Introduction
This week I’m joined by Mark Thiele, the CEO and Founder of Edgevana.
We start by going back in time. In 2009, Mark, Jan Wiersma and I presented ‘Death of the Data Center’ at a data center conference in London. At the time, it was pretty provocative. In this episode, I ask Mark the question: Is the data center still dead? Mark outlines what has changed and the role of today’s corporate data center. In our discussion, Mark shares his perspective on how edge and customers play a role in these decisions going forward.
Speaker Profiles
Mark Thiele Twitter: https://twitter.com/mthiele10
Mark Thiele LinkedIn: https://www.linkedin.com/in/markthiele/
Edgevana: https://www.edgevana.com
Audio File
Featured Hashtags
#CIO #Edge #IoT #DataCenter #Cloud #DigitalTransformation #CxOitk
Podcast Transcript
Tim Crawford 0:00
Companies are looking for new ways to transform their business. Technology plays a critical role in this transformation. Speed and innovation in both technology and thinking are key to this shift. Hello, and welcome to the Cxo in the Know podcast, where I take a provocative but pragmatic look at the intersection of business and technology through the lens of leading CxO executives. I’m your host Tim Crawford, a CIO and strategic advisor at Avoa. This week, I’m joined by Mark Thiele, the CEO and founder of Edgevana. We start by going back in time. In 2009, Mark, Jan W iersma, and I presented “Death of the Data Center” at a data center conference in London. At the time, it was pretty provocative. In this episode, I ask Mark the question: Is the data center still dead? Mark outlines what has changed and the role of today’s corporate data center. In our discussion, Mark shares his perspective on how edge and customers play a role in these decisions going forward. Mark Thiele, welcome back to the program.
Mark Thiele 1:06
Hey Tim, thanks man, good to be back.
Tim Crawford 1:08
So, CEO and founder of Edgevana, and for this episode, we’re going to take a turn and talk a little bit about the data center. But before we get into that, I want to take us back in time, back to 2009 when you, me, and a good friend of ours, Jan Wiersma, were all in London. All folks that weren’t originally from London, but we’re in London at a data center conference. And what are we there to present? A presentation called “Death of the Data Center. You may recall that time.
Mark Thiele 1:43
Yes, I I do I do. I mean, it’s um it was an interesting idea to create that kind of title and that kind of topic for people who had had for the most part the three of us had had branded ourselves as data center people to some large degree by that point, and here we are at a data center conference, specifically talking about the death of the data center. That was an interesting time, but it was a great starter to the conversation, and that conversation continues till today. And so I’m I’m looking to expand on you know what it meant then and and what it means now.
Tim Crawford 2:16
So let’s let’s get into that. I mean, is the corporate data center dead, or what has changed since then?
Mark Thiele 2:22
Yeah, it’s it’s interesting. You know, if we if we were to remove the last 11 years and just come to today, I would actually say the data center might be slightly more important now than it was five or six years ago. Not by 11 years ago, because the fact is, 11 years ago, there was no way to have foreseen the potential opportunity that might come about, allowing companies to do more or all of their functions via the cloud, and and what kind of resources would be available for that, and how big the cloud companies would grow, and how much diversity and scale they would offer. There was no way to know that in 2009, and so as you know, we’ve already spoken the idea of the death of the data center was more of a forward-looking idea and meant to to create discussion around the topic. But interestingly enough, and and as we go through the you know the rest of the talk, I’ll explain more of my thinking. But interestingly enough, some level of data center capacity or capability might actually be more important today than it was 11 years ago.
Tim Crawford 3:24
It was incredibly provocative back then, and even today, just in some circles, it’s still incredibly provocative. But you’re right; several things have changed in 11 years. So, is it? So, let me just kind of flat out ask the question: Is an enterprise today do I still need a data center, a corporate data center?
Mark Thiele 3:44
Yeah, I think the the question is really dependent on the type of business you’re in and the type of scale that you run at, and and frankly, where you are. To just give an example, even if you’ve decided to go mostly cloud, if you’re running some applications that require the benefit of being closer to the customer base, and that customer base is a campus full of employees. Then you still may find that some level of local capacity to run something like Azure Stack or an Outpost or Google’s Cloud at Home, Anthos, or something to that effect might still be a very interesting opportunity for you, and the thing that that worries me, and I realize that it’s very difficult to make an investment or keep an investment in something without having a known plan for how it will return opportunity to you, especially something as capex heavy as, and for that matter, opex heavy as owning a data center. We still are at the formative days, and this timing may end up being very fortuitous. The fact that we’re only at the formative days now, or we are at the formative days now, while we haven’t completely moved out of data centers already, is that we may find that there’s an interesting overlap in timing between the potential opportunities generated by new. Technology that is meant to exploit the edge and lower latency and localized analytics and AI and things like that, virtual reality, augmented reality, robotics-all of these things will and won’t have more impact on your business, depending on the type of business and/or your end customers. And so, making assumptions now for what you won’t need five years from now, I think, is a is a big risk, frankly. And so, the fact that we still have overlap with existing data center capacity again might prove fortuitous for many organizations.
Tim Crawford 5:33
Interesting. Do you think that the impact, or that there is impact from the virus and economic crisis, that’s going to impact this strategy is you think about things like colocation and leveraging colocation or leveraging cloud as part of your footprint and where you go.
Mark Thiele 5:50
Yeah, there’s there’s a couple of points relative to that, Tim, and and I think you know the first one is that when I say having data center capacity, that data center capacity doesn’t necessarily have to be an enterprise data center, but it might need to be an enterprise data center if you don’t have a decent localized service provider for managed hosting or colocation or whatever. But colocation offers a significant amount of benefits for companies who who don’t want to continue to manage data center capacity, but still plan to have some amount of work environments that they manage or need to be close to home colocation offers some serious benefits there relative to specifically the pandemic. Is that this is something that most of us are supposed to do as part of DR or business continuity planning? But unfortunately, few of us historically have us the Royal SF companies have done a very good job at it, and that’s planning for sustainability. And sustainability isn’t just am I recycling cans or am I using fossil fuels or not. Sustainability is the ability to maintain your operations and all the things that go into maintaining your operations. Well, it may sound simple and trite, but maintaining your operations involves people, and in many cases, when it comes to a data center, it involves people that are able to go into a data center. And in theory, when you buy from professionals, a colocation operator, managed hosting provider, a cloud operator, they are creating a service that is meant by design to factor in that kind of risk, and if you’re not sure, then you should make sure during contract negotiations. But by by virtue, this provides you a potential opportunity to put workloads in an environment where not only are you able to maintain your capacity strategy, your manufacturing oriented focus on how IT is owned and developed and expanded or contracted using available resources of cloud and colo and your own private space if you still have it, but also to leverage the the guarantees that are provided by colo providers and other service providers to continue to maintain operations for you even when your folks can’t go into the data center.
Tim Crawford 8:03
So, does that mean that if if I look at the penultimate kind of end of the spectrum, and that being cloud, does that mean that cloud potentially replaces the need for my corporate data center? I mean, this is a conversation that and narrative that I’m sure you’ve heard, I’ve heard, and others have heard is that cloud is just another data center. It’s just someone else’s data center. I don’t think that’s necessarily the case, but to your point, I think some things might change there.
Mark Thiele 8:30
Yeah, and and and the the key question really is, and again, the you know that we we unfortunately we have a relatively brief period of time, and going into all of the potential different business models or workload types that, based on vertical and size of company and age of company and growth trajectory of company and skill set of employee members and and operational performance skill sets, all of those things factor into what should be a complex decision making process of determining whether or not you should keep a workload or a set of workloads in your own data center, but the fact is, is that at least right now there are still demands on certain workloads, and and some of the more obvious ones are ones that all of us talk about, things like mainframes or environments where the conversion of that environment would be too cost prohibitive. You have sunk investment in your data center that you’re unable to recover, and the cost of cloud on top of that sunk investment doesn’t make sense for the enterprise right now, regardless of the perceived downstream benefits of of focus of team members, et cetera. So there’s a lot of reasons why that may still apply. But for those companies where cloud can accommodate the vast majority of your workloads, then owning a data center boils down to the potential to extend cloud offerings, as we talked about a little bit earlier in the show. Extend those cloud offerings into that data center space, even if that data center space is dramatically smaller than it was, to allow you to do things that protect the things that you believe need protecting. Whether that’s access to specific data types, whether it’s to guarantee. Specific latency requirements to protect against sovereignty issues, or any number of other things that might call for having what the cloud providers are euphemistically calling an edge solution in your data center.
Tim Crawford 10:13
But I think the the net net that I take away from that is the conversation and the decision points that you consider when you look at cloud and when you think about it in relation to your corporate data center are many, and it is complicated. It’s not a one size fits all decision.
Mark Thiele 10:31
No, not at all. And I think, I think you you know if I if you just kept talking for two more minutes, you probably would say the same thing that I’m about to say, and that that that the current environment is is both fantastic and scary at the same time. It’s fantastic because for hundreds of new opportunities for deploying and methods of deploying and app designs and chip usage, et cetera, et cetera, data center usage, distribution opportunities, network capabilities, you name it. For hundreds of things, just over the course of the last four or five years, we’ve created opportunities for what I would euphemistically call best fit IT, right? And and from my perspective, best fit IT is is the ideal environment. But best fit IT isn’t just that you’re able to buy the exact type of laptop that you always want, or that you’re always able to buy the exact type of cloud. Best fit IT requires a certain recognition of what the IT you’re buying is as compared to who you are as an organization and how the two of you blend together most effectively. To me, that creates best fit IT, and so the position we’re in now, where companies can make full decisions right at the beginning, like a small startup like myself could say, “I just want to be all cloud in the beginning and worry about making different decisions at a at a different growth trajectory of my life. And a larger company might might say, “Well, you know, I I want to keep the data to myself. I feel like I can better control performance. I can better control cost growth, etc. So I’m going to keep storage in my own data center, but I’m going to utilize localized cloud providers and colo providers for compute, etc. These are all options, along with you know, again, from a best fit standpoint, being able to use GPUs or ARM chips or Raspberry Pis, etc. and to be able to more effectively code against them. So in an era where flexibility of IT adoption has never been better, there are risks. Of course, there are always risks in making the wrong decision, but there are opportunities abound for making decisions that are less pragmatic and more perfect fit for your business.
Tim Crawford 12:34
Got it. So, what does the corporate data center of today look like then? If you were to to try and put some constructs around how to think about your corporate data center, and I use the term corporate data center to differentiate from other types of general data centers, and I know you and I over the years have talked about the fact that it’s not just the big monolithic data center; it could be engineering spaces, it could be labs. It could be any number of different spaces that have computing equipment. But what does the corporate data center of today look like, or what should it look like?
Mark Thiele 13:09
Much of it depends on who you are as a as a company and what you’ve built already, and what you plan to build. And that includes not just the physical structure of the data center and the supporting infrastructure of mechanical, electrical, and cooling, but also what type of workloads and applications might be maintained in that facility. So it’s hard for me to say to build, you know, the perfect data center for yourself in the modern era would look like this when part of your requirement is still to maintain, you know, 10 petabytes of storage and a mainframe in your environment; those have different kinds of supporting requirements than I would recommend for what might be otherwise considered an edge data center. But if we were to simplify that question just a little bit and say, if I’m building for modern applications and building some amount of accommodation on my campus for specific workloads via something like a Anthos or Azure Stack or something like that. Then what I would be trying to do is is find the lowest common denominator for delivering that service and using Edge as a assumption buster relative to what a traditional data center requirement was. So what comes with having one of these hybrid or cloud extension platforms, depending on the company, they take on different definitions in your data center. What does it mean? What kind of applications are you likely to be running on there? Is that an extension of what you’re running in the rest of the public cloud, or are they unique workloads? Are they unique workloads that were designed specifically for cloud, and therefore cloud native, and have additional failover and resiliency qualities that otherwise wouldn’t occur in applications in a legacy data center. If those things are true, then you can start considering things around you know what kind of cooling do you need? Do you need cooling? How much space do you need? You certainly don’t need raised. Can you decommission your data center and and turn what was a mechanical electrical closet or room into the data center and accept the fact that it runs at you know 85 degrees during 99.9% of the year without worrying about it? Those are the kinds of considerations where I would say whether you’re thinking about edge computing and what that requires, or whether you’re thinking about on-premises data centers going forward and what those require, my biggest first caution is know know who you are as a as a company and as a team and an ability to support and operationalize an environment. And two, don’t make any assumptions for having to carry over the legacy that you think is how things have to be delivered by from a historical standpoint,
Tim Crawford 15:43
yeah. And there are two pieces there I want to highlight. One is you’re thinking more than just the technology; you’re thinking about the team and the capability of the team to be able to support this moving forward. But then the second piece is not bringing forward some of that legacy thinking that the data center that you built and designed and and might have operated over the last decade or or multiple decades is not necessarily the same data center. We might still call these things data centers, but they might look very different moving forward. Sure, you might have that mainframe in there still. You might have some legacy footprint in there, but you also have this these new requirements that are coming in as well, and that doesn’t necessarily bode well with the old style of data centers that we might be accustomed to.
Mark Thiele 16:28
No, you think about how most data center rack environments are created today, and still the vast majority of data center racks have somewhere between three and seven kW worth of demand load in them, and that’s a very low power footprint as compared to what would be considered more high density in modern cloud environments, and even in in just more modern data center builds, which can be double, triple, or quadruple those numbers-you know, 20 to even 30 kilowatts per cabinet. So when you take that in perspective, and you take the facts-not what feels good, but the facts about what an operating environment can withstand today, all of a sudden you’re put in a position where maybe if I just guarantee good airflow, I can actually get rid of the the high cost, low wattage square footage of the vast majority of my data center and deliver everything that it used to deliver in one quarter of the space, and still have plenty of room to work in, plenty of room to expand, but create a lower cost ownership model because I’m getting rid of things like excess air conditioning and excess power support, et cetera, et cetera. Because I’m relying more on cloud as a backup and resiliency in my app designs.
Tim Crawford 17:39
So let’s talk about those designs for a minute. I mean, there is a school of thought that says data centers are becoming more complicated. It’s no longer just a room that has additional power, cooling, and network connectivity, but rather I have to think about efficiency on a whole different scale. Let’s say, for instance, that I’m of the mode that look running data centers is not my cup of tea. I’m not going to be able to get to or anywhere near the efficiency that, say, a Google, Amazon, Microsoft, Facebook might be able to get. If I’m thinking about going down that path and I want to move away from this corporate data center environment, where would you suggest someone start to think, or or start to go down that path, or should they, or maybe there’s something else they should they should be thinking about?
Mark Thiele 18:30
Yeah,
Mark Thiele 18:30
I mean, under the working assumption that even as I talked about corporate data centers and potentially using some smaller percentage of square footage, is it comes with the caveat that if there are local providers that can absolve you from having to worry about it, and they can accommodate this risk, they can accommodate the ownership model going forward, etc. then there’s really no reason for you to continue to do it yourself. But in a model where you’ve found that either the suppliers locally aren’t sufficient, or there is some other reason why your environment must stay on property. Then those are some of your options. But to think through what the operating model opportunities are associated with going to a colo or something, a managed hosting provider or something like that, and they’re multi-fold. And you know, at a time like this, it’s really hard to go to the CFO and say, “Oh, but Mark will be able to focus more on those things that are more important to the company long term. And the CFO is going to look at you and go, “I don’t give a crap. What I care about is how are you managing costs right now when we’re selling 30% less than we normally do. That’s what the CFO is going to say nine times out of 10, right? So you have to consider that when making those kind of decisions. But if that might be considered fluffy, even if it’s not fluffy in real terms, a lot of things aren’t so fluffy. Things like the ability to get access to additional networks that might offer you redundancy, might offer you better scale, might offer you better cost on your on your network expenses. The ability to get access to other service models that wouldn’t be available to you in your campus, and certainly wouldn’t be available to you in campus. During something like a pandemic, so somebody that can do managed private cloud, or somebody that can do managed private hosting, or even legacy managed hosting, and myriad other services, smart hands, et cetera, et cetera, or connections to cloud providers as appropriate, and better guarantees for those and cost models for supplying those. Those are things that the potential ecosystem of a good supplier can offer real differentiation for you. For other than trying to build it in your own facility,
Tim Crawford 20:27
and as companies start to think about becoming more diverse in their supply chains, not just in terms of manufacturing, but also could be people supply chains. You know, outsourcing development to other countries, to other locations, even their staff working from home. You know, even after the pandemic starts to subside and we start thinking about the economic crisis and getting through it, and we look at-I hate to say the term “new normal, but whatever that is-we probably will have more people working from home. How much does the network play a role in all of these dynamics starting to change, where we’re looking at greater flexibility around our ecosystems, whether it be our users, our partners, our manufacturers, developers, and the rest.
Mark Thiele 21:15
Right. Yeah. There’s no all or nothing answer to that one, Tim. But I think it’s a fantastic question.
Tim Crawford 21:21
Oh, but I want an all-or-nothing.
Mark Thiele 21:24
Well, I’ll give you one with the caveat that it’s not all-or-nothing. But the answer that I would provide to that question is that I see that there being no doubt that a wider percentage of the normal workforce audience will be working from home and eventually working from anywhere, and the reason I said eventually working from anywhere because those those are two very different delivery models from a from the standpoint of being able to get access to the customer and provide them with a service level that allows them to be successful in their home, and these combinations are important. And I realize it’s complex, but for apps that many of us still run today, Oracle, SAP apps, other apps like that that we still run in our data centers, we can want them to be microservices. We can want them to be delivered via Lambda, but they aren’t, and they won’t be until the owner of the apps, SAP and Oracle and companies like that, decide to build them that way. So it makes no difference what we want. It makes no difference what the cloud companies try to sell us. They are not better in the cloud at delivering that service to your workers, who may all be within a 30-minute radius of your office, your your your original office. So, being able to deliver that service out of your computer room might actually, or a local hoster might actually provide a much higher level of service for those applications for those employees. That being said, it’s striking the right balance and finding again the best fit. Is there are many apps as we’ve all discovered over the course of the last three months? There are many apps that are more critical for enabling remote workforce, whether it’s remote from anywhere or remote just working from home, and those apps oftentimes are better suited for delivery via the cloud because as people distribute, as they scale, as they need network in new locations, that’s not something you have to worry about as a delivery model. That’s something that the architecting appropriately for cloud delivery, or the companies that are actually building the surface themselves already offer via the cloud. So, very important distinctions between apps that enable someone to work from home and protect them while working from home, et cetera, et cetera, and apps that are allowing for people to be productive in doing the day-to-day work that is required of them in the office, right? And those are two very different categories of work environments.
Tim Crawford 23:50
Yeah, yeah, I can see that. So, where does edge fit in? I mean, does how does edge start to impact strategies and thinking around this space?
Mark Thiele 23:59
I mean, obviously, you know you and I have talked a lot. You know I have some pretty strong feelings about what edge is and what it can be, and and and where it’s going or how long it’ll take to get there. I think what you can’t ignore, in my mind, to to kind of take the simplest approach to this opportunity space, is what you can’t ignore is that edge really, to me, is where the internet was in 1992. Tim, you’re barely old enough to remember what 1992 was like in the workplace, but I’m old enough to remember looking around at Vi scripts and trying to find people’s dog pictures and whatever other crap you could search for on the internet before there were things like Google or even Alta Vista and before we had a consistent set of standards and tools for gaining access to the information,
Tim Crawford 24:45
I actually remember those times. I had an MCI mail address and CompuServe address, and yeah, long before the internet.
Mark Thiele 24:53
Yeah, so you know we’re we’re using Mosaic as a tool. We’re poking and hunting around for things. Now imagine one. Perspective. Imagine that the internet had been developed where you had to have a specific tool to get to specific web pages, right? So, for instance, Mosaic got you to 15% but you need to be on Google or Microsoft or somebody else’s browser to get to another 15% et cetera, et cetera. Imagine the complexity and the delay in growth that would occur associated with that. Imagine, on the other side of the coin, if internet had not become more and more a viable opportunity for the home user and the mobile user, and handheld devices that were actual laptops as opposed to lugables like the original compaq lugable, in theory, laptop computer was. If those things had not occurred, we wouldn’t have a wide volume of people that now had reasonable cost access to the internet, and that wouldn’t have driven people to put content on the internet, and so the internet would have never occurred in any real way, or at least it would have been a lot smaller and very different from what we know it today. Well, I see the edge as being the internet in 1992. All of the opportunities that we found after five to seven years playing on the internet that started to look obvious were science fiction in 1992. Not you know the 1% or less of us were thinking about creating an Amazon or or a travel site or something like that in 1992. By 1997, five years later, maybe even only four years later, the average organization that didn’t have a web page didn’t exist.
Tim Crawford 26:28
Yeah,
Mark Thiele 26:28
and so how can you possibly ignore the potential that that is true relative to the edge? Now, does that mean you go out and stake your claim and spend $500 million building something before anybody’s asking for it, no. But it does mean that you look for how modern technologies and your interest in gaining access to customers, gaining access to new markets, collecting more information to make your products more valuable, make your products more usable, et cetera, et cetera. That all of those things are likely to be enabled, and 1000s and 1000s of other opportunities by greater interaction with IT, and I see that certainly through Edge. And you know, to kind of cap that answer with one more thing is just there has never been a time in the history of IT where when we’ve given people more access to something, whether it was more speed, more bandwidth, more compute power, more space, more software, whatever that people haven’t figured out how to use more than we’ve created. A great example that many of us in the industry will remember is that around 2001 2002 I distinctly recall a lot of people saying, “Oh man, we’ve buried more fiber than we’ll ever use. Well, I don’t know that what I’m about to say is 100% factual, but I would be willing to bet a couple of $100 that it’s pretty close. And I would bet today that on the average year, we deposit more network in the ground today than there was in the ground at the end of 2001 in total, that tells you something about when you lower the barriers to entry for utilization and you create the economics that allow people to exploit the technologies. There is no limit to what people will do, and the first people exploiting the edge, yeah, some of them are going to look like the Chevy, the GM EV one, but it’s not going to be very long before people come out with a Neo or a Tesla, and if you’re not there to do that, you’re going to miss out. That’s my perspective.
Tim Crawford 28:30
That’s a great place to end our episode. This has been great, Mark. Thanks so much for your time today.
Mark Thiele 28:36
No, thank you, Tim. This is fun. I really appreciate you inviting me on.
Tim Crawford 28:40
For more information on the Cxo in the Know podcast, visit us online at cxo in the know.com You can also find us on Apple Podcasts or wherever you listen to your podcasts. Please subscribe and thank you for listening.
Discover more from AVOA
Subscribe to get the latest posts sent to your email.