over five years, and then stretch out the salaries, benefits, power, cooling costs, and all of that over the same five year stretch, and the solution to implement SDDC was not the big expense. The big expense was any of the operational waste that your organization was seeing through a lack of efficient use of resources, lack of efficient use of personnel time. All these things really add up and cost you more than you know in the end, and I would say most organizations don't know how to calculate those real costs.

Robb Boyd:                   Interesting, yeah. 

James Harless:              And to add onto that, I think what was kind of interesting along there was when you're thinking about those costs, again, the finance departments, purchasing, they're used to purchasing a single project, and so when you come to them with a bigger more aspirational goal like software-defined data center, you might start at a cost of four or five times what an individual project might cost, right?

Robb Boyd:                   Right. 

James Harless:              But you're going to provide resources for eight or 10 projects that are yet to be defined, that are still out in the future. So, that kind of mind shift is really difficult as well.

Robb Boyd:                   Yeah. Well, and so I want to be conscious of the fact we've just got about six minutes left in this conversation if we stick to our theme of TEC37 here, in which we should, but the last two in the article James wrote, remember this is the Six Challenges to Address when Preparing for SDDC, but the last two, and I'm going to put them together and we can comment as needed here, but treating it as a greenfield environment, and then adapting to the rapid pace of private cloud. Two of those, they're not necessarily related, I just want to make sure we get them out before we run out of time, but comments on this greenfield environment from anyone?

Shanon Berry:               Well, I think that this is one area, after taking a looking at that, that James and I might potentially disagree. 

Robb Boyd:                   Oh, I like that.

Shanon Berry:               So, I'd like for us to kind of hash it out a little bit here. Conceptually, yeah, the new infrastructure and software for the software-defined data center would likely be a greenfield, but I think that a lot of organizations when you think about greenfield, you're thinking about specifically new applications, and new workloads, and that's not necessarily exactly where SDDC has to live. One of the benefits about software-defined data centers is the ability to be able to start small and scale out as you need to move more workloads over, or you need to be able to deliver more services to your customers. 

                                    So, I wouldn't want to pigeonhole SDDC specifically just to greenfield, and the concept of new applications and new workloads. SDDC can start that way, but it can also be a great landing zone, a great migration space for existing workloads as deemed appropriate. So, greenfield to a certain extent, but also I don't think that we should look past brownfield, because I think that there is value for organizations. A lot of those legacy workloads are where customers are spending a tremendous amount of their time trying to keep the lights on, and if we can simplify that through automation that's where we're going to start to see a lot of the TCO, a lot of the cost reduction come from.

Robb Boyd:                   Well, James-

Sean Hicks:                   Yeah, and I think we should be careful about terms like greenfield, brownfield, because a lot of it will have to do with what layer you're talking about. If we're talking about the underlying physical infrastructure, because even though we're talking about extracting things to a software layer, there's still hardware running underneath the covers, right? We're not foolish [crosstalk]- 

Robb Boyd:                   Sure.

Sean Hicks:                   ... completely getting rid of that aspect, but the idea is that a lot of the solutions that enable this type of operating model, a lot of solutions that enable SDDC were actually designed to run on top of any ... as Intel's right here on the call with us, any x86 infrastructure, so the idea was that I could reuse existing infrastructure. So, maybe that's ... So, when we talk about greenfield, we're talking about building out something entirely different, but that's very hardware level conversation.

Robb Boyd:                   Well-

Sean Hicks:                   If we start talking about other areas like skillsets, right-

Robb Boyd:                   Yeah. 

Sean Hicks:                   ... or we start talking about things like the underlayment, the SDDC itself, so where we use these terms greenfield or brownfield, I think we have to pay attention to what part of the stack we're talking about people, process, and technology.

Robb Boyd:                   So, we got three and a half minutes left, and I just want to say though, I have a feeling James would agree with both of you, because one, you made great points, but I think James' specific point in the article, if I interpreted correctly, is the fact that we don't want to treat it like a greenfield environment, that would be a mistake that many people he has seen in his experience make. James, am I putting you on the right track there, or no?

James Harless:              So, I agree with mostly what was said-

Robb Boyd:                   Ah, good, yeah.

James Harless:              ... so, largely my experience is that most of the time when customers are talking about software-defined data center, most of the time they're looking for a public cloud, or an IaaS, or a SaaS outcome. So, they have this kind of specific goal that they're building towards, and SDDC is the foundational layer there. So, with that in mind, when I'm talking about greenfield in the article what I was largely referring to is the hardware. Sean makes a great point about the people and process, that's absolutely not going to be greenfield, we're going to brownfield that part of the organization for sure. 

                                    But what I would recommend customers do is do put a new environment up in the corner of the data center, and do build a team, could be dot matrixed in from your other teams, you could pull somebody from network, pull somebody from storage, somebody from compute, but do assemble that team, particularly if your outcome that you're focusing on is private cloud, do try to give them their own hardware, and their own kind of corner of the data center, because what happens when you go into a data center and you say, "Hey, we've got 10,000 servers here-"

Robb Boyd:                   Oh, yeah. 

James Harless:              "... that are all virtualized, we're going to just start layering SDDC on top of this."

Robb Boyd:                   Not in my data center, yeah.

James Harless:              It's not going to line up, right?

Robb Boyd:                   Yeah.

James Harless:              You're going to have all kinds of problems, and particularly if that team is trying to build a platform, now they're interfering with production applications. You got to give them a place where they can grow this environment, and they have to be able to build an automation engine where they can ingest service requests, and kind of pump those out. So, [crosstalk]-

Robb Boyd:                   Well, I tell you what-

Sean Hicks:                   That's a good point, because we've actually seen customers do this with their public cloud themes, they've built out teams in parallel to their traditional IT teams so that they can focus on the innovations that are available in that ... not even like a cloud team for all clouds, we see customers building out an AWS team, a Google team, and so let's take that same thing, we're trying to build a private cloud in-house with the SDDC is the underlay for that, let's take that same approach. Let's take a team of people, let's call it the coalition of the willing, people that want to learn things, that want to do innovative things, and let's put them on that path towards building out a true private cloud that's in feature parity with our public cloud provider.

Robb Boyd:                   Yeah. Well, I tell you what, we're right at our last minute here, and so I want to wrap things up. The number six on the list, which we already mentioned verbally, we didn't comment too much on, but I think it's also a statement, it's the need to adapt to the rapid pace of private cloud, because that's ... as we started this off, this was set up with the notion that there's a lot of things that have been accomplished, and there's a lot of attractiveness to what can happen in the cloud, but there's a lot of things we need to do on-premises, and their ideas to build towards that ability to do this on-premises.

                                    And this is not about doing cloud or not, it's just simply about taking what's the advantages, and learning them, and then being able to do that in your own organization, because at the end of the day, we're all here ... every company has got IT in it, every company's got some data center, and we all want to implement faster, we need to respond faster, the markets change so quickly, we need to be agile and flexible as we do it. 

                                    I do want to wrap up, and I'm going to give the final word to our sponsors here, and also just because you guys represent some great knowledge with both your backgrounds, but Robert, let's start with you as a hardware guy, final word of recommendation before we wrap up this business-oriented episode, and of course sell everybody on the fact they need to come back for the how to do it episode.

Robert Looney:             Yeah, very much so. What I would say is get started. If you haven't headed down the path, the bevy of solutions that enable software-defined data center have matured so much in the last six, seven, eight years, the bevy of partners, OEMs from Dell, our SI partners like WWT, if you haven't engaged WWT, and around the ATC and all the work you've been doing on software-defined, tremendous time to get started if you haven't. If you are along the path, start thinking through how this becomes cloud strategy for you, both your private, but also extending who you are in terms of IT services into other resources in the future. It's a fun time to be in IT, it's a fun time to be embracing software-defined.

Robb Boyd:                   Let me ask you this, Shanon, as a final word, so first of all, Dell Technologies does an awesome job with convergence, and especially hyper convergence, and you guys have come a long way in terms of offering this value, and you've also been working closely with both Intel and WWT to build out these capabilities and showcase them within their labs, whether someone can physically be in St. Louis or someone wants to access remotely. I've been looking forward to get into the how, because I know it's going to include some of the technologies and going deeper on that, but I'm just curious, last word recommendations from Dell Technologies and Shanon.

Shanon Berry:               Yeah, I think my thoughts are very similar. If you haven't already started on the journey you absolutely need to start investigating it, and I think that the best place to start is with a good partner like WWT. I've seen them work with a lot of customers, and from a consultative standpoint, I think that Dell helped to answer most of those six points that we just kind of discussed previously, and from there you have to be able to trust exactly what they're saying. They, along with Dell Technologies, we like to try and be prescriptive. 

                                    I mentioned earlier that there's no one solution that magically solves all problems, that's why they, as well as Dell Technologies, offer a broad portfolio of solutions, and we're the ones that can help an organization understand how to stitch the different solutions together so that instead of creating additional silos with their SDDC, or with their multi-cloud strategy, they have the ability to be able to create a common operating model across all of the different stacks, because that common operating model is what's going to lead to the efficiencies, is what's going to lead to the innovation.

Robb Boyd:                   I'm so glad you brought that up, because that's a great way to circle back around on where we are in all of this, which is it's not just about doing one thing in SDDC, it's certainly not a product, I don't think there's a line item you're going to find no matter who your rep is from World Wide Technology asking for, "Well, how much is this? Give me the SKU." It is something that is worth going forward on, and it is something that you can do with the help of people like World Wide Technology, and these technology partners that are here to make all that easier, because you guys have worked together to prove it. Guys, I want to thank you all for your time, I certainly appreciate it. Thank you, especially of course to Intel and Dell Technologies for sponsoring this episode, as well as the next one. 

                                    Guys, don't forget, that if you don't at least go to wwt.com and check out some of the resources they have available from James' papers to others, as well as the labs and the ability to get hands-on, and the ability to reach out and talk to a human being like some of these people here on the call to say, "How would this work in my situation?", these are the things that you could potentially do next, or at the very least you should sign up, make sure you're going to become aware of the very next episode where we get into the details of exactly how this can be done in your own organization. You've been watching TEC37, the podcast for technology, education, and collaboration. My name is Robb Boyd. Thank you so much, we'll see you on the next one. 

                                    Hang on just one more moment, I want to formally recognize two World Wide Technology partnerships that made today's episode possible. First up, Dell. Determining if SDDC is the right fit for you requires deep exploration from an experienced advisor who can identify the best technologies to support it. Well, WWT has the expertise and industry relationships to help you make that determination working closely with key infrastructure, OEM partners like Dell and VMware. VMware Cloud Foundation on Dell EMC VxRail is an agile and responsive HCI solution that powers SDDC, and together this jointly engineered solution offers reduced complexity and cost, as well as accelerated development cycles. Learn more about the benefits by downloading their whitepaper, we'll put the link there below the video.

                                    But also, Intel. Assembling the world's leading technology partners is key to creating the ideal SDDC solution for your organization. At WWT, our SDDC solutions are based on Intel architecture for the modernized data center. Intel applies advanced capabilities and deep ecosystem partner relationships to deliver transformative scalability through technologies like Intel Xeon Scalable Processors for your most demanding workloads, to machine learning, AI, analytics, and more. Intel Optane Persistent Memory, a new nonvolatile memory technology offering data persistence, and Intel Optane SSDs, and Intel 3D NAND SSDs to extend high performance storage. With Intel Technologies at the heart of nearly every solution WWT architects, we're helping to turn ideas into outcomes. So, thank you to Dell and Intel, we're going to get more into their technologies and how they work with SDDC in our part two episode-

Comments