Welcome to this special Telemetry Now where we kick off the first episode of our customer series. And in this episode, featuring Joe DePalo from Netskope, a veteran of the networking industry and specifically in the service provider and CDN spaces. We're also joined by our very own Avi Friedman, founder and CEO at Kentik, and Jezzibell Gilmore, Kentik's VP and general manager of our service provider business. Now in this episode, we're gonna discuss one of the unsung heroes of the Internet and specifically how we can consume content today, content delivery networks or CDNs.
We're gonna dig into some of the history, technology, and the future of CDNs, especially as we move into 2025. My name is Philip Gervasi, and this is Telemetry Now.
Avi, Jezzibell, Joe, it is so great to have all of you on the podcast today. This is a special treat. It's usually just me and some other person nerding out for forty five minutes or an hour, often about something that I know nothing about. So it's a great learning experience, and I really hope that today that is the case as well. Now before we really get started, I heard through the grapevine that the Internet is just a gigantic black box, that it's a magical back black box that we all connect into, and that's how we get our content.
And, being a father of children, of several children upstairs who don't care about technology, they just wanna know that it works, that they can get to their YouTube videos, and they can get to their well, now it's YouTube shorts. I guess that's the thing for kids. I don't know.
But but, honestly, that that's how it is for so many people, and that's one of the things that I wanna broach with you today. But before we do, I happen to know that, Jezzibell knows you both, Avi and Joe, for many, many years. I won't divulge how many, of course. So I'd like to hear about that. Yeah. I'd like to hear about that. And, of course, Jezzibell, if you have any silly stories that would embarrass Avi and Joe, of course, I wanna hear those first.
Oh, first of all, I I have to say Avi gave Avi, gave me the opportunity to be in this industry.
So, you know, everything that I have done here, started with Avi. So, Avi, I appreciate that, and I hope that I'm here to make you proud.
But also that, you know, I want to introduce Avi and Joe both as Internet luminaries.
They are the people who help build the Internet as we know it today. I'll start with Avi. Right? Avi started one of well, the first ISP in the Philadelphia area. And he, since then, have been instrumental in building large scale networks, such as Abovenet Communications, Akamai, being their chief network architect, and chief scientist. And then from there, he's gone on and built the ServerCentral/Deft cloud solutions and network.
And then, of course, he's here as the CEO and founder of Kentik, building a solution that the industry absolutely needs, to solve the sophisticated and complex problems that we are facing as network operators today.
And something that you don't know, about Avi or most people don't know about Avi Oh.
It's not even embarrassing.
Actually, this is something that's very fun. Avi is the only nerd that I know that has appeared on ESPN before.
I think you were in the audience for that, weren't you?
Were were you were you in the audience for that? I forget.
I was. I was, and I was very fortunate, but I wasn't actually having the camera focusing on me.
You know, so that the the focus was on Avi multiple times, and it was, so just to give you the background, Avi, not just an incredible technologist and a leader in technology, but he is a, winning poker player in the World Series of Series of Poker. And so that shows that good decision making doesn't just stick to business, and he leveraged that. He practices that in his poker playing, and so he's a winner, in a totally different arena there. And let me come to Joe because Joe's actually not any shabbier than that.
So obvious obviously, with a ton of shining glories. But, not only that Joe started in the Internet industry in building the hosting industry, which is now what we call cloud solutions. Right? So he started on the other end.
But he had participated in the founding of one of the largest CDNs and then gone on to AWS and build their Internet service business with them before he landed at Netskope in really combining networking and security into a single solution that had not been done in the market before.
And that enables the enterprise customers to have a single sophisticated solution in solving their business objectives rather than piecing together different types of technology to come up with something to meet their requirements.
So I will say something about Joe that nobody knows. I actually have several things I can say about Joe that nobody knows, but I will give a personal story. How Joe and I met, many years ago was through a very challenging conflict where my responsibility was to ban Joe and his company from attending a very important industry conference for a year, for violating some of the regulations.
Nothing too bad. It was just a conflict of interest.
And so Joe was able to turn that very difficult situation around into a lifelong friendship with me, and we've been friends since then. And I think that shows, as a leader, Joe has always had the ability to turn a challenging situation around into something that is a win win solution for everybody. So well, there's and, you know, if you know Joe, you know that he's got a team that follows him around everywhere just like Avi. So they're both incredible leaders, really caring people and incredible technologists.
So, but, you know, all the fun aside, I am going to I'm super lucky. Kentik is super lucky to have both of them here today to talk about the future of the Internet and from their perspective, how and where that we need to go. But before we get started, I have one last thing that I want to ask Joe because it's very rare that we get to have someone like Joe as a customer of Kentik to give a little bit of input on, his experience leveraging the platform. So, Joe, do you mind, talk about just a little bit of your experience?
Yeah. Sure. So I've I've been, a fan of all the work of Avi and of Jezzibell throughout the years, and we partnered in a lot of different aspects of business. And the most recent is with Kentik.
I was a customer at Amazon and, at AWS, and we're a customer now at Netskope.
You know, our global platform requires the visibility and the the data gathering that Kentik gives us, the the streamline interface. It's a big part of our network. We do a lot of route control and a lot of tuning performance management on our global infrastructure, and Kentik is a big part of that for us. And as we work with enterprises, we are strong referral for Kentik and these enterprise networks that are growing more and more, demanding of use cases like, you know, flow data and routing and running their own BGP and visibility into that. And so, like Phil said, in in the earlier part of the call, the Internet is is not really a place or a thing. It is a collection of different networks. And having that visibility, is your responsibility as a network operator, and, you know, Kentik has been that tool for us for and for me for for many years.
And I'll just, say something, to piggyback on Jezzibell. When when Joe and I met, we were competitors, but I think it's always possible, especially in the Internet, which as Joe said, is a network of networks.
You know, it it's always possible to build relationships, and sometimes even competitors need to cooperate or coopitate or whatever, however you would say it. And then, of course, if you if you're gonna be in the industry for decades, you know, those relationships will will help and sometimes are actually critical even to making the Internet, work and fighting, protecting our customers and and growing. So, I love the industry, you know, for that because that's generally the attitude that I found.
Yeah. The network operators have have a role, I think, maybe even a little bit bigger than the loyalty to their company and the fact that if the Internet doesn't work together, like, even at Amazon, we peered with other cloud providers. Like, there is the notion of connecting and working together and, for the collective. A lot of the IETF, a lot of the industry standards, you know, have to come from people with competing business factors, but also, you know, the underlying infrastructure is important that we can all use it. So, yeah, that's a good point, Avi.
Yeah. We are talking about a network of networks that no one really owns, and therein lies some of the problem with operating this, especially at at large scale. I get that. So, again, you know, being a little bit silly about how long you've been in the industry, the reality is that that's very valuable in this conversation because you have that wealth and breadth of of experience and knowledge. So, you know, my my kids, again, don't care about how it works. They just want, access to the content to the applications.
And I know, having been a network engineer, that it's always a process of reducing latency. It's a it's a process of lowering round trip times. And, really, how do I deliver content well over this big system of systems that's ultimately an application delivery mechanism, right, or a content delivery mechanism.
And that's that's one of the hard problems that I want to dive into right now. You mentioned CDNs several times. I'd like to level set what are CDNs, what problem do they solve, and, considering your background, why was that, the solution to solve this content delivery problem?
And, Avi, if you don't mind, I'd like to start with you.
Sure. So, I'll actually flip back with a question to the audience.
What was the first, we'll say, commercially successful or commercially used CDN?
Anyone know?
Digital island.
One could argue digital island. I might argue Sandpiper, but didn't they become sort of the same thing? But yeah.
Mirror image.
Yeah. Or well, no. No. Mirror we call them Mirage or Mirror Image, but no. Yes.
I would say Sandpiper in a distributed sense. Digital Island bought yeah. So Digital Island, I think, bought sand bought Sam, bought Sandpiper.
And, you know, actually, I would argue there's even another one. Anyone remember when we used to get shareware from?
Tucows.
What did Tucows do? They're still around.
Two I know, but they've shut down the distribution network. They've actually got some they've they've some of the insight, I don't think it was Netflix was inspired by them, but they did some of the same things as Netflix. Because we think of Netflix as a CDN, but it's really an rsync mechanism for the nerds. It's not really like caching and caching. But, you know, Tucals would get people to buy the server, put them in their own networks, replicate a shareware base of data, have a central database for it with some ads. I don't even know whether they did, whether they did rev share. And then users would get the access to the content inside their network.
I call these things almost or CDN's magic packet transporters because Internet people think of BGP and ASs and prefixes of stuff we can talk about. But Yep. This is with DNS or HTTP being an overlay on top of the network living at the are you ready? The edge of the network bypassing what are often bottlenecks between networks that have limited capacity. So content closer to user, content's faster, more scale at the edge. That's the idea of CDNs, and it was done simply with Tucows.
It was done a little bit just for objects, not for all sites with this thing called HTTP redirects, which is much slower by Sandpiper.
Akamai came about and was one of the first, in fact, you know, maybe the only to do intelligent DNS for quite some time. Everyone else did this Internet thing called Anycast. But content at the edge for objects, then sites, then dynamic content, And then it turns out, which Joe can I'll let Joe talk more about maybe the security evolution. Once you're in that position, there's a lot you can do besides make it faster to give visibility analytics, to add security.
So it's it's I call it magic packet transporters, overlay networks, networks that are not quite networks. When I hired Andrew Ku, who does a bill knows also, he said, oh, no. I'm I'm sorry. It was Rob Seastream who said, oh, you're the world's largest nonnetwork when I was at Akamai at the time. Yes. Exactly. So that's my summary of what a CDN is.
So, for me, the the CDN so if you if you think about it from the early days, CDN solved the problem.
Web hosting solved the problem. Cloud solved the problem. So there was it was data, right, in in the early days of the Internet, and then it became GIFs and images, and then it became videos, then it became live content. And so a CDN evolved.
Right? It became well, we have a middle mile problem with carriers. We have an interconnect problem with carriers. So let's get the content as close as we can to the end users.
Right? And, and so the CDN was was born to distribute the content in regions where you're in a single network or not a network at all, and you're the the closer you can get the content to the end user, the better the problem is. Because everybody has plenty of connectivity leaving their home, but once they get into the middle, it's where it gets congested. And then you you evolve that into, teaming the cloud space.
Right? Amazon, AWS was born because there was no place to put Amazon dot com, and they developed cloud. And someday, Avi, you'll have to tell me why Akamai isn't the first cloud provider, but maybe that's for another podcast. And then today, the, the evolution of the security piece now where, the capacity, the throughput, the reach, and the lack of checks and balances online, there's now we're going through the another evolution of technology that needs to be created to protect users and enterprises.
And so to me, these things happen through, a problem. You get a lot of smart people together and and poof an industry is born. So that's how CDN came for me.
Okay. I mean, we've described this problem from a high level, though. What was wrong with the Internet at the time years and years ago that required this? I mean, you're talking about moving the content to the edge.
And in my experience, it's so you can have a more performant application delivery. So you don't have the buffering and all that stuff, especially for streaming. What was wrong with the Internet? Was it just bandwidth or or lack of bandwidth, I should say?
Yeah. I mean, the as Joe said, the middle mile problem is think of the Internet as a network of networks. They don't all connect to each other, and that's called peering when networks connect to each other. And even if they do, there's all sorts of I don't know.
Do we call it layer zero or layer eight politics and economic incentives that, that limit as well as physical issues. Like, there may not be more capacity on my network in that city to connect to this other network, but that's where the users are. And, so as a result, people make decisions often based on cost and not performance. And if your peers making decisions that way, you have some ability to route around it, but, ultimately, you know, there still may be limited bandwidth between you and them.
The routing technologies, even today, give you limited ability to be flexible about at what times you move content, at what, you know, what types of content can go where, video versus web to the to routers. It's all just packets with IP addresses and port numbers. Right. And so the ability to have this kind of overlay that you can that you can map and route and, again, like, preserve some of that internetwork bandwidth by keeping it local doing caching.
Right? If I've pulled it once, keep it there, pull it again. This is something that people were selling boxes for. And, for example, universities were doing because they were getting slammed on their transit.
I mean, to give you an idea of the cost of bandwidth today, I don't know, Joe. What would you say? Two cents per megabit per second?
Yeah. Yeah. Pennies. Yeah.
Yeah. And what was it, you know, for for you could think of this kilowatts per fortnight or bananas per year or whatever. I mean, just as a unit, you know, back in nineteen ninety even when I joined Akamai, probably ninety nine, you know, four hundred dollars a megabit a second was okay.
Hundreds.
Yeah. Yeah.
Yeah. Yeah. And it was thousands before. It was a thousand for fractional t one. So caching was something that providers used, but even, you know, they didn't really have an understanding of the site.
The content providers lost visibility. They didn't like that. So, you know, the networks themselves are trying to solve this problem even before the content providers, you know, were offered CDNs that helped solve this. So it was cost and capacity.
And, again, because, you know, if the four of us are networks, well, I like Jezzibell, but I don't like Phil, and I have to connect to you. So maybe I buy from Joe to get to Phil, but then Phil fills up his pipe to Joe and you know? It's just the way Internet routing works. It's you you mentioned it, Philip.
Like, the Internet, is so important. It's a black box. But the more important the the more profound thing, the more you study the Internet is not that it breaks or gets congested sometimes, but that it never works in the first place for this, what we're connected on. Right?
I mean, it's it's a bunch of best efforts, decentralized interest cooperating to make a reliable service that is pretty reliable now.
With a layer eight issue wrapping the entire thing, I mean, it's built on trust. I mean, I know there's some mechanisms in BGP and things like that to secure things. Certainly built on trust, but we can't really, not really, solve layer eight with technology, though.
Like, we can't we can't fix, with some algorithm, you know, people having disputes and Avi not liking Phil, and therefore, there are problems there. So, you know, that that to me suggests that we still need CDNs, that we haven't solved the problem just because bandwidth has become cheaper and, you know, speeds and feeds have become so much greater. There there are still those bottlenecks from a technical perspective and then the bottleneck of, you know, people being involved with this.
Just as an example, I know, from working at Kentik, some of the, blog posts and analytics that, Doug Madory highlights after, like, a big football game. And so, you know, I just one of the points in our note notes here that I've been itching to get to for this entire fifteen, twenty minutes we've been discussing are some of the lessons that we've learned just in recent days on streaming NFL games or any other examples that you want. You know, there was the, the the boxing match just a few weeks ago as well that was all over the news, making it making it into mainstream media. Yeah.
So it you know, Avi touched on it a little bit.
It does come down to business and money as well. And so if you pay a hundred dollars a month for your Internet access and you pay twenty dollars a month for your Netflix, who is incentivized to to give you more capacity?
Right? So the reality is there's no more money. I I I have this thing that I try to explain to people is that there's no money in the middle. Right?
There's no money transporting this infrastructure. Like, the the reality was if the Internet was built, today, your home Internet would be metered. Right? There's a reason why cell phones are adding more and more capacity like five g and growing because your cell phone bill's going up.
Right? My your Internet bill is not. Right? Or your subscription service is not. And so Netflix tried to get providers to give them capacity.
I was the CDN for Netflix at the time, and the provider said, we don't have any more to give you. You're you're exceeding the middle. You need to share some of your revenue, right, in net neutrality. And so it became a problem where Netflix created open caching to move their servers into the networks and save on that middle mile.
And so your CDN problem is a business problem. And so, like I said earlier, you create technology solutions around these limitations, your layer eight example. And so that that is unfortunate that the the there's the evolution of the businesses faster than the technology.
But because we got smart people working on it, it, it catches up and eventually works around it. And so you talk about the boxing match, or the NFL games.
When I was at Amazon, we we started that NFL contract, and you had to explain to a lot of business people why the Internet capacity, the Internet reliability, the protocols aren't performant.
You're not getting the same level of of scale that you would with, you know, your home, cable networks or, you know, even, even infrastructure. And the the best example I give is we're old enough to have landlines. Those never went down. Your cell phone drops calls all the time, so you just pick up and call another.
Right? Whereas it's just a a different scale of infrastructure. And so it's a combination of business growing too fast, the consumers, and then very limited underlying technology that needs to adapt. And so the CDN was the first example of that of that adoption.
Yeah. I'm but it's amazing.
It works at all still at the scale that it does.
Believe me. I know.
Compared to ten. Yeah.
A hundred million people watching a boxing match. No way.
It was the at home.
We had a call with them at Akamai because what was the biggest event on the Internet back in the late nineties was the Victoria's Secret event.
And we congested basically every pure and transit provider that At Home had, and they said, yes. That's on purpose. You need to pay to be in our network.
And, again, we're seeing some of that not quite the same way, but in Europe, there's a lot of telcos that say, even just that last mile. As you were saying, Joe, you know, the the cellular, you know, upgrades, like, that should be partially funded by the content providers. So this has always been a war of their aid of influence and settlements and all that. But the technology of edge distribution is important.
And I I I think you hinted at it, Joe, earlier, but this is something we see with the enterprise. Like, a long time ago, peering was about cost. As you mentioned, the cost were very expensive. Now it's about taking control of your packet destiny.
If the revenue of your business is your applications and you just rely on the two people you connect to to get them to the other end and you're having revenue impact, that's a big problem. If you don't have an office and your your people can't work and interact because they can't get to Office 365 or or because they can't, you know, work with each other, that's the revenue of your business. So the enterprise is actually getting a little bit more intimate and familiar, at least with the problem set, if not manually manipulating all these things.
And, you know, it's been interesting interesting to see, and I think we're both, you know, the industry is helping people along with services. I still laugh when people say network as a service. Like, it was not a service before. What?
Right. Yeah. Right. Still you know?
But, yeah, it's a much richer toolset than you had in the nineties for the enterprise to try to tackle these problems.
Yeah. So I have to I have to do my plug. But for we learned that early days of Netskope was that we wanted to take control of our destiny and not build it in a public cloud and have access to the different networks and peers. Because if you're relying on I think you said black box, we call it gray network.
Right? Every provider has a gray network cloud for the Internet. And so if you're relying on another infrastructure to deliver your service, you're you're at the mercy, and we all know how limited that underlying technology is. So Netskope, we specifically decided that we wanted to build our own edge sites, have our own path peering connectivity, our own route, technology over the top of it.
Because if you're relying on public Internet or a carrier or a cloud, you're at the mercy of of that provider. You look at someone like GCP, they move a lot of bits with YouTube and mail and other services, and there's a cloud mixed in there somewhere. And so you're gonna have to take control of your destiny. In an enterprise, we see more and more is they used to just pay a carrier like AT and T for managed Internet, which is basically just just default.
And and they're they're struggling with application performance and and, and reach and and especially other markets. There are no more global carriers. And so as enterprises, they're gonna have to take some control of their own destiny as well. And that's where I think part of these podcasts, hopefully, we give a little bit of insight for the enterprises to be able to control their destiny because you can't rely on anybody on the Internet without at least some level of control or visibility, whether it's through Kentik or Netskope or your your own service.
I was gonna say, I'm gonna throw in a little wrench that as the geopolitics continue to evolve, you know, data sovereignty and security and cybersecurity is becoming increasingly important. And that obviously applies to both service providers and enterprises in, you know, between Kentik and Netskope. I think that those services, you know, that we're solving some of these critical challenges that the end service providers and enterprises are facing. And I think the CDN, and, you know, having the visibility and having that security really helps. So, you know, what how do you see that impacting?
Yeah. That's a great question. And the one thing the the one area people miss with data sovereignty is they they assume storage.
And so they worry about where their data lives and stores, but they don't understand how it transports, Right? Where it's cashed, maybe, where it's processed. And so those are lots of components. If you have sensitive data, you don't wanna leave your country or your company, but you're transiting out a third party carrier who's backhauling it to another location.
There's a lot of complexity there. Is somebody caching it? If they're processing, are they writing it to disk? And so, there's a lot that goes into it. Again, someone like like a Kentik to provide visibility, a provider like Netskope that can also do inspection.
But data sovereignty is more than just where is it stored or where is it kept or where does it sleep. And there's a lot of other factors that the enterprises is learning the hard way, as these things happen. You know? And so, unfortunately, people view the Internet as geographically dispersed, right, by country or city, but it's not.
Like, Avi said, it's AS based. And so, the if you ever Google a picture of the Internet, it's just a big glob of looks like a brain. Right? Like a a neuropathways.
And so being able to track where your data transports, who's looking at it, where it's cached, where it's written, where it's managed is, is a huge, weakness of enterprises that I've run across in the last few years.
Yeah. For us, it affects us internally one way, which is that some of our service provider customers have bans from their governments and sending their telemetry, like, outside. Sometimes even, you know, multiple conflicting bans in terms of they're an entity that's in multiple countries. They can't send data outside.
But, so, you know, it forces us sometimes to run copies of Kentik for people to run inside their own premises, which we've made work, even though, you know, it's not particularly the way they wanna they wanna deploy it or or or we want to, but it's a software we could make that work. But on top of it, it's really something that is critical as a primitive for the right insights or and analytics is to be able to understand your telemetry. This is a telemetry podcast as well in terms of the service and customer. And if you're an enterprise, maybe division group, etcetera, that is generating that, because to do as Joe was saying, to understand, like, is this customer comma application maybe even that level of of of of analysis and be able to say, hey.
You know, there's been a routing change underneath, and this is now traversing a path that it shouldn't. In the old world, it was just performance, I will say. Just performance. Right?
It was like, why is my South Africa going to United States and back to get to, you know, Zimbabwe? And you're like, well, we're both legal or whatever. But, you know, for for data sovereignty, sometimes that's just not, you know, allowed. And then, you know, as people increasingly use cloud, it's almost equally important because you can get really outrageous cloud, you know, charges.
You know? I've got we have some enterprise customers that have terabits per location of cloud interconnectivity.
And and, again, they take control of their own packet destiny because of cost, but also because they have restrictions on where they're allowed to send that data into and out from. So it's just it's more and more, you know, complexity that the network folks have to deal with. And, you know, as one of our customers said, network is the water of the clouds. It's what clouds are made of. If there's no network, there's no cloud.
So, you know, lots of constraints and and and solutions that people need help with.
The nature of application delivery, you know, lent itself to CDNs and, focusing more on the edge, both in terms of the actual delivery application, the app and of the content, but also in sec in terms of security.
So it has is that now a solved problem, or is there still a performance thing that we're looking at, solving? Is there a security problem that we're still trying to solve? Where where do you see us going next?
Yeah. That's a that's a great question. And so what's happening is that with five g, with fiber to the home, with, as a result of COVID, less enterprises in central locations, I don't even have an office anywhere anymore for the first time in my career. You're just kind of roaming the Internet. The edge, as Avi says, has dispersed even more. So you now have more throughput than you've ever had. You have more dispersion than you've ever had.
And and we're using a firewall or using a centralized location is no longer. And so, again, like I said earlier, we have an evolution, period in the Internet where before we had storage with CDN, then we had compute, and now we have security as that that next, evolution, period where the the data is everywhere, the zero trust concept, and bring your own devices. And it's, it's very much the wild west, and I don't wanna create too much FUD, but the the bad guys, the bad actors are way ahead of enterprises and and the providers, because of that dispersion of infrastructure. And, you know, when everything backhauls to your office and you run it through a big old firewall, you have a a private IP network at, you know, RFC nineteen eighteen, you're in good shape.
But when you have your users that have a one gig to the Internet from a Starbucks that has access to just about anything, you're gonna have some, some pretty scary problems. And so the evolution, I don't think, is there's there's garnered coins SSE or SASE. I still think we're in the very early phases of those types of, implementations, the integration of security and network. And so now is the time as the network enterprise operator, is to take control of your destiny and to make sure that you're aware of layer one, layer two, layer three, layer four, of what's happening, where the packets are going because that's that's the first step in in securing your workflow, securing your users, and protecting your networks.
Yeah. I think that we've we've seen writing about, and I think we've done some blogs about also some simple stats that that paint the picture. But for some of our broadband customers, north of ninety percent of the content comes from CDNs, and north of two thirds of that comes from parts of those CDNs that are embedded in the networks themselves. Sometimes more than that. We call embedded caches again even though I don't think Netflix is actually a cache. But, and I don't think that's gonna accelerate to ninety nine percent, but that there's nothing showing that that's going to change.
And I think that the, from an enterprise perspective, I joke about network as a service, but whether it's SASE, CASB, all these composed services are gonna be tools that enterprise can use, and then it's gonna be necessary from cost and control perspective to have some of this taken control.
But, you know, whether it's cloud abstractions that's more as a service or more as an abstraction layer to go between clouds, some amount of your own interconnectivity management, some amount of, you know, almost everybody using something to help cloud manage, access to endpoints or between campuses and branches instead of the old MPLS technology. It's gonna be a set of services that you use to manage your network, which is a presence on the Internet, and make sure that the high value things that, represent the revenue business, you know, are protected. So I think a lot more the same. The only good news is, people are a little bit less worried about, you know, why aren't we all I p v six.
One of the great, people that's been tracking that finally said, yeah. I guess with Nat and and and these overlay kinds of things, maybe people one less stressor. Because a lot of people, I think, have been overly concerned for the last ten years. Like, why aren't we all IPV6 yet?
And it's like, it's fine. The network works. It's gonna continue to work. It'll continue to be both.
It's fine. So don't don't at least worry about that. We have enough on the security front, you know, to worry about.
So The the surprising thing, Avi, is that I don't think the enterprise people understand that the visibility you're talking about or the control I'm talking about is actually cheaper than what they're doing today.
And so getting that visibility, controlling your destiny, you know, managing your packets, is is much more, affordable than it used to be and is even more more affordable than your overlays, your MPLS, your SD WAN, your managed service providers. And so it requires some level of technical understanding, but the control you get, is actually much more affordable. I was working with a customer that was, you know, getting a ten gig wave. And even five years ago, that thing would have cost a hundred grand a month, and now it was three thousand dollars. You know? And so whether it's fiber, whether it's metro fiber, whether it's dark fiber, whether it's transit or, transport services, the the cost is much more manageable and well within the budget of an enterprise.
But then you get the benefit of all the things we're talking about, the control, the capability. And so there there's a little bit of both.
Yeah. For us, we saw three years ago a big driver for I'll go with enterprise side of our business, which was they were sort of struggling with and maybe spending a little bit more on capacity and, you know, poking at old Windows and tools and federated appliances. And it was really when they needed to take control of the cloud that a lot of folks have said, okay. And they've started seeing cloud bills.
Right? It's like, oh my god. This is not gonna work anymore. So how do we stop having eight different solutions for each part of our network and and get it all together?
And I think we're seeing the same thing on the control. Right? The governance and control and security substrates where it's all network, and it's all critical to make the business run. And, you know, the the the trick is for executives to understand it's not, you know, fire the network people and they go home.
It's you're giving them the tools to make more of an impact and make the business run better. And I I think we started seeing that in the web companies and service providers where they really always had an understanding that the package is the revenue and the enterprise cloud, you know, a few years ago really was for for in our journey, what started getting people realizing how important it was, to, magnify the network people, especially as it's gotten so much more that you need to understand at least a little bit of. Right? I mean, cloud networking is just other people's routing tables and firewalls and tunnels within your games and bugs.
But, you know, if your applications are running across it, you've got the Internet, you've got the cloud, you've got your own stuff, you got the LAN. You know?
And with the right base investment, yeah, it can be a lot cheaper for all the things that you need to do to make the business run.
As we conclude this episode, I have a last question, for both of you, actually.
I, you know, there had been distinct personas in the purchasing of digital services, whether it's network buyer or security buyer. But I think that there had always been sort of understanding that the security buyer, gets to make the decision.
But things seems to be very different, now, and things are shifting between, the security buyer and the network buyer and how that they have they're collaborating and they're working together.
Do you have some sort of take on that? How how are things working out for each of you from your perspective?
Yeah. That's that's something that I've learned over the last couple years.
Before the security team was compliance, they could tell a scary story about a bad guy and get funding and kind of leave it to be. The network person, as long as there was connectivity, and and no performance complaints, they could ignore the security. Well, now, we have seen more and more of this merging of the network person. You know, I I call it your chocolate's on my peanut butter and your peanut butter's on my chocolate.
The network person can no longer care, can no longer, excuse me. The network person can no longer ignore the security cape components and risk of their network, and the security person can no longer just pick anything and hope it gets deployed. And so the security person has to be, conscious of the performance, the impact, the workflow, and the network person. And so what I've seen is I've seen security roll under network.
I've seen network roll under security. I I've been on calls where I've introduced them to each other even at the the same company. And so as I said earlier, there is definitely this evolution, that's happening in the market with the next phase of network and security, and and more and more, the decisions, the buying, the authority is merging between the two. We don't know who's gonna win, but, they're gonna have to work together for sure.
I I think it's a journey that's been accelerating, but it's still maybe, at the fifty percent mark, in terms of the companies. Whereas you said, one's working with the other. They know each other. They're in touch. I think the the criticality of in protecting embedding these services, both reliable networking and inter networking services and the security substrate, government compliance governance compliance enforcement has driven a lot of that. I think another thing that that's driven it is the way we're writing applications even in the very traditional enterprise.
If you're dynamically orchestrating and provisioning things, that looks like an attack to a security person unless they have visibility and understanding of what those things are, unless you have telemetry analytics that understands this is an app this application customer, etcetera.
And if you're, you know, the network team and security has shut something down and you don't have visibility about it, you could run around for for literally days, try thinking it's an outage or someone blaming you, and your mean time to innocence is very high even though it's the same organization.
So I think there's been that understanding. At the same time, there's still cultural there's just language. Right? The language that people use to describe the same things can be very different, and you still got you know, well, there actually is no enterprise network conference anymore. Interop is no more. Cisco Live doesn't welcome half the industry.
You know, the mugs tend to be very Internet focused. You've got AutoCon, which is great, but, you know, you know, focus on that one segment of it. It's a it's an area we need to do better. And then RSA and, you know, Black Hat and, you know, are off of there.
Still a lot of work to be done, but, yeah, I'm I'm really encouraged by the the trend is super positive, whereas for twenty years, it was more balkanized. I mean, twenty years ago, we were all just nerd, and everyone did everything. It was a little simpler. But now that everything specializes, it's become a little bit more complex.
So, and I think that's gonna be really helpful as people understand how these converged services, you know, can help both teams and, you know, be better for the enterprise budget.
Great. Well, it it even twenty years later, Avi, I still feel like I am a nerd.
So, but, this has been an excellent episode. Thank you, Avi, Joe, and Jezzibell for joining me today. And for our audience, please stay tuned for more episodes with Avi, Joe, and Jezzibell where we're gonna discuss more topics around things like security, of course, AI. We can't not talk about AI.
And perhaps even how to take control of your IT operations, especially as we move into 2025. So for now, if you have an idea for an episode for Telemetry Now or you'd like to be a guest, I'd love to hear from you. Please reach out at telemetrynow@kentik.com. So for now, thanks so much for listening.
Bye bye.