TEC37 E11: The Future of Network Architecture


In today’s ever-changing world, do you ever think about how your organization’s network must evolve with it? This discussion will talk about the macro factors to consider when looking at the future of networking and what happens if we think differently about the problem. We will also cover how sales and services models need to evolve with the network.

Please view transcript below:


Robb Boyd:                   How fast can you name five monumental changes affecting the way we work, live, play and learn? 5G, Wi-Fi 6, edge computing, artificial intelligence, machine learning, cloud, SaaS. And I bet each one of you came up with a different list. All distinct technology changes, the common denominator, a network. A network that has not been keeping pace. It's not the connectivity requirements, but the demands. Today's network architecture was largely formed before we really knew how it would be used. That architecture is now a bottleneck slowing our ability to embrace the benefits that these changes represent. The future is already here, but for many of us, the network is still not ready. 

                                    Well, we have two leaders from World Wide Technology, each with over 20 years of network experience across multiple disciplines. Bill Thompson is director of Global Solutions Development. And Neil Anderson, a senior director for Network Solutions. My name is Robb Boyd, and you're watching TEC37, the podcast covering technology, education and collaboration from World Wide Technology. 

                                    Well, gentlemen, thank you so much for taking the time to join us. I've been looking forward to this conversation. I thought it might be good, because this keeps me from messing anything up. But let's start with some introductions of you guys, explain what you do. You both work for World Wide Technology, but you both come here with years of experience in the industry and with a lot of different customer situations which are going to be valuable for today's conversation. But Bill, let's start with you, how do you describe what you do to the lay person?

Bill Thompson:              Sure. Thanks, Robb. So my name is Bill Thompson. I work at World Wide Technology and I lead up the Global Solutions development team. We're a team of experts that like to dig into the ones and zeros of technology and figure out how it works. We have skill sets across data center to include compute, data center networking and storage, all the way to network across the WAN. Things like SD-WAN and into the campus where our users are using Wi-Fi and researching other things, such as 5G and how can we connect to the network.

Robb Boyd:                   All right, excellent. And Neil, what are you responsible for, there at WWT? 

Neil Anderson:              So I lead up the Network Solutions team at World Wide. So we have a team of architects that really goes out and talks to customers a lot about what they're trying to do, what are their business goals, and then we map networking technologies and architectures to help solve their problems. Pretty big networks we're talking about here on a global scale. 

Robb Boyd:                   Yeah. In my experience with you guys, you've got your fingers in a lot of different network pies, so to speak, at a bunch of different levels from service provider down to mom-and-pop and everything in between. And I think that's perfect for what we're talking about today as we talk about the future of network architecture. And specifically around architecture. As we were preparing for this, and I don't know who wants to lead off here, but the notion of macro disruptors, the big things that shouldn't probably be a surprise to anyone in this audience, but it's how we're dealing with these things and where they're driving the need for change is what I really want to get into, but I wonder if you could speak to the macro disruptors as you are seeing them, and then we'll get deeper into how we begin to address some of these things. Because disruption is a good thing in this case, I think.

Neil Anderson:              Yeah, so I'll take that one Bill. And essentially, we see a whole bunch of disruptors, but the big four that we are talking about right now are really around what we call multi cloud architecture. Applications have moved to public cloud to SaaS, and still private data centers, but the combination of all of that we call multi-cloud. And so most networks were not built for that. And so that's a huge disrupter to the networking architecture.

                                    The next one is around high-speed edge. So if think about Wi-Fi 6 and 5G coming online. Those are pushing tremendous speeds to the edge of the network. And again, the network in between the applications and those high-speed edges really wasn't built for that kind of capacity today. Then we also have automation. Automation is a huge thing that all of our customers are looking at. They don't just want to replace their networks with new routers and switches, they want to change the way they're really conducting IT. And that takes automation. So that's a huge one right now. 

                                    And then the final one is business continuity, or in a situation we're all in, we're doing this broadcast remote as is just about everybody. But that's tremendously on customers' minds right now, "How do I flip the workforce out into people's home offices and still make them productive, still make them secure, still get them the quality of experience to the applications they're looking for?" All of those four have just huge implications on the network architecture.

Robb Boyd:                   Well, you know what? I'm going to give this one to Bill then, because I think that's a good setup. And I'm curious because this is what I think we can build on is, and I know there's no typical network anywhere. Everybody's network is extremely unique. But Bill, from an architectural perspective, where would you say we are now and how is that maybe not exactly where architecture needs to be for dealing with these disruptors?

Bill Thompson:              Yeah, you could say in a way, I compare it to how we have looked at the world in the past. Meaning at one time everyone thought that the world was flat when looking at a map or traveling. And-

Robb Boyd:                   It's not though.

Bill Thompson:              ... we know that it is a round and almost 3D type of model. So thinking about networks that way, traditionally we've thought, we have these applications and they reside in a centralized location, probably with some redundancy there. But we always had to design our architectures around that model where we just got to get everyone back to the central place. Our users, typically were in a branch location. But now we're learning that the earth is indeed round and we have to start thinking differently about how we architect these networks. As applications leave the private data center and start being consumed like Cloud-as-a-Service or public cloud offerings, now all of a sudden we have these applications residing everywhere and the way we design networks has to shift to take that into account.

                                    Even more recently, our users are being distributed everywhere. When we look at the work from home standard that we're in today, now we've got this entirely distributed system of users everywhere, applications everywhere. And it changes how we have to think about designing and architecting these networks.

Robb Boyd:                   Yeah, I think what's weird, with the lockdown that we've been under and everybody working from home and I think overall and I'm sure you guys were involved in this with several different customers, there was this huge shift of traffic patterns when you look at service provider traffic and it wasn't from Netflix and other things, those are things that I think are pretty easy to deliver, store and forward type videos not a big deal. But this type of thing where we're actually streaming synchronously back and forth becomes a whole different model. And now these are coming from different sources. They're coming from our home networks which may or may not be experiencing whatever they experience from neighbors who may be doing very similar things. Who knows what's going on behind all these closed doors around us? It's too hot to go outside so I'm not going to find out for another couple of months regardless.

                                    But I feel like some of the patterns that Neil's talking about with regards to the macro disruptors also, they all have this interrelated nature that are driving each, whereas this push to the edge feels like a response to the fact that as things move to the cloud applications are no longer in the data center and then it causes us to go, "Wait, how do I deal with continuity now because it's in the cloud? Do I still need to worry about continuity? Maybe the cloud takes care of it." Probably not the best attitude to take. But nonetheless these questions come up. And it feels like as we're all under this, I'm curious if you guys agree, it feels like the campus is under pressure to really change because we've spent years designing high-end super campus networks that now everybody is saying is the real estate still as necessary, which means that then is the network still as necessary? Is that going to drive changes with the Wi-Fi?

                                    But with all that said, as I continue speechifying here, where's this going to bite us? So as we take this architecture that Bill's talking about, that hasn't changed as much as it should, over time, and these macro disruptors are saying, "Wait, everything's different now," where are we feeling it? Are we feeling in this thing already? Or is this something that we need to be prepared to see or maybe some combination?

Neil Anderson:              I think it's really interesting. And I mean, no one has the crystal ball. But if we look at what our customer base is talking to us about and what they're planning for, intuitively, you would think that, well, no one's going to build campuses anymore. Everybody is going to focus on the home office environments. Most people believe that, and what we're modeling is somewhere between 20 and 30% of the workforce may stay permanently at home. But we are still seeing a dramatic number of projects that are coming to us around campus refreshes. And you could wonder why is that? I think it's a couple of things. One is that they have the ability to do it without downtime right now. So while the-

Robb Boyd:                   Excellent time to get-

Neil Anderson:              ... buildings are empty-

Robb Boyd:                   Yeah. 

Neil Anderson:              ... let's catch up on our technical debt that would have taken a lot of maintenance cycles to get to, they can do it pretty fast right now. So we're seeing some of that at our company. The second thing that, I was just talking with someone earlier today, and there's a feeling that what happens with the workforce if they stay home and they never come back to a campus? You lose a little bit, potentially of loyalty between the employer and the-

Robb Boyd:                   That's a good point.

Neil Anderson:              ... employee. I can work for corporation X just as well as I can work for corporation B if my home office isn't changing. So it makes people a lot more portable, maybe a lot less loyal. And so that's something that some of the companies are starting to think about, is employee retention becomes a huge, huge issue here.

Robb Boyd:                   Well, and I think anytime, as soon as I make a guess at anything, it's going to be wrong. And I don't think campuses are going away because I think as soon as everybody goes one direction, then it becomes the competitive differentiator to do the opposite regardless. And I think as much as I'm a big proponent of remote collaboration, tooling and stuff like that it has its place, but there's something about face-to-face creativity and how you get things done in serendipitous discovery that I don't think is going to be replaced anytime soon. But I'd love, if there's anything to love about what we're going through, is this has forced all of us to get better at certain things that we didn't have to be that great at before. And so it's really changed that game.

                                    But yeah, mentally I need us to get back together on an irregular basis at the very least, because I need that variation. I've worked at home for years. But I'm curious when it comes to these network patterns that you're seeing in the disruptors, are we seeing scaling issues? I think the big thing is that everybody is always talking about the market. If you've got Wi-Fi 6, which is we're getting so much more just bandwidth changes that come from the amount of stuff that can be sent from devices, that then has to be progressively larger as it goes. If you took a traditional core aggregation edge type of layouts.

                                    But then you've also got 5G, which is going to eventually be here, though it seems like it's more talk sometimes and elsewhere. But as we all know service providers are working very hard on how are they re-architecting as they move from traditional Backhaul to Midhaul, Fronthaul and other types of ways of getting that additional traffic and driving additional services. But where do you think we're going to see cracks in that foundation, first? Or are we seeing it now?

Bill Thompson:              I'll take that. Robb, to your point, that's something that we refer to as bandwidth competition. And traditionally, that's been looked at as competition between applications. But to your point about scaling issues, I think another dynamic is being added here, especially from say a work from home standpoint. Now I'm not only competing between applications that are relevant for me to do my job, but I also have, my spouse is doing her work, my kids are on their virtual. We're not only competing against applications but also with other people. How do we prioritize those, so everyone can do their job? 

                                    That's where 5G and Wi-Fi 6 come into play. Now in a residential place, not sure how much Wi-Fi 6 will be applicable. But these are certainly things that others are looking at, as we get back into work and go back to offices, what if it's shared retail space to help reduce those costs, and we're competing against the other people. So I think that's where we start looking at new technologies such as Wi-Fi 6, 5G for just simply connectivity to the network in ways to reduce the amount of competition there.

Neil Anderson:              And I think, to add to that Bill, I think there's a lot of angst about how do I measure that the person in that home office is getting a good experience? Do I know that their applications are performing well? Do I know that that person's productive? Do I know that's secure? All three of those are really causing people to look at this. And then when you move upstream from that, like the service provider backbones, they've definitely had some strains and areas that were not intuitively obvious. Everybody thought, "Oh, well, Netflix and streaming must be off the charts." It really didn't affect them that much. People are streaming pretty much the same as they were. But what's really, obviously the broadband networks are under a pretty big strain right now and they've actually had to shift a bit of capacity over from the cell phone networks, which you can imagine, people aren't moving around as much. So the cell phone capacity is relatively low compared to the broadband capacity. So they've had to do some quick dancing there to shift some of their own capacity just to handle that.

                                    And I think some of that, I think you're going to see long-term effects on as they roll out 5G, they'll be thinking about both of those audiences. What is it to the consumer audience? What does it mean to a business audience? What does it mean to a home office, perhaps? 

Robb Boyd:                   Yeah. Because I think, the biggest thing I'm excited about on 5G is not necessarily as a consumer, increased bandwidth or anything like this, but it's the new services that are being driven out of this, which also take into account changes in terms of how we deploy stuff out to the edge. I think one reason why your Amazon Prime's, your Netflix and 100 other streaming services that we have access to these days. One reason I think those guys don't blink too much is that largely we've already figured out how to catch that stuff at the edge. It doesn't have to travel dramatically differently. But as we go forward, these new services that have increased timing requirements at the edge for new applications, whether it be automobiles or robotics or healthcare, and things like this, those are the things that are exciting to me because we're really pushing to do things different, when we talked about ultra low latency, wirelessly and things like this that are coming about with these changes in 5G. 

                                    But you mentioned applications, and I really want to double down on this one because I feel like applications are driving a need for change. But I'm curious, repeat some of the points you said, and let's talk about these, why are applications... Let me word it this way, how are applications driving this need for architectural evolution? Or revolution? Yeah.

Neil Anderson:              So I think first of all, we've seen almost a tipping point with a lot of customers where the majority of their application traffic is now outside their private data center. When you think about SAS things like WebEx and Zoom and Office 365 certainly, has really tipped that traffic profile to where now the majority of my traffic isn't going to my data center. But a lot of people are still routing their traffic there first and then going out to those applications. And that's just not the optimal way to do it. It leads to all sorts of challenges. And again, the point that we said at the top of the broadcast, the networks were never built for that. They weren't built that way. They were built for, I'm going to aggregate my traffic someplace and then if it needs to go somewhere I'll go ahead and route it but that's just not the way traffic patterns are matching right now.

Robb Boyd:                   How much, Bill, you mentioned your teams are taking care of working with a lot of customers on we're moving from the typical WAN architectures hinted at here, where everything is coming in one location, are you seeing a big uptake in terms of SD-WAN? And how that may be changing, because it feels like with SD-WAN it's just silly, the whole point I feel like of SD-WAN is to re-examine your security model, re-examine where applications are being accessed. But implies doing business a bit differently. And it's driving architectural change. Are customers picking up on SD-WAN and taking advantage of that, or are they just doing SD-WAN in the same old models?

Bill Thompson:              Yeah. I think SD-WAN in general gives us the ability to be very agile in how we shift those different connections. So one such thing is we look at as applications move out towards the cloud and closer proximity towards the users, what we're seeing is it's easier to use automated and centrally managed systems like SD-WAN to deploy those quickly and adapt to the user's needs. I see it as the universe is always expanding, while the architecture is always expanding outwards, to get better proximity towards the user. SD-WAN is one of the enablers that allow us to do that, I guess simply in a manner of speaking.

Robb Boyd:                   And it changes the, I feel like it changes the security model too because I think one reason that we got that trombone effect, Neil, that you're referring to there is because we've always had the, and maybe it was really the only way to really do things. But it was that old, I forget what the right historical reference is. But you want to funnel all your traffic into one place and then secure the heck out of that one place. And so it made sense in a headquarter location when your data center is in one location, that you're then going to spend your extra dollars on security at that one place. And everybody's going to come through there before they go back out to the big internet.

                                    Well, now, as you mentioned, with applications in the cloud, that doesn't make sense, it's not the most efficient way to do it. But that also implies, if you're going to do it differently, then you've got to pick up all the security mechanisms that hopefully were in place, need to be carried just as efficiently and securely in all these smaller locations. It feels like that must be one of the big drivers you guys are seeing from an architectural evolution. 

Neil Anderson:              Yeah, absolutely. And this isn't new, the dynamic between I want to have a centralized internet edge and bring all my traffic there where I place my central DMZ. The InfoSec teams love that, because it's a nice model for them to be able to secure things in one place. But there's definitely pressure to distribute that out towards the edge. So you have this pendulum swing between, "Okay, do I have a central internet edge? Do I manage thousands of locations each with an internet edge?" Security teams hate that. But there's a happy middle ground emerging that we're seeing where you have regional security models and take advantage of things like the Secure Access Services Edge, the SASE model that's quickly becoming a sister with SD-WAN. We're seeing a lot of those two architectures coming together where people are deploying SD-WAN, they're also thinking about the new security model to your point, Robb, and a lot of times we're seeing that those strategies have to align. 

                                    So the SD-WAN team wants ultimate flexibility, "I want any-to-any connectivity." The security team says, "Wait a second, we're talking about internet access here, we got to be a little bit more cautious than that." And we think that the middle ground emerging is this regional hub model. 

Robb Boyd:                   Yeah. Well, let's go more into that. Because I think as we were preparing for this, you guys were sharing just some different ways of thinking about how this change is driving and where maybe our audience needs to be thinking about how they are responding to these changes and preparing for them, I guess, is really the right way to do it. Because the idea is not to be too reactive. But how do we begin to think differently about these challenges? You had some what ifs, I think, if I remember correctly, that were interesting ways to say imagine how you might respond to this. Do you mind covering some of those? Do you remember that part of the conversation, I'm hoping?

Neil Anderson:              Yeah, absolutely. I can start Bill, and then you can take on. But we were supposing, what if the traditional DMZ goes away? Well, you need distributed security model, you need to be able to essentially place your security dynamically between users and applications. And that's how we're starting to think of this now, is how do I bring applications as close as possible in proximity to users? So it's almost an any-to-any fabric is the way I think about it, rather than a traditional hub-and-spoke topology. I've got to think of it almost like a fabric where I can get on and off at any point between users and applications.

                                    And so security needs to be a part of that. And so one of the what ifs we were thinking about as what if, that model of fabric is the right way to solve the problem, what does that mean for the security dynamics? What does that mean for how do I treat application performance and monitoring between that because I no longer have that tidy model where everything's coming over my network? I may be traversing all kinds of different things and going through Equinix data centers and regional colo models. It really just changes the way you have to think about how you're connecting your users to the applications and where the workloads are. 

Robb Boyd:                   Yeah, Bill, and I'm curious, do you think is this something that needs to be consciously thought of in terms of when the network administrators, are they really unders... I shouldn't say it this way. I'm trying to think, how much is this decision being made? Because I feel like we've got global cloud providers that are also starting to route traffic through their excess capacity, where they almost become service providers in this sense, because it also makes sense perhaps, if they're serving like a 360 Microsoft type of thing, that might be the best way to get to their applications, you hop on and off their cloud, but you could be traveling a good expanse of time or space. What am I thinking of? Territory. Wow, I can't think of the right word. But how much is that as a conscious decision these days? Because I agree this mesh notion of a fabric orientation versus hard static routes, perhaps as diametrically opposed examples. Does this mean we need to think about it differently? Or is this stuff going to be happening somewhat naturally?

Bill Thompson:              I think we have to think about it differently. We certainly want those easy buttons that we can provide a great user experience to everyone, hence connecting directly into like a Microsoft fabric or something to that effect to get a quality O 365 experience. But as I mentioned earlier, the network is constantly expanding, we always have to think about it going beyond those regional borders and eventually looking at future things like edge compute, expanding the cloud into the private data center, expanding the cloud and applications almost next to the users thinking applications and users are co-located with each other. 

Robb Boyd:                   Ah, interesting. Okay.

Bill Thompson:              I think the biggest thing we have to think about, at least as network engineers or network administrators is gaining an understanding of where those users are, what applications are important to them and where those applications reside. It will be a constantly shifting, almost like a kaleidoscope of how that looks. But we have to get our heads wrapped around that so that we can design these architectures appropriately so it's optimized for the best quality experience we can provide.

Robb Boyd:                   Do you feel like some of these architectures are starting... Before I take this maybe off course, what were you going to say, Neil? Sorry. 

Neil Anderson:              I was just going to say, I think the perfect example of that is when you look at 5G and Wi-Fi 6 and Wi-Fi. The way we used to build wireless architectures was we would put radios out there, either cellular radios or Wi-Fi radios and we'd bring all that traffic back to a central place. And Wi-Fi it's the Wi-Fi controller and LTE it's the mobile packet core. They're having to rethink both of those architectures, because you physically just cannot bring all that traffic back to a central place anymore. So you are seeing that disruption happening in 5G networks and Wi-Fi 6, you're seeing distributed controller models embedded in the network, because you simply cannot think about the traffic patterns the way that they used to be. So I think that's a great example of just what Bill's talking about, having to bring those services closer to the user, is really the only way to architect it now. 

Robb Boyd:                   Do you think, I feel like sometimes one of the things that's been interesting to grapple with as individuals, whether it's on the consumer side or in the business side, because we've often seen is changes coming so hard and fast on the consumer side, so to speak, have really driven a change in expectations for what you get when you're on your IT network and a frustration with the fact that the network maybe isn't as fast at the office like it used to be. It used to be you go to the office because it was going to be faster and more reliable. Then all of a sudden, it seemed like it was faster at home, but there's all kinds of things happening behind the scenes. And it feels like also, we've got this level of customization as individuals these days where the way I see an app, you and I could be using the same applications, but they may look dramatically different in terms of how we interact with it.

                                    It feels like what you're saying here in terms of evolution in network architecture, that there's going to be some type of a customization, Bill, as you talk about the application almost coexisting with the user. Almost like I could have a different, not that I would feel it, but I would have a different architectural experience maybe than you would in the future. Because that's just what's needed based on how my usage patterns are being monitored and being responding to it. So we have that fabric now responding more directly. Maybe a little bit squishy, but is that what it feels like we're going towards?

Bill Thompson:              From my opinion, yes. I mean, over the years if you looked at how we've done the network in campus, it's been we just constantly build bigger pipes. We went from, what was it? 10 Meg to Fast Ethernet, to Gig Ethernet now we're on 10 Gig. All those different things. And I think those days, I don't think we'll be able to scale to the next generation of bandwidth availability cost effectively. 

Robb Boyd:                   That's a good point.

Bill Thompson:              The best way we're going to be able to scale out is not only adding larger pipes, but also more pipes in there. Building the fabric if you will, so that we can go many directions to get to the applications we need. 

Robb Boyd:                   I want to go through, as we're starting to wrap things up here, you guys were sharing with me some of the things that need to be considered for success moving forward, what our audience should be looking at. We've really been talking a lot about network architecture. But let's start with that one is the first of four. So well, network architecture, we've talked about the need for that to evolve and a lot of other things. But another one on the point here, Neil, why don't you jump in on this one, is that we need to consider these changes around consumption models? What do you mean by that, in terms of what's the expectation we should be setting there?

Neil Anderson:              Yeah, I think it's no surprise that we're seeing, especially with SaaS services and other things. We're getting much, much more used to subscription based models. We're starting to see a lot more of the network being consumed that same way as subscription services, as a service, purely as a service, with managed services and other things. So I would say there's a transformation that's right in the middle of happening right now where our customers are just getting used to that model for network access. They're just not used to purchasing it that way, they've been used to permanent licenses on equipment and I'll replace it when it fails. But that model is really changing quite rapidly. 

                                    And I think that just benefits customers. Once they're more comfortable with it, they have to think of it that way. You may be spinning networks up in software now in different places around the globe, in CNF data centers and things like that, where you're not really visiting there and placing a physical RU in a rack, you're just spinning it up in software on compute somewhere. And it's going to be a lot more flexible, I think to think about it in terms of subscription licensing, what am I consuming right now, versus physical hardware?

Robb Boyd:                   Interesting. Well, that may be a good way to just go into the next one, and the third one that you mentioned. So again, our network architecture evolving, consumption model is evolving, but you also mentioned IT technology. It sounds like it plays right into your last point, but how is IT technology itself evolving in response to these big macro trends we first started off talking about?

Neil Anderson:              Yeah, so I'll kick it off and Bill, jump in. I think we mentioned that IT is looking at this as, okay, I can't just uplift and replace one for one, router for switch and go on about my day. I got to think about more automation, I got to think about having better telemetry. If you think about the architecture we're talking about here, it's no longer a nice tidy little telemetry, where I can do some pings on my network and figure out what's going on. I'm going over multiple networks and multiple provider networks and to many different cloud providers, many different SaaS providers. And so that telemetry becomes super, super important to get the visibility to understand, what's going on? Is that my network? Is it my provider's network? Is it the SAS provider or my cloud provider? Where is the slowdown in the application happening? That's going to be a super, super important piece going forward.

Robb Boyd:                   Yeah. I think and the telemetry, I'm glad you brought that up. Because I feel like there's never been more tools. I feel like there's been an incredible growth, where we're way past SNMP Traps, and things like this, where we're finding information out every five to 10 minutes of what happened and it's been pre-aggregated for us. And we're guessing at what that means, to where devices can actually tell us this is what's happened. There's a wealth of data, because now we can handle a whole lot more data. But with that, we need to understand well, what is the network seeing in the context that's going to come from many different places and have to be correlated. And then we have application tools that are going to tell us how the applications are responding, so that we don't just say, "It's not the network." We go, "It wasn't the network. But let me show you what was happening in this specific instance with this application, how it was responding, and what caused maybe that issue that we can't replicate, but yet we still need to resolve."

                                    Because these days, I feel like to your continuity point earlier, we're building networks that are very stable. And performance may differ quite a bit, but they're stable. Things can fail and users won't necessarily know it. But it's incumbent upon us as operators, I feel like, I've put myself into that category, where we need to be preparing for, well, why did that fail? And how do I use that? Before the failure perhaps even predict or know that there are conditions that are leading that are trending towards something I need to be more aware of? Anything there you guys disagree with or would add more color to?

Bill Thompson:              No, I'd add, what the telemetry gives us today is that understanding that we really don't have. Where are those users connecting to what apps and what is really important? But even more so important is what is their experience with that, that's going to influence how we design the different networks in the future? And secondly, how we automate to auto magically, call it, fix problems as they arise? Lessen the dependence on a human to actually go in and understand these things and let machines start doing it.

Robb Boyd:                   Yeah, that's a good point. I always think about the trust we need to have in our systems to do that. Because I think it's hard for any of us that are used to knowing everything we did, because we built it ourselves. We know how it supposed to operate, we know how it will fail to a certain extent, at least we used to. But now it feels like things have gotten way too dynamic and it's not something that we can insert a human in the middle of the process and still expect to keep up. That becomes an instant bottleneck if it's even attempted. Take log reading, which coming from my security background, that was something we never did, who wants to read all those logs and lining up timestamps trying to figure out what happened? We need machines that are going to be able to do it. And so coexisting then is you've got the automation piece that's on top of this increased visibility in the streaming telemetry and that type of thing to make these decisions and let as much of this stuff be handled by the system as possible as we build trust in these things.

                                    The final thing you guys were mentioning in terms of success moving forward, we must evolve in terms of IT services models. Can you tell me what you mean by them?

Neil Anderson:              Yeah, and I think some of that is tied into a little bit of the consumption models we were talking about. I need to make some decisions about what am I going to own or operate? What am I going to consume as a managed service, or completely as a service? And then the things that I'm going to operate, how am I going to offer that in a way that's going to benefit a monthly user community? And it's going to get more infinitely complex for the IT staff, as we talked about the way that these networks are built. And you mentioned tools, there's so many tools out there. We have a joke internally that if you have the words analytics and AI in a business proposal, you can get VC money.

Robb Boyd:                   Ding, ding. Ding. Yeah.

Neil Anderson:              It's just that we have dozens of companies that come at us every week with a new tool, a new tool, a new tool, please look at us. We've picked some that we think are really good tools, that stand out from the crowd. And those are the ones that we go to market with, that we think are going to really help to solve and address some of these problems. But I don't think anybody really has the silver bullet for this quite yet. But there are some that we think are really moving in the right direction. 

Robb Boyd:                   It's just interesting how all this stuff really points back at itself, so to speak, when you talk about these macro trends and the need for change. Because I think about the fact that ML or let's say machine learning is a subset of your artificial intelligence, and you talk about, or we're pushing with 5G and the fact that there's applications that are demanding that the machine learning that's happening has to be so responsive and has to be out at the edge, because as you mentioned, it can't go all the way back through a core to get an answer, be processed and then come back. By that point, someone's cars has already run off the road, the surgeon has already pulled the wrong organ. I'm not a surgeon. Something bad's happened. There's blood everywhere. Let's paint the big negative picture. 

                                    Okay. So let's assume, as we wrap this up here then, both you gentlemen have done a lot. You work for an incredible organization. I've always enjoyed working with World Wide Technology over the years because of the fact that you guys are very so solutions oriented. And I imagine that as an organization, you have to constantly, as you're looking at your customers adapting or your customers look to you for advice on where to go next, that you're constantly having to reinvent yourselves in terms of the services that you provide and these type of things. 

                                    Assuming because, anybody that's listening to this in our audience is going, "Wait a second, I don't feel fully prepared to take that next step based on these things that are happening, whether they were recognized or not." I'm sure that's a common refrain. What type of resources does World Wide Technology provide for someone who needs maybe a little bit of hand holding for lack of a better way to put it in terms of taking that next step?

Neil Anderson:              Yeah, we really do have a lot to offer. Everything from workshops upfront, where we will get in a room and whiteboard with a customer, really understand what are your challenges? Where are your applications? Where are your users? Let's think about how your architecture is today. And then what's the end state you want to be. And then we have, obviously a lot of deployment services to help them get there and augment staff at IT, which is a big one right now. We're seeing a lot of interest in our staff augmentation services, just, "Hey, I need some extra hands right now to get me through this," that really understand the technology. And then I would say, one of the things that Bill's team works on constantly are our customer facing labs. So Bill, you want to describe what we offer there.

Bill Thompson:              Yeah, sure. So once you get to that point, figure out what that architecture or what that automation or those services look like, we have a ton of different hands-on labs that you can go in and educate yourself at no cost, by the way, on wwt.com. These are things educating about features that are part of say an SD-WAN or how do I implement security on end user devices? And even going down to the basics on how do I start with automation? What are some of the tools that are very commonly used across the industry? It's a great resource to go to, to start learning how to get to that next level and the future of networking and automation and all the things we just talked about. 

Robb Boyd:                   Oh, that's perfect.

Neil Anderson:              Pretty unique also.

Robb Boyd:                   Yeah.

Neil Anderson:              Like Bill said, it's no cost, customers can simply come and register for an account and you have access to all this whether you're a WWT customer or not, it's pretty unique and I don't know of anything else like it out there. 

Robb Boyd:                   I don't either. And so I was just going to put this on as a button here at the end. But first I want to say thank you guys, I appreciate you taking the time to talk through these issues and raise some of the stuff we need to be thinking about in terms of our evolution from a networking perspective. And of course, I want to encourage our audience, it's wwt.com, it's not hard to remember. And indeed, these resources, it always surprises me when they remind me that there's so much of this that is free, and it's about education and learning. And that's probably my favorite reasons for getting engaged. 

                                    And I do encourage you, it's free to sign up for an account. These guys have built their ability to interact with you in a way so that you can be notified, you can follow certain subjects, you can get involved. A lot of the intelligent people at World Wide are actually writing and sharing information on a regular basis and encouraging interaction. And so there's a lot of ways to get involved that don't mean, "Oh, gosh, I need to pay these guys or something." It's really not about that. It's very solutions oriented and they continue to invest in that which is amazing, because back when we were traveling a lot, I spent some time out wandering through and I know it's completely changed since then. 

                                    What they call the Advanced Technology Center, the ATC, that everything you can do at the ATC physically in St. Louis is available worldwide through the web because they've made a conscious effort to reach out. And that is fantastic for going deeper and deeper in these labs where you want to get hands-on and work with some of these smart individuals. Because I've worked with the cloud team, I've worked with the data center team, your service provider resources, it goes on and on and on. I'm just constantly amazed with the people that are there. But either way, thank you guys so much. Appreciate you sharing this. And thank you guys for watching. This is the TEC37 podcast, your source for technology, education and collaboration from World Wide Technology. My name is Robb Boyd, we'll see you on the next one.