 This is JSA TV, the newsroom for tech and telecom professionals, and JSA radio, your voice for tech and telecom on IHOT radio. I'm Jamie Scott of Gattias, CEO of JSA, and on behalf of my senior JSA, welcome to our monthly virtual CEO roundtable. These virtual roundtables lead us up to our on-site CEO roundtables at our C-level networking event, the Telecom Exchange. This went up June 20th in Hoboken, New Jersey at the W Hotel, and later in the year, November 9th, Telecom Exchange LA in Beverly Hills. If you're interested in more info, check us out at btelecomexchange.com. So let's go ahead and get started. We are gathered here today to talk IOT and the data center. We have a C-level lineup from three absolutely innovative companies, all with different interesting perspectives, COG systems, Data Bank, and IPG. And in true roundtable format, our live viewers can now type in their questions in the question box, and time-permitting will pose your questions to our speakers. And we should also note that today's roundtable was organized by the very fabulous Jonah Till Johnson. She is the CEO, of course, of Numerity's Research, and our guest moderator. Since 2002, Numerity's has been the leader in advisory and strategic consulting, specializing in analyzing and quantifying the business value of emerging technologies. And of course, highly informed on the topic of Internet of Things. So thank you Jonah for joining us as our star guest moderator today. And of course, the floor is yours. Great. Thank you, Jamie. And welcome everyone. Thank you very much for tuning in. I think you will not be disappointed. This is a great topic, and we've got some great insights to share. So without further ado, I'd like our panelists to introduce themselves, give a little bit about themselves and their companies, and Raul, I think I'll go ahead and start with you. That's good. Good morning. Good afternoon. Good evening, depending on where you are. My name is Raul Martinek. I'm the CEO of Data Bank. Data Bank is a Dallas-based national data center and managed services provider. We operate 14 enterprise-class data centers in eight U.S. markets. The topic around IoT and Edge is particularly interesting to Data Bank as our strategy has been to expand into what we consider secondary tier two markets, where we consider that being the Edge. We're starting to see the types of customers and workloads that are interested in deploying in markets like Kansas City, Minneapolis, Salt Lake, Cleveland, Pittsburgh, Baltimore, away from the primary tier one data center markets. Great. Daniel? Good morning, everyone. I'm Daniel Putz, the CEO of CogSystems. CogSystems is a cybersecurity software company. Our mission is to secure IoT. We do that by working with device manufacturers all over the world to help them build in better security, resilience, and functional richness into the devices that they're building and deploying. For us, really, that's all about ensuring the devices are behaving correctly so that they in one sense place less pressure on our infrastructure and our data centers. Thank you. It's great to be here. Great. It's great to have you. Tom? Well, thanks for the invitation to be here. My name is Tom Sivo. I'm the Chief Technology Officer for the Interpublic Group. The Interpublic Group is a global advertising and marketing services company based in New York City. About $8 billion in revenue, 50,000 employees, and I'm responsible for the core infrastructure. This is running the global network, our authentication platform, our unified communications platform, our help desks, our enterprise data centers, and setting the technology standards. Okay, great. So folks, as you can see, we have three different aspects of the look at data centers and IoT for counting the research advisory component. I think what I'd like to do is go ahead and address the elephant in the room. We've been talking about this a bit beforehand, but the real question and the reason for this panel is does IoT fundamentally change the paradigm of data centers by moving from a centralized, standardized, consolidated paradigm to a distributed, as Raoul said, edge paradigm? I'm going to go ahead and, Raoul, I know you have some thoughts about this, so I'm going to go ahead and ask you if that's what you're seeing in your market and then just sort of open it up to everyone else. Are we actually seeing a paradigm shift or is it too early to tell? Yeah, I think we're definitely seeing an evolution. As I mentioned, Data Bank has expanded and built facilities in kind of what we consider a tier two NFL cities, tier two markets, and what we see that as today in 2018, that's where the edge is. That's where kind of cloud and content players are deploying production nodes to get to their end users, in addition to obviously building their large core nodes. We see a lot of interest beyond that as far as the cell tower, but so far those kind of discussions are preliminary and there really isn't kind of a use case to apply that. But if you follow the logic of data, it could happen. Well, and Raoul, before I turn this over to the other panelists, one quick follow-on question on that. Are you moving into the ancillary markets or whatever you want to call the sort of smaller cities, is the driver there a latency, i.e., you're trying to get closer to the end user, or is it real estate? The cost of real estate is simply lower. Yeah, I wouldn't use the word latency. There's only a small subset of applications that are really latency sensitive, but proximity is probably a better term. Having kind of that infrastructure in a market like, let's say Kansas City, it's a 2 million MSA market, you can have a cloud and content player delivering that content from Dallas or Chicago or Oregon where power is cheap, and they just see such a volume of content going to that market that by putting a small for them node in that market, it just means that it's a better end user experience, more reliable, cheaper for them, more distributed. If you think about Facebook, it's not really a latency sensitive application, right, your video, your pictures can take some time to get there. Facebook isn't, but I would argue that IoT is. But let me go ahead and ask some of our other panelists. Tom, I know you guys at IPG have invested quite a lot of time and energy into centralizing data centers. Are you even considering or looking at the possibility of distributing them and is, if so, is one of the drivers IoT? I mean, what are your thoughts? Yeah, absolutely. From a data center operations core infrastructure perspective, IoT obviously is a real challenge and it's happening to the enterprise as opposed to the enterprise out in front, really designing and preparing properly for IoT. So I think it's happening to a lot of data centers, just like ours at interpublic, and it's going to be your worst nightmare. It is going to highlight and exacerbate the things that you do not do well from a data center operations perspective, like, do you have a robust asset management system? Do you have a change control, change management, change management database process and system? Are you able to vulnerability scan your environment quickly and patch updates and upgrades right away? For your IT landscapes, are you preparing IoT in a development QA testing, pre-production, production release environment, or is IoT just happening at the endpoint, devices attached right to your production network and you're reacting and responding to it? These are the real challenges that are facing agencies within interpublic and I would think on a broader scale across the sector. I can absolutely guarantee that the rest of our clients are wrestling with those operational challenges, Tom, and I love that you brought those up. Coming back to the whole edge versus centralized, Daniel, are you seeing any trend towards decentralizing data centers as a result of IoT? Yeah, I think IoT is such an interesting space because of the way it changes the variables, right? I love the idea of talking about proximity as well and thought about that. We talk a lot about latency and for us, what we were seeing is a lot of the driver in IoT use cases and applications is to try and do more things at the edges of the network, more time-sensitive applications in some cases, and a lot of those in truth don't really exist yet because the infrastructure and the capabilities are not there to support them. So that's the exciting thing around the data center expanding closer to the very edges of the network and into the home potentially, is that we're able to respond quicker from an application point of view. Which is great. Tom, you were raising the whole issue of operational impact and as Daniel was speaking, I was thinking, good God, there's the whole question of how to secure the IoT, which of course is Daniel's area, but also, Tom, you said it's the IT department's worst nightmare because the advent of IoT will expose any flaws that you have in data center operations and you led with asset management and the lack of asset management as being something that would be a huge issue. Can you expand on that a little bit? Why is that? Sure, it's a big problem. So if you think about a centralized data center organization, right, there are a lot of math majors, one thing IT organizations don't do well is count, right? So we can count desktops and network endpoint devices and the servers and storage in real time, which is what you're going to have to do in an IoT environment. These devices are scanning and monitoring, collecting information real time and you have to know where they are so that you can at least manage, mitigate some issues that they will create for the enterprise. So in central IT, we are doing some fundamental network redesign to handle the IoT world like creating a separate dedicated network for IoT devices with limited capabilities of what they could access, basically the internet and some ports that we allow as a start. But there's a whole program behind getting the enterprise and the user base aware of what these devices can do to enable business, it's needed for business, but also to protect the business as well. Hey, Tom, I was gonna comment that, that's a great comment and what we've seen some of our larger enterprise users do is kind of bifurcate how they do that similar to what you're describing. So for example, we have a customer, that's a large industrial manufacturer of kind of electrical type of components. Obviously those are very ripe for potential IoT type of applications and what they've decided to do is separate kind of the collection, assimilation and analysis of that IoT information into the public cloud and keep it 100% away from their kind of production environment inside their, in their case, two data centers. So that's another way I think enterprises are kind of handling IoT is by saying, okay, I'm gonna make it Amazon's problem or Azure's problem in terms of this flow of traffic going in, being ingested into some network that we then need to figure out what to do with. So, yeah. It's your point about a slow evolution, right? So first we'll set up this dedicated environment but eventually it has to be, IoT has to be integrated into the business and business processes and departments and organizations. So we will find a way of hydrating the IoT infrastructure just like we've hydrated the cloud. So that's the new world, but it's real time. And that's my point about our organizations really ready for real time monitoring and sensing and most importantly reacting, right? So think about what happened this week. Alexa devices randomly laughing, right? What if those Alexa devices on your enterprise network were randomly sending out malware? What are you doing? Do you know, can you detect them? Can you isolate them? Can you back up and recover to a better point? Those are things that are on our minds and what we're wrestling with. And that back up and recovery is very interesting but when you led with asset management, I wanna come back to that. And Daniel, I wanted to invite you to weigh in on this because there's another area where asset management becomes critical and that's when you're implementing a zero trust security model. So essentially you cannot implement zero trust unless you know you have an up to date accurate sense of every device, every service, every virtual machine in your network in real time which changes the problem of asset management considerably. And Daniel, since you guys are in the business of IOT security, is there anything you wanna comment on the whole notion of IOT potentially driving folks heavily towards that zero trust model or is it just not something that's arising yet? I know it's really, I mean, it's an interesting problem space, you know, the, what we see is a lot of fragmentation and if you think about almost everything that's being built today, every appliance even becoming connected, that's a lot of fragmentation, right? And a lot of companies that lack security experts and because there aren't that many building these devices and connecting them to the network. So starting with that premise of yeah, zero trust is probably the way to go. You know, the other thing I'd add is it's, what we're seeing is really interesting from, you know, denial of services right now, you know, the die indeed, us attacks, using Mirai and so on. And, you know, it's a peak of what's to come. Yeah, I think it was last week, you know, GitHub got hit by a record rate. Yeah. Yeah. And it's impressive, you know, was at 1.35 terabits per second, you know, 127 million packets per second, basically into their infrastructure into a data center. And then they stood up to it, which is amazing, which is, you know, well done. But I think, you know, every couple of weeks we'll start to see that number push up. And so one of the big challenges is, how can we ensure that our edge devices aren't participating in that, which, you know, is gonna be this sort of arms race process of trying to improve them. And, you know, at the same time, making sure our data centers are impacted by it and out of the problem as well. Which is an excellent point to raise, and I wanna circle back for anyone listening. Essentially what Daniel just said was, when you're rolling out IoT, you've just given the bad guys an order of magnitude or two order of magnitude increase in their botnet armies. Because now you have all these little devices that can be used to hammer GitHub or whoever you're going after. So, you know, instead of one or two devices per human, which is the norm in most enterprises, you've got thousands of devices per human in an enterprise, each of which can be infected with malware and be used as an army. So, one of the challenges is making sure that doesn't happen, because aside from the fact that that incriminates you in bad efforts, it also means that your devices aren't doing what they were supposed to have been doing, whether it was sensing or controlling. So you're not getting the benefit of them if they're out there participating in this army. Ro, I wanted to circle back on something you had said earlier about the whole notion of bifurcating IT. Tom had pointed out that that's a standard approach. Anytime there's the new technology that's big enough to be game changing, the best approach is to dedicate a team to focus on it for a while and understand what the issues are and then sort of eventually blend back into larger IT. Can you elaborate a bit on how you would see that evolving for your clients? And, you know, for the rest of you, if you have thoughts on that as well, I'm asking Ro because he brought it up first. Well, just to take a quick step back, you know, most of the enterprises we deal with which are, you know, mid the large size enterprises, they're in one of two phases. Number one, believe it or not, the majority of enterprises are still kind of own and operate their own data center assets. I'm talking about the physical components of the data center, the generators, the UPS, the access control, the electrical system. So there's a trend to outsource that to third party providers. When they finally get that done, then they start thinking about redundancy, where, you know, we've done our primary location, what do we do from a redundant location? You know, most companies, even very large ones, you know, try to focus, try to reduce it to a reasonable number, kind of two, three, four data center locations from which they then support their users and their customers. I think what IoT is now kind of is putting into question, you know, should there be further bifurcation of that? You know, in our case, we're, you know, we're providing the physical layer of then all these systems that Daniel and Tom's groups then manage. So we're responding to their kind of requirements around, you know, is geography important, where that geography should be? And I think that we will see kind of a splintering of that consolidation and a desire to stand up, you know, different types of workloads for different types of uses like IoT, like we're seeing it for things like artificial intelligence or deep learning. These are workloads that didn't exist, you know, three, four years ago. And now all of a sudden they're showing up in our data center. So, you know, the network, I'll obviously interconnect all that stuff, but ultimately the users like Tom will end up deciding what makes the most sense in terms of locations. Tom, Daniel, any thoughts on that? Well, Raul brings up a good point. You know, IoT isn't just about data collection. You know, you're collecting data, you're sensing data, you're monitoring data to do something, right? So the processing on your enterprise is going to be great, right? Because you're gonna have all this data that is eventually going to be feeding artificial intelligence systems, programmatic systems. And do you have the network bandwidth internally and externally for that? Do you have the processing power for all that data? Do you have the storage and backup and recovery capabilities so that your AI systems are real time? These are things that you have to look at comprehensively beyond the first stage of just putting out IoT devices and sensing things. There's a ripple effect that the business expects from that investment and you have to make sure your infrastructure handles those necessary business expectations. And of course, being in the advertising world, you think a lot about analytics. So, yeah, where do you get it? It's a core part of our business. The analytics translate into programmatic media placement and media buy. So there's a lot of data, a lot of churn, a lot of impact on the infrastructure that we are just recognizing now and it will greatly improve. Yeah, Tom, if I could give a kind of a potential real world example of that. We're opening up a data center in Atlanta where having discussion with a very large national logistics company that collects so much information based on all their IoT devices and sensing devices that they literally just throw away the information after a couple of months because they just can't see it. So the question is for that type of user, what's the best in our case data center strategy? Where do you wanna collect that data so that you can get or make useful decisions out of it? And it's probably not, you're not gonna collect that into two locations, you're probably gonna wanna collect that into a more regional type of architecture and then kind of figure out that computing aspect of it and the analytics and getting that back to the user. So, it's an exciting time to be in our business because of all these new kind of technologies and ideas just starting. And I think we're gonna see a lot of development on this in the near future. Definitely. I know we're coming to the close here, but I wanna toss another small grenade into the room here. Rolly, you said something interesting a little while ago and about making all of this, Amazon's or Microsoft's or IBM's problem. And by the this, you meant the analytics of a lot of the data that's being collected. I'm interested in how IoT potentially transforms relationships with third parties, not just from a data center perspective, but also from a hosted analytics perspective. And what I mean, that's kind of a consultancy way to say, good heavens, is it safe to put all your stuff in Amazon? What happens if Amazon screws up? What's my company's liability? What, how do we make this decision? On the one hand, IBM or Amazon is better equipped to do the analytics than we are and to do the backup and the recovery. On the other hand, you're essentially putting all your eggs into one basket. Does anyone have any thoughts on the best way to approach this challenge? Put five people in a room and you'll get 10 thoughts, I think. I think this is a challenge of enterprises, the folks we speak to, they all have their different approaches. Some believe strongly in the public cloud being a solution. Some believe strongly that it's not the solution. I don't know if there is a right answer. I think it depends on the entity, their compliance regime, their security regime, their preference for CAPEX over OPEX. There's a lot of variables that I think go into that decision. And I think what we see today is a lot of people trying a lot of different things in a lot of different ways. As we kind of learn collectively, what's the right approach? What's the best paradigm to tackle this with? Yeah, I would add too. Yeah. Sorry, I would add to Raul's point the whole concept of information protection, data privacy, as you begin to turn on these IoT devices, you've got to be real aware, cognizant of what rules and regulations at the country level, if you're running a global business are evolving. And just like you have to capture emails and instant messaging and produce that information for court cases, you'll be doing that for IoT devices and what they record. So you've got to take that seriously. Yeah, I agree with that too. And I think back to your question, having your own data set infrastructure and cloud solutions working together to interoperate is going to be, I think, one of the criteria for success in the future. So being aware of, as you said, security implications, locality, jurisdiction, all that kind of thing as you're building this out is critical. Okay, well, folks, I know we can be talking about this all morning and afternoon. Raul, I liked your comment about five people, 10 opinions. I'm thinking, well, I should get at least six opinions then from asking this question. But what I would like to do is kind of sum up very quickly. If each of you could just boil up everything we've talked about into something actionable, whether it's for a provider or for an end user or both, anywhere from one to three things, just key takeaway for someone who's been watching this and saying, great, we've discussed a lot of things, but what does that all mean? What should I do differently having watched this than I would have done half an hour ago? And Daniel, since you've been last before, I'm gonna start with you. What are some of your wrap-ups takeaways that you would recommend for either end users or providers or both? Yeah, I think for the consumer, for the end user, be the discerning consumer. One of the things that we need to look out for is what devices we choose to put in our homes and in our enterprises. So the brand and whether that company's talking about their security posture, that kind of thing, whether they're doing updates, they're important sort of table stakes for selecting what you purchase and put in your home. Okay, great. So be discerning, make sure that your provider cares about security both publicly and privately. Tom? Yeah, right. I would say from an enterprise data center operations perspective, if you haven't thought about setting IoT policies, you better start doing that now and seek third-party professional help in doing that. And as you establish policies for IoT, the standards to be used and how they are to be used, you need to educate the user base, whether those are your employees, your contractors, the clients you collaborate with, a good education training program so that they are aware of the technology, its capabilities, but also its liabilities. Okay, so be aware of the challenges of IoT, set your policies now before you think you need them. I like getting help. Yeah, policies and training. Policies and training. And I like that notion of third-party help because, hey, I'm a third party who can help. So I appreciate the plan. And if you can wrap up, that would be great. Yeah, I'm sorry, just froze for a second. I would have two major takeaways, which is from a data center operator perspective, I think we'll see the rise of significant data center assets in the secondary tertiary market, specifically for applications like IoT. And then to echo on Tom's point, I think what you'll see is end users embracing hybrid type of approaches where it's a combination of public cloud, private cloud, on-prem and co-location. So I think there's gonna be more integrated solutions and then more diverse options. Okay, so think hybrid, think secondary. And Jamie, handing back over to you two, wrap this one up. Well, thank you everyone. This has been such a hot, provoking, insightful roundtable on IoT and the data center. Again, our all-star panelist, Mr. Daniel Pog, CEO of Pog Systems, Rahul Matinek, CEO of Data Bank and Tom Sevo, CTO of RPG and of course, our star guest moderator, Ms. Jonah Till Johnson, CEO of Numerity's Research. This wraps up our latest virtual CEO roundtable. Come meet us in person, June 20th, Telecom Exchange, Hoboken, New Jersey, November 7th, Telecom Exchange, L.A. To feature your C-level here, next time, email us pr at jsa.net. Thanks for tuning in to JSA TV, the newsroom protect and telecom professionals and JSA radio, your voice protect and telecom on iHeart radio. Until then, happy networking.