Network Intelligence in the AI Era: Balancing Cost, Complexity, and Performance

Okay. Hello, everyone, and thank you very much for joining this webinar, Network Intelligence in the AI Era, is done in collaboration with Capacity Media and Kentik. My name is Amber Jackson, and I'm a senior reporter at Capacity Media, and I'll be hosting the webinar today. We're all very excited, to have you join us for what promises to be a timely and insightful discussion. So I'll I'll give a little bit of background to start. So AI, multi cloud, and data hungry apps are driving an explosion in network traffic and complexity. But the problem is that traditional monitoring tools weren't built for this era. So providers are under pressure to deliver performance, keep costs in check, and meet those rising customer expectations. So that is where next gen network intelligence comes in. In today's session, we'll spend an hour looking at how leading providers are using it to optimize, secure, and scale their networks. We'll cover what network intelligence really means in twenty twenty five, how to balance efficiency with performance, and how to turn network data into real business value. We'll also show you a live demo of Kentik's latest AI driven insights and cost intelligence tools. We'll have time for a fifteen minute q and a at the end, so please send questions in the questions tab throughout the webinar. And just a little bit of housekeeping, if you experience any technical difficulties throughout the hour, just refresh your window and this should correct it. Simple case of turning it off and on again. So I'll introduce our panel for today. We have Jezzibell Gilmore and Steve Meuse from Kentik. We also have Dave Siegel, CEO of NetMavens, and we have Steve Orlando, CEO of Seaborn Networks. Thank you all very much for joining, and I'm now gonna pass over to Jezzibell to kick things off. Over to you. Hello, everybody. Welcome. Thank you all for joining this afternoon to talk about network intelligence in the AI era. We're gonna be talking about balancing three, challenges that has always challenged all the network service providers, and that's cost, reliability slash complexity, and performance. So running a network is not for the faint of heart. No child I know, that has ever stood in front of his or her class on career day said, hey. I want to be a network operator when I grow up. Why is that? So number one, I think it's because that it's running a network. It's seriously not glamorous. Right? Nobody's ever won an Oscar for running a network. And, in fact, some of you actually compare running a network to plumbing. And, two, I think it is extraordinarily difficult, especially in the age of the rise of AI, multi cloud, the data hungry applications, which Amber just talked about. Right? The complexity of managing the network has grown exponentially. After decades of hardware platform evolution and the combination combination of ever changing protocols and standards, if we haven't made running a network hard enough already, we're just doing it more by applying AI. The network operators, the engineers, and the architects have been really intelligent, gritty, and resourceful people. They are the innovators and the pioneers, and they're not going to let anything stop them from making progress, and that had been the way, and that will continue to be the way. Over time, they've overcome so much and achieve so much. But, you know, if you could think about what it would have been like if we hadn't created the Internet, or even worse if we'd stopped making progress just because we didn't have the resources or the tools, to allow Internet to proliferate. So and that's the way that we've pushed through as network operators. But, let me talk about something that's a little more. It's not something you don't already know. Almost all successful technological endeavors were achieved through the assistance of high quality instruments. Can you imagine flying an airplane without instruments? Well, I wouldn't either. So, why would you imagine, operating a network without one? Because operating a network is just as complex, if not more. So with that said, since the networking is the foundation of all the craze in digital AI, in digital infra and AI today, do you really want to run your network without it? And how are you going to have the most effective tools to ensure the optimal operational efficiency of your network? Today, we have invited invited a couple of prestigious guests to be here with me, to talk about how they balance the difficult act between reliability, performance, and cost in network operations. So let me introduce them, and then I would like to ask them to talk a little bit about their companies. So first, Dave Siegel, who is the CEO and founder of NetMavens. He's got over thirty years of experience in building, operating, and marketing, as well as selling Internet networks. His expertise is wide and deep, impuring network architecture and design, OSS designs and implementation, as well as product and product strategy. You know, as if that's not overachieving enough, he's got multiple patents. And, you may have passed paths with him doing his years served at Frontier, Global Crossing, level three, CenturyLink. I hope that I didn't forget something or somewhere, Dave, but welcome. If you could please, maybe introduce NetMavens to us, a bit before I introduce Steve. Sure. Yeah. I I I founded NetMavens to provide expertise to service providers. So, companies like Seaborn Networks, they could be middle miles, it could be broadband, And, you know, there's there's a a definite need for expertise in this space. People that have have done the work before. So with NetMavens bringing together a lot of consultants that cover just a really broad array of disciplines in the network engineering and architecture, business, product management fields, all related around how to help a service provider, excel, grow their business, optimize the way that they're looking for, all those kinds of things. So, that's that's what we're all about, helping service providers be successful. Fantastic. Welcome. And I'm looking forward for you to tell us more about how, what you see of Kentik and how you're helping your customers achieve success with Kentik. Steve. Steve Orlando is the CEO of Seaborn Networks. Steve is no stranger to challenges. He joined Seaborn as their CEO in the height of COVID. And Steve has also got almost thirty years of experience. Prior to joining Seabourn, he's held senior leadership positions at Zayo, CenturyLink level three, Global Crossing, and even Verizon Business. And if you hear a little bit of overlap, between Steve and Dave, that's because they had worked in the past together. And, they have actually, Steve has actually held roles in both domestic and international sales, led strategic planning, product development, operations, and many more areas, and hence made him a superb candidate to be a successful CEO at Seabourn. And, Steve, would you please introduce Seabourn to us a bit? Sure. Thanks, Jezzibell. I appreciate the opportunity. You know, Seaborn Networks is a global technology and infrastructure company focused across the Americas. We provide solutions for the wholesale, high-tech enterprise client space, people that need to transmit latency sensitive data between data centers, innovation, and peering centers across Latin America. I'll talk a little bit later today about the journey that we have. We're a pure play subsea transport company that launched in twenty seventeen. And over the last five years, we diversified our portfolio, building a world class IP ecosystem. And, that's not easy along with accelerated growth, comes a whole bunch of challenges and opportunities, and I'll talk about those a little bit later today. Fantastic. I'm gonna dive into, our presentation, and then, you know, without me spending too much well, actually, I was going to also introduce Steve Meuse, who's our solutions architect. And he's going to, be giving us a demo of some of the new capabilities in both traffic cost analysis at Kentik, but and also a sneak peek into our, AI capabilities. So, Steve. Thank you, Jezzibell. Looking forward to it. Fantastic. Alright. Well, let's dive in, and we're gonna talk about, let me make sure that we are able to move these slides. There we go. We're going to be talking about how AI is changing network demands and what is what it means to have network intelligence in two thousand twenty five. And then we're gonna get Dave and Steve to talk about what they're seeing from the network intelligence perspective and how to achieve, value from with network intelligence. So first, let's talk about the changes in the market, and I think this slide speaks for itself. The complexity in running a network is accelerating. Change is accelerating. Some of the data that let me present to you, right, is that we are anticipating six times the increase in data center to data center bandwidth demand over the next five years. That's with forty three percent of new data center facilities expecting to dedicate to AI workloads. That's almost half of all new data center facilities to be allocated towards AI. Incredible. Right? And projections suggest that nearly two hundred million fiber miles, both metro and long haul, will be needed to meet the capacity projection by two thousand and thirty. That's four years from now. Two hundred million fiber miles to avoid bandwidth shortage. And so if you, at some point, thought, oh, we have plenty of fiber, and we don't need, to have network operational efficiency achievements anymore because we're gonna just light up more fiber, you might want to think again. And lastly, I'll give you this factoid. Twenty nine percent of the communications service providers believe AI will contribute more than half of the long haul traffic over the next three years. Half. So that is an incredible amount of growth. Right? So, if you take a look at the changes over the last two decades, and you will come to recognize that things are going to go even faster, and we need to have even more ability to actually up up up optimize woah. Optimize the operational efficiency of all networks. So let's talk about the exponential traffic growth that we have, and we are anticipating in the next few years. And the only point that I will highlight here, for everybody is to see that the global WAN traffic is projected to increase five to nine times by two thousand and thirty three. That is, if what we're seeing today isn't a ton. Right? If you take a look at the top left tile here where the mobile data traffic has hit a hundred and eighty exabyte per month for this year, for the second quarter this year. That's twenty percent year over year. And sounds like a lot, but five to nine times. Right? It's incredible on how much traffic that we're going to be seeing. So let's talk about what is network intelligence. So network intelligence, has been evolving. Right? To me, network intelligence, I know that here, it says two power, but I think it's it's the power. I think it's two power, the designing, the operations, and protection of networks at scale. And we started in seeing traditional monitoring. You see some of the traditional monitoring companies, and they still exist today. Right? But, that is to allow you to see what's happening in your network. And Kentik started as the network observability platform that helps you to understand why things are happening and what we believe and where we're taking the company is to network intelligence that will guide you on how to achieve operational efficiency, how to make your network better, how to run it better, and what to do next. So, Camtek has always been powered by its data engine and start in two thousand and fifteen, and that still holds true today. Right? To the we've been add we've been able to add so many solution to the platform because of the flexibility and the scalability of the Chantix data engine. Solving the big data problem of network intelligence has helped allow us to enable many of the workflows of our customers. What I want to highlight here is addition of Kentik AI, and AI has been the core focus of our innovation right now. And our AI has been built into the vision of network intelligence. We are finding that AI is a very powerful way to bring the products that we've already have together in a way that having a huge impact for our customers' workflows. This unmatched environment with built that that are built by the network experts, makes advanced operational solutions and analytics tasks quicker and easier for the operators. The perfect foundation for, high impact reasoning for AI. I can't talk about this all day, but I I know that that is not what you're looking forward to. So, let me now hand the mic over to Dave and Steve so they can talk about how they're using Kentik to achieve success. So, I will start with Dave, and so he can tell us about, you know, why Kentik is crucial for NetMavens. Thanks, Jezzibell. And hi, everyone. Thanks for joining today. So NetMavens helps clients with a lot of different things, but one of the key things we get tapped on is has to do with pairing. And it's really much broader than just pairing. It's it's about looking at the entire network and developing a plan to figure out how to optimize that network. Pairing is just one component of that. So it typically starts with an assessment. We're trying to understand, you know, how much pairing is currently in place, where the transit's being purchased from. Are they buying from the right transit providers? Are they buying it from the right transit providers in the right locations? How have past decisions impacted how traffic flows across the network? Is there more backbone traffic than there needs to be? All of these things come into play. And without looking inside the traffic that sits on those connections, it's flowing across the backbone or how it comes in on existing transit or IXPs. You can't develop a strategy. So it's pivotable. It's it's it's so important to have something like Kentik, to do this. And Kentik, you know, does as does a better job than any tool I've ever worked with. I have multiple clients using Kentik. We get off and running really quickly when we use Kentik, to help develop this strategy. So in the strategy development section, we can start looking at, who do we need to add in terms of peers. If there are existing peers, we can figure out if they're delivering in a way that's localized or if we need to engage with them to improve the localization with that network. And then having developed that strategy, we move into an execute execution phase. We start talking to people, setting up new, private pairing, new public pairing, maybe ordering some new transit connections, disconnecting others. All of the the the the outputs from the strategy work happen here. And then we develop reports that show the progress during that execution phase. So so every week, there's there's there's some great progress there to to show for. For for an example, when we started working with with Seabourn, we found some immediate opportunities, some some quick wins. And they were, like, mostly transit when we when we came in, to the picture. And within three months of of, given the go ahead to execute on this strategy, with just four p and I partnerships, we moved forty percent of the traffic from transit to pairing. And without Kentik, that would be complete guesswork. If we go to the next page, we can see, you know, how we've been able to help, Seaborne over the years. This is a CADRE ranking. For those of you not familiar, Kate has been around for ages out of, University of, California San Diego. And, they take a look at, like, what's going on in the Internet ecosystem everywhere, and they produce a global ranking of providers based on the number of peers that they have, transit providers, and downstream customers. And when we started working with Seaborn back in late twenty twenty one, they were sitting at around fifty in the global ranking, you know, which is, you know, still fairly respectable for you know, as a global ranking for somebody that operated only out of Brazil and sold only to customers in Brazil. But, fast forward to today, and they're now eighteenth globally, still selling only in Brazil. So it's pretty remarkable results. And, I think that last slide here, next slide, is, wanted to talk a little bit about a new feature that Kentik has released. So another thing that that we've had clients retain us for is they particularly if they're a middle mile provider, which means that they sell to other Internet providers. They're not necessarily selling to the residential users, but they have different ISPs that they sell to. Each ISP is in exactly the same maturity curve, as, you know, seaborne would be, but in a different stage. Right? So you start out buying transit, then you start adding public pairing, then you add some private pairing. And over time, the way that you depend on your upstream provider changes, and that has a huge impact on the cost burden that you put on that provider. And what we've done is we've we've created a process where we're able to get Kentik set up so that we can extract these reports about how each customer uses the network and then bump them up against some big cost tables, put it into an Excel spreadsheet, use VLOOKUPs to map the prices in and create a per customer cost for each customer. And then you bump that up against the price, and it tells you whether or not you're profitable with that customer or not. This is really, really beneficial for a p and l owner of such a business so that they can see you see stark differences. You think the differences might be subtle. They can be huge. Like, the way one customer uses the network might be making them unprofitable, where another customer at the same traffic levels and the same price might be wildly profitable. So we used to do this manually. It took us hundreds of hours to put these these together across, you know, hundreds of customers that a network might have. And, this latest cost modeling that Kentik has is is really fantastic. We're able to go in and get the the the all of the pricing for all of the costs for the upstreams, for the peers, for the public exchanges. We're able to get this all set up in Kentik so that you just go through and you set up the customer once, and then it's automatically calculating an updated cost for that customer every single month, which is really, really useful because customers are coming up for renewal every single month, some subset of the entire customer base. So you can go in and you can see that this customer, is up for renewal, and they've asked for this, really crazy price point. And this tool will help you figure out if you can make money at that requested price point or not. And if not, how to adjust your negotiation with the customer. So, huge time saver. And I I think that's it for my section. Well Does it know if you want to move on to Steve? Absolutely. So, you know, before I move on, I just wanted to ask a quest quick question. Do most of your customer who utilize Kentik have the same use case? No. So I'd say that the the most middle mile carriers have a very similar use case, and most broadband ISPs have a similar use case. But they're a little bit different, because the way that you approach developing a network is a bit different. So, like, in a middle mile network where you're selling to other ISPs, one thing you're not doing is you're not embedding caches inside your network. So a broadband carrier definitely wants to embed caches. Right? They're gonna work directly with, you know, an Amazon, for example, or a Netflix, and they're gonna figure out where to strategically locate those devices in inside their network to maximize performance for the broadband subscribers. That's That makes sense. They're willing and they're willing to contribute the space and power in order to do that. Where middle mile is gonna tend to wanna consolidate things down into major third party data centers for those interconnections because that's gonna be more economical for them than donating space and power to someone that does caches. So those kind of nuances affect the use cases quite a bit. So, like, a middle mile, for example, doesn't care at all about the OTT capabilities that Kentik has, where you're able to separate out who the CDN is from who the actual application provider is. Right? So Amazon Prime is an application. Amazon Prime also has a CDN, but Prime itself is delivered across, you know, five different CDNs based on performance around the world. Kentik has a tool that can help you see, you know, not just how much is coming from the Amazon CDN or the Akamai CDN or the Fastly CDN, but how much is coming from HBO Max, Netflix, the Disney app, all those different pieces. And those can that can be very useful information for a broadband provider to figure out. Well, do you know, should I go and talk to, this person directly, this OTT provider directly, to see how I can improve performance for any one of my customers who happens to be using Disney as an example. So the use cases and the way that you approach problem solving is different, and that that drives a little bit different use of the tool itself. But the tool has all the capabilities to to to solve whichever use case you're trying to fit or whichever approach to improving performance or profitability you're trying to get to. Thank you. And, Dave, you know, I I know you're an expert in Kentik, but, you know, we appreciate, the way that you are facilitating your customer using Kentik. And, I know that you've been a champion of Kentik, so thank you. I don't feel really feel like an expert in Kentik because I'm still learning something new that it I didn't know it did before every single day. We all are we all are, Dave. Yes. I'm constantly innovating. Right? And it's the beauty of a SaaS platform. We continue to launch new capabilities and new features. So we're all learning every day too, but thank you for pointing that out. Now let me move over to Steve. So, Steve, I know that oh, what just happened? There we go. So, I know that you joined Seabourn about five years ago. So, you know, what what has it been like for five years? We're, how about this? I'll let you start telling us about Seabourn and you know, how you have transformed Seabourn, into a totally new company. Sounds great. You can actually if you'd like, you can go to the next slide. Oh, we're already going. Alright. Yeah. So, I talked a little bit in the introduction that, you know, we were pure play subsea launching in twenty seventeen. And over the last five years, we've diversified the portfolio, and started to sell Ethernet and IP services, in the Americas with the big focus of Brazil for IP transit. And along with that, huge challenges and opportunities. Right? With growth and network strategies, you need to think about costs. You need to think about preserving margin. And for my peers or colleagues that are in the market in Brazil, you all know on the call that it is a hyper competitive environment, both in IP transit and also high supply subsea environment, for the, you know, transoceanic, transport layers that are out there. I'll talk a bit about the journey with our partner, Kentik. I'll first state the obvious, which is, Dave, great job with the the presentation. I think people think Kentik, they think it's a IP engineering tool. We have a different view. My view is is different. We believe it's a commercial analytics tool, and I'm gonna tell you the reasons why here in a bit. But as you can imagine, competing in the wholesale IP business in Brazil, you know, has a set of challenges. You know, we we partner, we compete, we enable companies. In fact, there's about twenty to twenty two of us that are probably chasing the top one thousand targets in the addressable market in Brazil. So there's lack of supplier consolidation. There's accelerated price compression. Not not a great environment to be in. And in the past, we would fly blind on how our customers that we'd bring on board, and we've brought on, you know, two hundred and fifty plus, new customers, how they would impact the cost of the network. And if you don't have visibility, you put yourself at a huge disadvantage. You could lose money with very educated clients in the marketplace. So we kind of implemented and adopted a transformational strategy in this space. So commercial analytics tool, not just for IP engineers. So from the executive team and our IP product leadership teams, we leverage the data from Kentik to make more intelligent decisions. And some of that's around the traffic cost analysis, margin analysis, what we do around our renewal strategies for our customers, who we target, and then ultimately who we try to establish peering with. So as we talk a little bit today, I'll briefly bring up two or three use cases that I think are important to most of us and show you how Kentik helps me see through it. The first one, Kentik allows me to see the top ten largest networks and how my customers are reaching those networks. And why is this important? I need to figure out traffic volume levels. I need to understand the prioritization of traffic. And more importantly, I need to make sure that I have proper peering, proper capacity on the backbone. And some of this information will also lead for sales opportunities so that we can target customers and sell transit to them in the future. Some of this will actually help me and assist the operational team with cost optimization and better performance. And again, all this leads to, you know, more client traffic and, and certainly top line revenue growth. So that's one, one big piece of it. The second piece Dave talked briefly about was the cost analytics tool that supports the sales and product teams on renewal strategies. Again, as I said in the past, we had zero insight on how customers impacted traffic. Was it public? Was it private? Was it customer traffic? Was it transit traffic? Did it stay in country? Was it centralized and localized peering? Did we use Subsea to go carry it across the world, etcetera? We really only had a network strategy for cost. But as we speak today, we know the very specific traffic. We know how it impacts cost, and Kentik gives me that data. So when we look at our top renewal strategies for the next ninety to one hundred and eighty days, I can use the tool to understand the traffic, how much margin I have, what the cost implications are, the trending associated with it so I can package up an aggressive offer to to certainly keep the client, but also maintain certain levels of profitability, or have a really hard conversation with the client that we're making zero margin or negative margin, and we either end the relationship or we try to improve the dynamics of the relationship to move forward in the future. Those are tougher conversations to have, but the data gives you that insight. We also have a bunch of clients in Brazil that wanna trial our services for a week or two before launching with us. Fine. So they trial our IP transit or remote peering services before they purchase. They make a decision go, no go if they wanna be a client. Now I can understand their traffic. I understand their traffic on cost and margin before I extend them a final offer or before we walk away from the table because we don't want them as a customer. So the tool itself will actually help me mitigate risk for future new clients. So those are some of the use cases that that, we see on the product side and the executive side. I think as a CEO, we define what success looks like weekly. So Kentik allows me to establish dashboards. By the way, they're they're user friendly dashboards for me. I'm not gonna be into the IP analytics like Dave or others that are on this call, but I'm looking at the health of the network. What's the balance of traffic of peering versus transit? I get to prioritize the opportunities where I understand the cost improvement and potentially lead generation. I've got renewal strategies to mitigate, the risk of churn and to maintain levels of profitability for us as a company. And I also have a future roadmap for us to look for potential private peering partnerships that could help us with performance and more cost optimization. For me, you know, the favorite tools for me that I look at on a weekly basis, I get them, Jezzibell, I get them, between midnight and six AM on Monday morning. Traffic mix is really important to me. The balance between private, peering, and paid transit, the top eyeball and content clients, I get to see a week over week report of traffic. I also get to look at the incumbent tier one traffic, down in Brazil, which is, you know, Embratel and Vivo traffic. And as that traffic continues to grow, Dave or myself or others can have real peering conversations with them in the future because we see that data and we see the growth of that of that traffic happening on a weekly basis. And then the the second source, ANN ASN source is important to me too. I wanna know what's on the other side, origination and termination traffic from client to carrier provider upstream and downstream. All that's really, really important for the addressable market and who we chase in the marketplace and continue to grow. So I'll I'll finish just with, you know, kind of the last piece for me. As I stated earlier, it's, again, it's not an IP engineering tool. To me, it's a commercial analytics tool, and the data's there for you to action, whether it's cost optimization, whether it's margin improvement, lead generation, or if it's about the future performance and who the next partner will be in the peering ecosystem. To me, that's what the tool does, and that's the measure of success. Well, thank you, Steve. That was so well said. I I don't know that I can top that. I, you know, when I when I joined the company, I thought that maybe the tool's a little difficult for CEOs, for the executives to see and understand, but, I see that you have a dashboard that put together for you, that you utilize with that information you so eloquently, you know, put all that data together to, achieve actionable insight for your business. That is absolutely the goal that Kentik is trying to achieve. So I'm so glad to hear that we are one of your favorite tools. Oh, Dave, you have something to add. If I may just jump in, I think Of course. I think I think for the new person, like, going into the user interface and trying to navigate, Kentik, it would be quite hard for an executive because they wouldn't really know what they they were looking at. So I mean, the approach that that that we took was we started by explaining what was going on. And we, you know, explaining what was important in in the data to the operation of the network, what was important from a commercial perspective, built you know, used reports from Kentik, you know, just in PowerPoint for, I don't know, like, you know, a year or two, went through those on a regular basis explaining what what what was being looked at. And and after a while, you know, the executive now knows when the second they see that report, they know exactly what it says, exactly what it means. And at that point, we just set up a weekly subscription inside the Kentik tool that just sends out that report, just emails it out as a PDF. And that's what that's what Steve looks at every Monday morning is stuff that he's been looking at for years. Now it's just automatically generated inside the tool, and he and he knows right where he wants to go in that report to find out what he wants. That is fantastic. So, both, you know, one can set up a dashboard, right, for the executives and their subscription to the data feed as as David said that, he's able to receive on a weekly basis. And, Dave, I'm sure that NetMaven's help set up the dashboard, and, obviously, Kentik's customer success team can also help set up dashboards, as teams or executives need them, and can be educated of them. So Yeah. Exactly. I mean, Steve doesn't even log in to Kentik, and I don't think most executives would unless they were real propeller heads. But the the subscription gives them, like, a just a super easy way to consume the information without having to, you know, remember their password or where to go or any of that kinda thing. Again, the the the the tools that we set them up are board board ready tools too. We've used them in board meetings as well. Very easy to explain, and, it's repetitive. Right? When they see it once and twice, they finally learn from it, and they know exactly what to look for. And it's good insight at the very beginning of the week. Well, for you, Steve, as the CEO of the company, it gives you the visibility, of what's happening on your network and, you know, give you an understanding of both maybe cost and margin, right, to allow you to see, at a glance what's changed from a week to week basis, through your subscription of the data. And from the from, the platform itself when you actually log into the portal, Dave and others and other operators have the ability to drill into the details. So without further ado, let's actually drill into the details with Steve so that we can take a look, at, you know, what would an operator see when they're looking at the traffic cost, which is the new feature that we've lost, we've launched. So, Steve, all you. Alright. Thank you, everybody. So, one thing to note, I know we talked a lot about usability of, you know, maybe a CEO wouldn't be logging into the system. But one thing that Kentik has really spent a lot of time on is what a term that we use called democratization of the data. We try to make the system usable for sort of any approach because as as we've talked about today, for any sort of network that is making money off of delivering of bits, the data the underlining data that we're collecting and and modifying is important to all aspects of the company. Of course, we always think about the engineering and operation side of it, but, of course, we're you know, Steve's the CEO. He needs to be able to, you know, make business decisions over that. The accounting team might need to be able to do use it to, you know, look at, invoices and make sure that they're correct and so forth. So there's many, many, many aspects of every persona within a a company like that who has access need for the underlying data. So we really spend a lot of time trying to make it easy for each one of those folks to have easy access to the data. Because typically, in my experience that this data has been locked away and complicated to use systems where there was only one or two experts, who inevitably accidentally become roadblocks to, get producing data out of systems, which which is really critical to running the business. That being said, one of the questions that was asked was of Dave earlier was, like, what we're using before, and Dave said spreadsheets. Right? So, unfortunately, that's been what has been used to kind of perform these operations for a very long time. So at Kentik, we took an approach of trying to really understand what our customers' day to day workflows are, what they're using these, the day the underlying data for, and try to figure out ways to, make their life easier. Look at those day to day tasks and and eliminate those, you know, need to have spreadsheets for these these operations. So it all starts with, understanding your connectivity cost. What it is that you're paying on a monthly, basis for, peering, paid peering, we settlement free peering because as we all know, peering isn't free. There are always costs associated with that, transit and so forth. Right? So, the idea being that what you're trying to get to at the end of the day is an understanding of your cost. So here in our connectivity cost model, we can see here at the end of the month, this is what our estimated monthly, spend is going to be on on bandwidth, what our cost per megabit is, how much traffic we have coming in and going out. And then also the ability to break that down by all the varying providers that we have. In this case, all the different providers that we were contracted with, what our cost is. Is that cost going up? Is it going down? What's the cost per megabit? And be able to really do some deep analysis of that traffic. Now the reality of this is that there's lots of technical underlying information that supports this, this data, but we really need to understand the business relationships between us and these connectivity partners. So we have, the concept of a provider. In this case, these are all the different, external connectivity partners that we have that we're paying services for. It could be transit providers. It could be paid peering. It could be an IX port. So here we can see we have, you know, ports at Equinix and France IX. But in this case, we'll take a look a little bit closer at some of our transit partners. Now what we need to do is map the business logic that we have with these relationships onto the technical platform. So in the case of NTT, as an example network here, we might have, a network that spans the continent, and we might have different connectivity contracts depending on what market it is or, some other potential, you know, whether it's Asia Pac or North America. So what what Kentik does is we allow you to actually model the contract. So in this particular case, you can see we have some, some flat spend fees that we have, and then we can, map our commit, which is our what what the cost per megabit is, but we have the ability to also tier that. So we can say, you know, my contract says it's seventy five cents per meg up to the first two gigs, but anything or, for, you know, minimum min committed two gigs, but anything when I hit to the eight gig mark, anything over eight gigs, suddenly is at fifty cents per meg and so forth. So that's that's these are contractual terms that we we need to worry about. So we can come in here and model all of that very easily through the the user interface. But also things like when does the contract start? What day of the month does it start? When is the contract end date? How the billing is actually calculated because some providers use different, calculation models. Is it a peak of sums or a sum of peaks? Is it done by individual interfaces? The interface is grouped together, what the cost per megabit, how you do the the banning and so forth. Right? Now once you have this contract model, you then need to map that to individual interfaces on the platform. So we can see routers in different parts of the the the platform get modeled, to that particular, cost group or or contract in this particular case. So at the end of the day, we can see we have multiple contracts, multiple cost groups. They each have their own static charges. In this particular case, it could be cross connect fees. In in years past, some providers charge just for having running BGP from the transit port. So there were all these, you know, minimal fees that you can, collect and and Some still do, Steve. Some still do. Yeah. That's unfortunate. But, I I can think of the name of a few, but we won't mention them. So but but, you you know, the silly things like that that cross connect fee that, you know, in in LA is is, you know, two hundred dollars, but in Ashburn, it's, suddenly creeping up to, like, five hundred dollars. Right? So that that'll play, into the cost of the of the delivery of that service. So you need to be able to model that into the total cost. Right? So we have the ability to do that. Now at the end of the day, per provider, we can give you lots of information about, the cost just strictly for that provider, what the blended cost per megabit is, you know, ingress, egress traffic, all that. So we can kind of kill you with information including, is that cost going up or is it going down based off of your utilization and so forth. Right? And and so the connectivity cost, really is sort of like an underlying dataset that enables you to do lots of, different business functions. Right? So, of course, just managing this cost alone is a huge task, and having this connectivity workflow eliminates that need for the spreadsheet, for the the the specific task of keeping track of these costs. Now I I've done this for a living, for a very large, cable MSO in this country. And I can say it was nerve wracking having to come up with a projected budget for what my transit costs were going to be for the course of the next twelve months, hand that to, an accounting team, and, and then looking at my own calculations and realizing that, oh, I did not account for something, and the bill is going to be a lot higher. So in this case, maybe I only, calculated twenty five k a month, and it's now turning out to be thirty seven k. Well, I can get ahead of that if I can get to my accounts payable team and say, hey. I I goofed up. You know? And it's better to, let them know ahead of time than have to, stand tall behind in front of the man many months later saying, why are you hundreds of thousands of dollars over your budget? Beg for forgiveness. But once we have this underlying data, we can then proceed to kind of to really start to do the analytical piece of it, which is looking at traffic cost as we apply it to some smaller increment of traffic. Now if you look here in our traffic cost module, you can see the different, dimensional data that we can use to break that traffic down. So, let's see. If we look at one particular one, we have a prebuilt report here for our customer port for our, fictional customer called Acme. Now what this is doing effectively is telling you what the estimated total cost for this this customer's traffic in this port. Now this is a a fictional network, so, you know, we we're not really worried about the actual cost here, but you can actually see what the total blended cost per megabit is based off of all the different edge connectivity contracts you have in the other modules. Right? So you can see as we follow the flow of traffic here, some's coming in through transit providers, some's coming in through, you know, a cuss an additional other customer port. Some is coming in through private paid private peering, and some is coming in for free private peering. So what we can do is then break it down. So for this, you know, fictional customer, we can go through and actually zoom in and look at, you know, for this free private peering port to charter, you know, the total ingress megabits per second is actually this is the, you know, the blended cost for that. So at the end of the day, what this is effectively allowing you to do is to say, is this customer cost effective? If they're asking me for, you know, a discount, should I give them the discount? Because I can now see very detailed costs, of of what what where this traffic is coming from, or is it going to, it might potentially, you know, going to move off of an old provider like Savas here. Maybe we know the support's going away, so that traffic will end up shifting to my transit. So I I have some visibility if I make changes to my network. I can understand where that traffic's going to, potentially. And is that going to, you know, you know, affect the the total profitability of this particular customer or not? So that's one particular example of of how to do that. Now if if we look at it, from another perspective, we can do things like look at, OTTs, OTT, options. So if you're a broadband service provider, you can understand you know, Kentik has an OTT capability that allows you to break out traffic by services, not necessarily CDNs, like, you know, Akamai or or any other or, you know, Amazon, you know, CloudFront or something like that, but actually look at the services writing on top of the network. So we can do something like break that down for an OTT service like Apple. You know, this could be Apple Video or, you know, any of the Apple services, Apple TV maybe. We can look at that traffic in extremely granular detail to understand where that's coming from. Is it coming in from all these number of different providers? And then look down here in the chart below to understand what of course, we still have the the total blended cost, so my cost is actually only five cents per meg. What my, actual cost, in terms of total dollar amount is, but then to be able to really get into really deep, deep detail in terms of how that traffic's delivered to me. Is the cost increasing? Is it is it decreasing? So, again, I can help understand the the profitability or, you know, if if I'm being asked for a discount from the customer, is this is a good decision or not? And, of course, we can snapshot that that cost over time so we can look backwards to see, okay. Well, in this particular case, the cost has actually increased a little bit over the course of the last few months since we started, capturing that data. So maybe we should, you know, you know, reconsider some business decision there. One one thing I will add, if I could interrupt, Steve, is that, it is really important to get help from someone, when you're inputting your costs into that traffic cost section. Yeah. Particularly if you're gonna be using this this cost module, because simply taking the contract that you have and then basically copying it word for word into the commercial structure that Kentik gives you might produce some unexpected results in the customer cost analysis piece. So I'll give you I'll give you an example. You know, let's say that you have a flat rate port or a private peering interconnection that you pay a fixed amount for, and it has extremely low utilization. Your cost will actually fluctuate based on the scale of how much traffic is on that connection. So there's there's a better way to set those up in the traffic cost model, which is to compute a a cost per meg based on the infrastructure cost and the available capacity that you consider maximum, usually, like, seventy five percent of the total circuit. And when you do that, then you will get much more accurate, customer cost models that don't penalize you if you happen to have fixed rate ports that are at very low scale. Yeah. Yeah. Of course. The math works out that it's assuming a hundred percent utilization. So, of course, yeah, you have to take that into consideration for sure. Now We Go ahead. I just just I just want to add for timing, wise, we might want to, wrap this up and to go to the AI portion because I think a lot of the audience may be very interested in what's to come, in AI. Yeah. Of course. Kentik has is about to release a new product called AI advisor that allows you to kind of ask questions about the network. Right? It's the putting the intelligence into, network intelligence. Right? So we can ask questions about the network. So I could do something like, say, how can I save cost with traffic? And and the AI will go out and interpret my question, try to understand what's happening in terms of the distribution of my traffic, and then it can go off and actually start to look at, various pieces of data into the network. So I'll expand it here. So it's processing a little bit of takes a little bit of time to understand...