Today’s episode features an interview between Matt Trifiro and Mehdi Daoudi, Co-founder and CEO of Catchpoint. In this interview, Mehdi discusses Catchpoint’s cutting-edge approach to monitoring, the user demands that will drive innovation in the third act of the internet, how the current pandemic will serve as a massive digital accelerator, and much more.
Today’s episode features an interview between Matt Trifiro and Mehdi Daoudi, Co-founder and CEO of Catchpoint.
Mehdi’s experience in IT leadership at Google and DoubleClick inspired him to build the digital experience platform he envisioned as a user, and Catchpoint has been at the forefront of digital experience monitoring since it’s launch in 2008.
In this interview, Mehdi discusses Catchpoint’s cutting-edge approach to monitoring, the user demands that will drive innovation in the third act of the internet, how the current pandemic will serve as a massive digital accelerator, and much more.
Key Quotes
“Our mission is that we want to monitor as many things from as many locations as possible...and monitoring needs to be part of business strategy, not monitoring for the sake of monitoring.”
“It doesn’t matter what you monitor as long as you monitor it from the right location. So it's very important to monitor things from where the end users are. That's what we focus on at Catchpoint is outcome-based monitoring.”
“We focus on four key criteria. Our technology allows you to look at reachability-- can I get to you? If I can get to you, are you up or down? If you're up, how fast or slow are you? And then how reliable are you? And reliability is extremely important. Reliability means are you able to deliver the same quality of service 24/7?”
“Having a single CDN vendor is no longer an option...I’m surprised by how fast that movement has taken off-- before it was reserved for a few large, super sophisticated companies...but today literally everybody I talk to is doing multi-CDN, multi-cloud.”
“I think this pandemic has shown the need for better data, better network connectivity, better end-user experiences. It is a digital accelerator that no industry has ever seen. It’s a kick in the butt for all of us to innovate, bring solutions, and deliver on this challenge.”
Sponsors
Over the Edge is brought to you by the generous sponsorship of Catchpoint, NetFoundry, Ori Industries, Packet, Seagate, Vapor IO, and Zenlayer.
The featured sponsor of this episode of Over the Edge is Catchpoint. Catchpoint gives critical knowledge to help optimize the digital experience of your customers and employees. Learn more at catchpoint.com and sign up for a free trial.
Links
Matt: Hello everybody. This is Matt Trifiro, CMO of edge infrastructure company, Vapor IO, and co chair of the Linux foundation state of the edge report. And today I'm here with Mehdi Daoudi, CEO and co founder of Catchpoint. We're going to talk about Mehdi's background in [00:08:00] technology, how he got started, what he thinks about edge computing and the future of digital experience monitoring.
[00:08:07] How are you doing today, Mehdi?
[00:08:11] Mehdi: For getting together. It's a, it has been a long time.
[00:08:13] Matt: it is, it is. And it's always a pleasure to talk with you. So I think this could be quite a fun, fun interview. I wanted to start. Cause you know, I actually don't know a lot about your background and you and I have only spent time together in business context. So I wonder how how'd you even get involved in technology?
[00:08:29]Mehdi: so when I came to the U S I was in a French business school, I used to a and M. And so in order for me, I fell in love with New York and I, in order for me to stay in the U S. I got the job at Reuters and the risk management software division. but they didn't need marketing people. And so they only wanted the Linux and Unix wax at the time Unix.
[00:08:53] And I said, well, you know, I can learn that. And I learned, yeah. And, in two months, but I've always been,
[00:08:59]
[00:09:00] [00:08:59] Matt: You started in the U S as a sales engineer for writers.
[00:09:03] Mehdi: Yeah, yeah. As a sales engineer. And then I moved from there to DoubleClick in 1997, again, as a sales engineer. And, I was the first se on this new platform, they were selling to publishers and, an incredible ride.
[00:09:17] I ended up. You know, spending 11 years there until Google bought us and then a year later, and I didn't see time, but, as an se, a DoubleClick, what got me into what I'm doing today was my frustration. So I was an se and, I would go on demos because there was no zoom or WebEx at the time. So you would fly to meet customers and the system would just crap out and was just like break
[00:09:40] down
[00:09:40] Matt: what, what timeframe is this?
[00:09:42]Mehdi: this is 1997, right? So we had, we had a single data center at 41 Madison in New York. We had the, an ISP at the time called VBN and U unit.
[00:09:53] Matt: Right. And there was barely a web browser.
[00:09:55] Mehdi: Right, right. Barely of web, right. I was running, we were selling this ad serving technology [00:10:00] and the system would just literally crash out every few minutes or every two hours.
[00:10:04] And as an se, I would go in front of these customers and I would get super frustrated because like I'm in the middle of a demo and just, boom, the thing disappears. So I said, we have, I have to get, I have to get early notification system. So in my apartment, I bought this. I bought this. And again, back then we had 28, eight modems, right.
[00:10:23] Or 56 K modem secretary. I still have mine. and, I bought the server and I bought myself a page. Sure. And this is not reimbursed by double pick at the time. And I bought this amazing software called SiteScope, which is this thing I found online. It was a monitoring tool and I built my first synthetic network.
[00:10:42]between my house and the friend's house, and it would literally monitor DoubleClick. And as soon as we would have a problem, we'll get paged. So as an se, I would have a nice transition to talking about the dog cats, weather, whatever it might be. And so I resigned from double click. I sent this angry email to my CEO after,
[00:11:00] [00:11:00] Matt: Wait, wait, so I'm going to pause you for a second. So, so you built in your apartment. Okay. The first thing you said, synthetic network, what the heck is this synthetic network
[00:11:08] Mehdi: So I want them to have an early detection mechanism to tell me that DoubleClick was having a
[00:11:13] Matt: so that you don't mess up the sales presentation?
[00:11:16] Mehdi: Exactly. Exactly. So I want that. So it wasn't called synthetic at the time. It was
[00:11:20] Matt: get a page, you get a page and you say, Oh, wait, before I show you a demo. Let me tell you about our high quality ad
[00:11:26] Mehdi: Right. Exactly. So it allowed me to have much better transitions but at some point it, well, I grew tired of that.
[00:11:34] So I sent an email to my CEO, Kevin O'Connor at the time, and I told him, listen, Kevin, this system sucks. I'm quitting. And that typical, Kevin, he replied back in a two sentence, email, stop bitching about it, fix it. Now you're in charge of monitoring a DoubleClick.
[00:11:51] So
[00:11:52] Matt: I love it. I love it.
[00:11:54] That must've been, that must've been yes. Right. With anybody to anybody with ambition, a challenge [00:12:00] is exactly the right way to
[00:12:01] reply it. Yeah.
[00:12:02] Mehdi: I love that. So that's how I got into the monitoring business. So I did that for 10 years at DoubleClick. And then, and then Catchpoint, which is just an extension of that, except we're doing it for other customers.
[00:12:12] Matt: So you say monitoring and, and the, the, the, the verbiage that I've seen, Catchpoint use is user experience monitoring. Can you describe to me what that, what that fully encompasses?
[00:12:24] Mehdi: Sure. so monitoring obviously is, is a loaded term. it has data central monitoring, network monitoring, temperature monitoring, all kinds of stuff. But so at DoubleClick, I was in charge of all of that And one day we had an outage and I walk into my knock. And everybody was chill. everybody was normal day in the knock and my hair was on fire.
[00:12:48] I mean, obviously I don't have any more of that because of that. It all burned off. Exactly. And I told them guys were down. It's like, so everybody looked at all the internal monitoring systems. We had, [00:13:00] the network was great. The firewalls were great. The databases where everything looks great, but from the outside perspective, It wasn't.
[00:13:08] And so the only way I knew about the issue is we use the tools
[00:13:12] at the time in between 97.
[00:13:14] Matt: PC in your apartment?
[00:13:16] Mehdi: Well, that, that exactly. But also we have the, at the time we had those, so companies like Gomez and keynote systems, which basically, created the first synthetic network monitoring system. So. So that, that basically gave me a huge lesson, which is the, as a matter of what you monitor as no, cause you monitor it from the right locations.
[00:13:38] So if I am, an ad serving company, I'm not serving ads to my other servers in the data center, I'm serving ads to end users. All around the world with very different latency profiles and whatnot. So it's very important to monitor things from where the end users are. So that's, that's what at Catchpoint we focus on is more the, the outcome based monitoring.
[00:13:59] So if [00:14:00] you're setting a. If you have a website that sells TVs, you want to make sure that you can sell TVs. You still need all the other monitoring. Don't get me wrong. Right. Obviously it's like, what are we monitoring? How is this impacting the business? The other thing that I learned through my DoubleClick experiences monitoring for the sake of monitoring, doesn't put food on my table.
[00:14:21] The monitoring needs to answer and to help business. So very early on, I walked into the typical it shop finger, pointing the business, yelling at the it, the its screaming like this business guys don't know what the hell they're talking about, but at the end of the day, we're here as an it shop to help a company succeed.
[00:14:41] And so I turned the monitoring to be more business centric. How do we help our sales people win more deals because we can show that we're better than the competition. How can we be the most reliable ad serving platform in the world, blah, blah, blah. So, so it was always like monitoring needs to be part of a business [00:15:00] strategy, not monitoring for the sake of monitoring.
[00:15:02] Matt: Yeah, that makes a lot of sense. And, so where did catch point emerge in this? The story?
[00:15:09] Mehdi: So, towards the last few years at, at DoubleClick before the Google acquisition, I also run an R and D team. So I got to know
[00:15:18] these
[00:15:19] Matt: So what, what year? What year are we in roughly now? Oh,
[00:15:22]Mehdi: we are in a. 2004, 2005. And so I got to know my cofounders of Catchpoint. So they came and worked with me in R and D at DoubleClick. And I loved working with these three individuals.
[00:15:37] We just got so much stuff done. We, we were cranking projects left and right. Some of them became real product. So, and so in 2008, after the Google acquisition, Google is an amazing company and, I just. Thought that this wasn't for me anymore. I was little bit done with the ad serving world and the media world.
[00:15:57]and I thought monitoring, with the [00:16:00] cloud explosion that I saw a little bit at Google, was going to take over the world. And if the cloud was going to take over the world, monitoring needed to change and monitoring is going to be more from an end user perspective. So We took the hardest decision is to leave Google.
[00:16:16]With all the perks and the nice salaries and the GSEs and everything and the free food. and off we went. So in 2008 and September, we launched Catchpoint, and, we thought we were going to go far away with the first $80,000. We raised among the four of us that got burned in two months, just in legal fees.
[00:16:36] And the rest is history.
[00:16:40] Matt: You filed for bankruptcy?
[00:16:41] The rest is history.
[00:16:42] Mehdi: no, we, the rest is this we're still here, but it has been, it has been an interesting ride. So, you know, 2008 was an amazing year to start the company, as you can remember.
[00:16:52] So it was the crisis. Yeah, financial crash. And so we had to, [00:17:00] and, you know, friends and family and just literally we, we worked very hard, the four of us
[00:17:06] Matt: So you're a largely bootstrapped. Are you, are you, were you ever venture funded or.
[00:17:10] Mehdi: We were, but only five years after we started the company. So when we raised the money from battery ventures, in 2013, we had about a hundred customers and we were, I believe three or $4 million, revenue, company. And we were hugely profitable
[00:17:28] because we, we
[00:17:30] Matt: You had to be. Yeah, you had to be, you had to be. so now it's 2020. can you tell me how big of a business you are now? Like how many employees like
[00:17:38] with
[00:17:38] the shape of your businesses?
[00:17:39] Mehdi: Sure.
[00:17:40] So we're 250 employees, today, and it's an awesome responsibility as you as you know, , but we're growing obviously. Yeah. and, we above 450 customers today. we're very lucky to be working with some of the most amazing brands in the world.
[00:17:57]some of them are mutual friends like backup [00:18:00] and many others, but the vast variety of company types, in the media entertainment, technology, SaaS companies, and it has been an amazing
[00:18:09] ride.
[00:18:10]Matt: maybe tell me about, a company that you can talk about. like w what problem they came to you with and how you provided a solution and what happened,
[00:18:19] after you provided a
[00:18:19] solution.
[00:18:20] Mehdi: so one of them recent is a very big organization, fortune 20 company kind of thing, big need to expand in China from a digital perspective. And so they want you to understand. how their performance and reliability and availability from their website's perspective is from China.
[00:18:41] So in that process, we, so again, we're in the syntax. I mean, one of the products we have is the synthetic monitoring business. So we rely on about 850 locations that we've built around the world. So these are data centers that we operate, the equinoxes, the packets, et cetera of the [00:19:00] world. And, so in China we have about 150 locations that are collecting telemetry that are used not only by technology teams or SRE teams, but also in this particular case, by a business organization that is trying to understand how is our digital presence in China doing?
[00:19:19] Do we have any challenges? How do we go and fix it? So we ultimately. Matt, what we do is we have the business, expand their revenue or protect their revenue. Because if you don't know how bad things are, how
[00:19:33] Matt: Yeah. I think the example you used in the intro was imagining you're a seller of TVs. What you really care about is how many people are buying TVs. Was it the direct function of how many people are capable of buying TVs? At least, you know, we're having a good experience because we know, you know, every, every couple of seconds of delay on a shopping cart, for instance means that many more abandons and.
[00:19:53] Yeah, that's interesting. Okay. So, so how can you, can you help our listeners understand how [00:20:00] your technology works? Like how do, how do you actually measure without being inside the head of the person?
[00:20:04] How do you actually
[00:20:05] measure
[00:20:05] user experience?
[00:20:07] Mehdi: Yeah, I saw. So you, you just used a very interesting term when we were talking about the person buying a TV. So the way I look at Catchpoint is where like digital mystery shoppers in a way. so except that we focus on four key key criteria is because on the internet you have to.
[00:20:24] So our technology allows you to look at a reachability problem. Can I get you, can me a user in Los Angeles on spectrum can get to
[00:20:35] Matt: So, so, so let me, let me pause you and see if I can say that back. So the actual end point, meaning someone that looks like a user on the other end of the last mile network. So at the end of a forgery network or at the end of, somebody's
[00:20:46] fiber or a cable plant. Okay.
[00:20:48] Mehdi: Yep. Yep. That's exactly it. So we have these 850 mystery shoppers that are connected to various networks that matter and on the internet, in the right, in the [00:21:00] right physical locations where populations are, are a key factor. And so reachability, can I get to you? If I can get you, you, are you up or down?
[00:21:09] If you're up, how fast or slow are you? And then how reliable are you? And the reliability is extremely important. Reliability means for the majority of our customers and for me as well. Yeah. Is, are you able to deliver the same quality of service to those users? 24 seven, because it's easy to deliver great performance, great availability at two o'clock in the morning, if nobody's on your side, but can you deliver the same one at 2:00 PM at noon?
[00:21:36] At any hour, because exactly or whenever is nobody's watching TV these days or Netflix, which is now 24 seven. So that, that level of consistency is reliability. And today you hear the word site, reliability, engineering, those kinds of things. And it's really about how can we put teams and tools and metrics, whatever to drive that reliably.
[00:21:57] So our customers care about those four. [00:22:00] four quadrants, if you want, or Cardinal points. And that's, we have them with our technology to answer those questions. Are you reachable? Are you available? Are you performing? And then how reliably are you
[00:22:10] doing that?
[00:22:10] Matt: Yeah. And when you think about an eCommerce site, I get the kind of mystery shopper examples. So you've got these little bots that are going out and. And testing the important parts of the site, 24 seven, and then reporting back. Do you, do you, I imagine you, you work with other segments, I mean, do you have like a, you know, synthetic, Fortnite players and, you know, do you go
[00:22:30] into
[00:22:30] gaming?
[00:22:31] Do you go
[00:22:31] into just about
[00:22:32] everything?
[00:22:32] Mehdi: Yeah, absolutely. Yeah. We go in and absolutely everything. So there is obviously the web part that is important, but the internet is more than just a webpage e-commerce page. So we have very good that, for example, monitoring as infrastructures, we have a lot of gaming companies that use us where latency is extremely important.
[00:22:51] So we have certain capabilities to be able to, yes, 11 year old walked into my office one day and said that our network socks there is [00:23:00] lag. Right. And I looked at them as like, where did you learn all of this? But that's exactly the importance. The age, when I looked at him, I said, this is why we must do better.
[00:23:11] Right. This is why even edge computing is more important today because it's about
[00:23:15] latency.
[00:23:16] Matt: And it's about through the eyes of the user, right?
[00:23:20] Mehdi: exact, my 11 year old has a one. Millis has a one millisecond timeout in his head.
[00:23:25] Matt: Right.
[00:23:27] Mehdi: like 56 more than she can wait a few
[00:23:29] Matt: sometimes you wait a half hour, right. To download a file. Yeah. Yeah. That's really interesting. And so, you know, as we move into this new era, that's really different and we'll come around edge computing on this edge computing podcast, but I sort of to want to stay here for a minute, So, again, this sort of general category of what you do is user experience monitoring, but this world, like the user is. It might be a machine. It might be a sensor or it might be something else. Are, are you, are you monitoring? Yes. Are you monitoring
[00:23:57] aren't like humans?
[00:24:00] [00:23:59] Mehdi: Absolutely. Absolutely. So when you think about, so we do web monitoring, DNS doing all that good stuff, but a lot of our, our second most popular, platform or tool that is used in our platform is our API monitoring. Because again, the majority of the stuff that is happening today is machine to machine one system, calling another system via an API to get something even, even actually on a webpage where you have a merchant gateway, but credit card is an API call.
[00:24:27] But there's one machine to another. we also worked with a bunch of other companies that are in the IOT. We built a, an MQTT monitor, for example, to be able to monitor and pretend that IOT device, trying to register with them with an M with the gateway. and so we are always innovating and trying to bring that level of visibility to a different.
[00:24:50] Parts of the intranet. That again are not just necessarily somebody
[00:24:55] Matt: that's neat. And it brings, it brings a new for me, a new [00:25:00] understanding of your company name. I don't know if this was the origin, but I can imagine literally you sending little, you know, putting little, little. Bots out there, so to speak that are catching the data points that are important. That's really, it's very, very interesting, very interesting image.
[00:25:14] so, so you know, you and I are both fond of the three acts of the internet, and this, this emergent third act, which, we're just starting to see now. can you tell us a little about, about, you know, your interpretation of the third act of the internet and how you're changing your business to adapt.
[00:25:35] Mehdi: Sure. So it's driven mostly by our, what we see from our customers and what they are worried about. for the future. So when I think about the internet, it was again, I mean, when, even going back to my DoubleClick experience, we had one data center, which ISP, no Akamai at the time, there was nothing. And then customers were complaining about how slow ad [00:26:00] serving was because then in an ad serving system, there is the ad matching and then we needed to deliver the banners right.
[00:26:09] Oh, it was, it was more than that. Trust me. And, we decided at the time it's like, okay, we need to distribute the biggest payload, which is the media, the actual gifts and the JPEGs at the time. So we started building our own media networks. So we went and opened like three or four or five other data centers.
[00:26:26] And we said, okay, we're going to use DNS. A round Robin people based on geography and whatnot. And then there was a company like Akamai that came before that we used another one called speed arrow. so the first part of the intranet was like a single data center, monolithic thing. The second part was like the CDN.
[00:26:46] Right. So, which is like let's, let's spread the content. As close as possible to the user. But when I look at some of the technologies that are out there today, SAS applications, games, I mean, [00:27:00] telemedicine cars and whatnot. You need to be even closer to that end user. You really need to reduce that latency.
[00:27:09]when you look at VRS, when you look at all these amazing telemedicine, especially in the COVID thing, I just had the telemedicine, thing with my doctor for my arm a few days ago, it was awful experience. I mean, doing it over zoom, I mean, really.
[00:27:25] Matt: How how, yeah. So, so how, how, how does Catchpoint make that better?
[00:27:32] Mehdi: Well, we don't have anything yet there. Right? However, we, we are seeing customers demand low-latency stuff. And so therefore what happens is you engage, you have to break away from just the CDN model and have capabilities much closer. It told us Angeles, for example, where I am, or I don't know where you are, these days map.
[00:27:55] Yeah. So you're in Berkeley. So Berkeley isn't is not San Jose. [00:28:00] You know, if, if you, most of the data centers are in San Jose or God knows what else you still have 20 minutes, second latency, maybe that's too much for certain applications. So how can we get even closer to the end user
[00:28:12] for even applications, not only static content or dumb content or whatever you want to call it.
[00:28:18] Matt: raise a really interesting point because, you know, I tend to think of the big driver of this third act of the internet.
[00:28:24] being that we're moving from a world of primarily humans talking to machines, to machines, talking to machines, and not that humans are gonna stop talking to machines.
[00:28:32] in fact, our demands are going to become. Are going to grow as humans, right. you know, right now we wait ones of seconds for Facebook to refresh. That's fine. But if there's a one second delay in a video call, it's, it's absolutely frustrating. It's like, An 11 year old, trying to play Fortnite with five milliseconds pink.
[00:28:53]and so, you know, like the e-commerce example. So I, I ran an eCommerce company and we had an interesting cart abandoned [00:29:00] problem, which was, you know, in order to provide the best information in the cart. In real time, we were constantly calculating shipping and tax. Based on what people are typing in.
[00:29:11] And so we had to make a bunch of, I calls and these were two different vendors. Cause we didn't know how to calculate sales taxes complicated. So we'd called yeah. Avalara. I think there's a company we're using. So we would, you know, and they would take 50 milliseconds and then yeah know the shipping alternatives, which we'd have to compare.
[00:29:24] We'd have to hit three shipping companies, you know, and suddenly your cart. It takes a few seconds to refresh. Right, right. And like people just say, okay, I'm Don by. and in, in an, in a world of machines talking to machines, you know, like ones of seconds or how humans evaluate their tolerances, but like milliseconds and microseconds or nanoseconds are how machines, measure their, their, you know, speed or latency or experience so to speak.
[00:29:52] And so, yeah, I can, I can see why, Your business is going to become even more important in this world because you know, at least with the human experience, you literally could hire [00:30:00] 800 people out in the field to report back. But
[00:30:03] Mehdi: not with machines.
[00:30:04] Matt: When you, when, when it's important to know the changing conditions like every second.
[00:30:09] Mehdi: Right. Yeah. And so, and so what we're seeing from our customers, especially the ones that are extremely latency sensitive, whether they're car manufacturers that are in the self driving business, those kinds of companies, they really need super low latency between. The car to a data center where telemetry is going to be sent and decisions are going to be returned.
[00:30:33] And so they're asking from us, it's like, Hey, Catchpoint I know you have all these locations, but. What are, what are some of the other locations? Can you, can you show me an edge location so I can choose which cloud provider to use, which CDN to use? How is my latency to a bunch of people in Santa Monica versus a bunch of people in downtown Los Angeles? is how far, how far some of
[00:30:59] Matt: Yeah, [00:31:00] really nuanced.
[00:31:01] Mehdi: absolutely. So we, we, we went from one. So, we went from a one country's latency to maybe a city latency. We're going to have to get down to the zip code. I mean, I'm exaggerating, but it's literally that
[00:31:22] Matt: So you're at 850 sensing locations today.
[00:31:27] Mehdi: Right. Well, I hope my CFO is not listening to this. Well, I think we're going to hit about a thousand, early next year, locations. and, the mission, our mission is we want to monitor as many things from as many places as possible. So we're doing a lot of things to, to add. We're focusing a lot right now on the edge, as you know, we are in talks with various companies, various ISP.
[00:31:53] Jeez. And I think a lot of them are realizing the need of, of edge compute and being able to provide, [00:32:00] compute at the towers, compute that, certain smaller facilities. I was on the phone with Cox the other day, which is like, they're starting to, they're looking at offering some edge compute capabilities for on the Cox network.
[00:32:14] For example, obviously the big guys are playing in this Google, AWS, et cetera, but companies like yourself, we're innovating and trying to come up with super, Scalable solutions when it comes to two edge platforms. And I think the key thing is going to be, you know, very modular, bare metal capabilities, but also compute VM containers, that kind of stuff.
[00:32:40] But the thing is going to be scaling. How, how many locations? So
[00:32:49] Matt: Yeah. I think that's, I think that's really true. Well, and you can tell your CFO, maybe there's partners that will help you pay for the deployments. well, it, you know, it also seems [00:33:00] that, you know, this kind of. Like, like, like what's your measuring potentially gets more complex, because there's a lot of there more points of failure or more points of delay.
[00:33:12] Yeah. So, so in addition to measuring kind of the end point, like the consumer, whether it's a device or a human,
[00:33:21]are you also looking to instrument
[00:33:25] Mehdi: The whole path, the whole path. Absolutely. So the customers that we're talking to today that are, again, these are existing customers that we're engaged with. They want to know, for example, somebody is in Santa Monica, They are talking to an Akamai, let's say, but then I come, I asked to go back to an origin somewhere else.
[00:33:45] And then I come, I asked you. Talk to another data center somewhere. So they're looking at, they want to understand the whole path, because again, our job in the monitoring, you know, you don't, again, you don't do monitoring for the sake of doing monitoring. You do [00:34:00] monitoring to collect data, to answer questions and to reduce problems.
[00:34:03] So when you think about meantime to repair, which is the majority of our, of our customers on the SRE side are using too. To measure themselves. Meantime to repair is detection, identification, escalation, et cetera. And so you need to detect the problem, but you need to identify as quickly as possible, who, who you need to call, because if you, if it's Akamai or Fastly or somebody else, or it's Equinex, you need to figure that stuff out very, very quickly or an ISP.
[00:34:35] Like if level three blew up a fuse, you need to know that as
[00:34:38] Matt: So to your customers, you're not only providing sort of end point alert. Like, Hey, the users in this area are probably having a bad experience. You're providing me some diagnostic tools to figure out what the root cause of that bad experiences.
[00:34:52] Mehdi: Absolutely. So we, you, you, you nailed it. So we, we, we show the Ariens data, which is the high level top of the [00:35:00] iceberg stuff. There you go. And also the diagnostic data, and last year we added the BGP capability as well, because that's the, that's. The other thing is like how, what are we seeing on the network?
[00:35:12]so we want to provide as much diagnostic because again, if you want to deliver stuff in milliseconds, you also now need to troubleshoot in milliseconds.
[00:35:20] And give answers. Why, why things are broken
[00:35:23] in milliseconds?
[00:35:24] Matt: Yeah, that's really interesting. So you mentioned, autonomous vehicles or vehicles in general, I think as one of the customers creating demand for these, these millisecond level, user experience measurements that you're introducing, what other industries and use cases are you seeing down at that, you know, out at that cutting edge?
[00:35:44] Mehdi: So a lot of, warehouse automation, robotics. Yeah. And the machine you, you said it earlier, right? She's talking to other machines. They need the answers very, very quickly. So we're seeing a lot of that there, we're also seeing a lot in the gaming [00:36:00] industry. Like some of these games are just, And again, especially with COVID.
[00:36:04] What, what, what we're seeing is gaming is becoming a social place form where people are interacting, where people are getting together. They can't, they can't hang out in a bar. So they're playing Fortnite or goggles. What does that mean? Playing? And, Sony is launching a new platform. Microsoft is following.
[00:36:23] So gaming is big. streaming is big. so again, with these are the things that we're seeing the most,
[00:36:33] Matt: Got it. this is kind of an out there question and, and, you may or may not have an answer, but I'm super curious. So, I mean, so catapult was founded in 2008. Is that right? Okay, so, so is there anything that Catchpoint has, has identified that was previously unknown or not Le not widely known that it was kind of like structural, like something that just needed to be fixed in the internet or across lots of companies where you just have this like [00:37:00] general insight from, you know, one customer's experience or a couple of customer's experience.
[00:37:04] Is there anything that stands out.
[00:37:07]Mehdi: having a single vendor is no longer an option anymore
[00:37:10] Matt: what does that mean? A single vendor is no longer an option.
[00:37:15] Mehdi: single CDN vendor having one yeah. Yes. That is, I think the, the thing I'm the most, I'm the most surprised by how fast that that movement has taken off, before it was reserved to a few large, super sophisticated companies.
[00:37:34] Like I think the first time I saw multi CDN was at LinkedIn, which is like on the cutting edge of user experience. But today it's literally everybody I talked to is doing multi CDN multicloud. I have one customer that doesn't use. One cloud vendor. They use five clouds vendors. They have a cloud vendor for the middle East.
[00:37:53] They have a different cloud vendor for China.
[00:38:00] [00:38:00] Matt: Now did it, do they use, data from Catchpoint to negotiate, renegotiate their contracts or, or like enforce SLS?
[00:38:07] Mehdi: SLS. Yes,
[00:38:08] Matt: Yes. I can say that's a really neat,
[00:38:11] Mehdi: Our data, our data can be an enforcer, but I think what they do with the data mostly is like, Hey, we're going to make a decision in China. We'll use Ali cloud versus. AWS for example, or Microsoft.
[00:38:30] Matt: Like Amazon could go in and add a new fiber route and suddenly like, yeah, really. Do you do any general monitoring yourself? That's not tied to specific customers to derive sort of insights like that.
[00:38:44] Mehdi: Yes we do, but we try to stay away from publishing some of that data because, I have, I have a strong belief about a benchmark and publishing public benchmark. I, in my opinion, I think if you're number one is great, if you're number two and three, it [00:39:00] sucks. So I would rather, so we do benchmarks, but we do them privately.
[00:39:05]we are, I don't think it serves anybody's purpose to, you know, put the number out there and say X, Y,
[00:39:11] Matt: you're a gardener and selling reports.
[00:39:14] Mehdi: But I don't sell data. Right. I would rather work with companies to improve things. I know it's just a personal belief.
[00:39:20] Matt: interesting. I, I, I that's a, I, I like that. You're opinionated on that.
[00:39:24] Mehdi: I, I do. I mean, I fight with my marketing team all the time because
[00:39:30] Matt: immediately thinking, Oh yeah, we could publish all that. We can. The state of the
[00:39:32] Mehdi: yeah.
[00:39:32] The state of the internet. We, we have, we have network latency, data, et
[00:39:36] cetera. I really think it, it sucks. I, it doesn't serve anybody's I don't know. you know, I told you, we start, we
[00:39:45] Matt: headlines, but is it really useful? Is it really useful?
[00:39:49] Mehdi: friend and then you make 2000
[00:39:51] Matt: sure. And there's always edge cases. There's always like, well, you tested the wrong thing. That's a big problem with benchmark is why that as well as got a hundred different chip benchmark.
[00:39:58] Mehdi: I mean, if you asked me to [00:40:00] benchmark the top websites, like the top eCommerce websites, right. It's in the numbers don't make any sense. I will tell you that the Apple is the fastest website. Yes. Apple is the fastest website, but you know
[00:40:12] why
[00:40:13] yeah, no, no. Apple is the fastest website, but you know, you know why they don't have any third parties on their pages.
[00:40:20] They are. Right. So does it compare apple.com to amazon.com? Come on. It's not the fair comparison, but people, somehow people want to see that, but apples, you cannot, you know, if you're going to do benchmark benchmark, apples to apples. So otherwise yeah. It becomes a food solid and, and it's, it's not great.
[00:40:42] And the other thing about benchmark that again, I have a strong opinion. We started the company in 2008. You know, those banks that were so big, they could have never failed.
[00:40:56] a grades. Right?
[00:40:58] Matt: yeah, yeah. You could end up with a [00:41:00] black guy because you missed something and
[00:41:01] Mehdi: Exactly.
[00:41:06] Unless, unless as an industry, we come together and we come up with the benchmarking standard that says, this is I'm going to benchmark AWS versus Azure versus X, Y, and Z.
[00:41:17] versus that e-commerce,
[00:41:19] that there is no
[00:41:19] Matt: Sure. And even that would be contentious. yeah. That's, that's, that's interesting. That's interesting. although there is a general, you know, it's it, there is, there is some generalized information that I think. People are interested and I'm certainly interested in, and it's like, what is, what are the overall trends in the internet?
[00:41:37] Right. And actually, I mean, can you, you must have a spidey sense of what's happened since 2008. What are the big, what are the, what have been the big inflection points in terms of like, Meaningful improvements to user experience. Granted that's a vague description, but you know, I mean, when, when you, when videos stopped buffering for me, high Rez video stopped offering me.
[00:41:56] That was a, that was an inflection point. Are there other inflection points? Because you're [00:42:00] closer to this that you've seen.
[00:42:01] Mehdi: I think one of the biggest inflection points I've seen was the dine outage, the DNS dine outage of 2000, I don't know, a few years ago, three, three, four years ago. Right. And that was, an incredible thing. So a few things happened again, resiliency where people had one DNS vendor, overnight, people had two or three myself.
[00:42:26] I went from having one DNS vendor to having three. Because I, we, I said, listen, if I'm the monitoring business, I can never, I can't fail. Right. So, so that was, that was a huge thing. The other one I've seen around DNS as well is a lot of companies realize that DNS is more important than, than ever. because it's your first experience as a, as a, as a user when you type a domain.
[00:42:52] Right? And so a lot of people try to do that at scale in house. And one trend I've I've seen is like, people are [00:43:00] starting to rely on companies like an, a Swan today or dine and, and others. So that's one trend. Again, the multicolor, the multi CDN is like, Taking off big time. I think another trend that I've seen is this concept of, observability being able to collect as much data as possible, but to answer questions and to be able to tie or connect the dots.
[00:43:23]we're seeing some companies doing amazing job, connecting the dots and answering those questions like what's broken when, why, et cetera. And then obviously the cloud adoption. Right. I never, I didn't think, I didn't think it would be that, that, popular, but, that has been an incredible the cloud, like AWS and Google GCP and, absolutely.
[00:43:47] Right. The only trend unfortunately that I have seen improve is the ability for at least in the U S for us to get better at high speed internet from homes. That's unfortunately, one trend [00:44:00] that I'm very concerned about when in Europe and Asia, there is a lot of innovation. There is no internet access or very limited.
[00:44:10]I mean, I get like, I'm lucky I have one gig down from spectrum, but I only have 20 Meg up. Now try to have four family members doing four zooms at the same time. It gets a little bit tricky and that's the maximum I can get. so the digital divide, I'm a little bit worried about, to be honest, math is how, especially in.
[00:44:32] You know, we've covered, people not being able to go back to schools in the Bronx, why what's going to happen to those kids? How are they going to have access to the internet? so the digital divide, I'm a little bit worried about because we're going to leave a generation of people and kids behind.
[00:44:47] If,
[00:44:50] Matt: Yeah, that's a, that's a really important trend to pay attention to. And, you know, I've, in some ways I have [00:45:00] reasons to be optimistic about that, you know, coming from an infrastructure provider because a lot of. What we, and I mean, collectively as an industry, not vapor IO necessarily, but we were collectively building at is a lot of infrastructure to support this, this third.
[00:45:16] Back to the internet, which is billions of devices generating zettabytes of data 24 seven. So, you know, the old asymmetrical internet, I mean, it's kind of cute that you only have 20 megabits per second up, right. That doesn't make any sense in today's world. It made sense a long time ago when we were primarily consuming content, but like you and I we're, we're generating, you know, a 10 ADP video feed, both directions.
[00:45:41] And like you said, you've got four people in the house doing that. But, but a lot of infrastructure being built and, you know, even in rural areas, you know, you're seeing like precision agriculture driving, fiber deployments and data center deployments in, in rural America. And I would hope, and I don't have any reason to know that this will will not happen, but I hope that that [00:46:00] infrastructure would create new, new economic equations for
[00:46:06] Mehdi: I hope so. I mean, that's what I'm hoping, because we also have a lot of people that are good with coffee that are moving away from cities that are, I mean, I have a lot of my employees that moved as far as Montana back where their families are from and the internet is not that great. So I'm hoping that, you know, this trend is going to,
[00:46:31] Matt: Well, and so you, you must actually see that, right? You must actually see that. Look, I don't, I don't care how much CDN or edge, edge computing you do in this area.
[00:46:48] Mehdi: That's the weakest. So we have to fix two things. Math we have to bring low-latency. So meaning more data centers, more compute capabilities as close as possible to [00:47:00] these end users. But if most of these end users are not driving a Tesla car in Santa Monica, but now you're dealing with the John Deere tractor.
[00:47:09] In the farm in Montana, how are you going to deal with that low latency, especially when people are going to move back. So we have these two spectrum where we need to bring low latency and more
[00:47:26] Matt: Yeah. Yeah. do you, do you have an example of a customer of yours that has gone from a, a non, I mean, edge computing is really early, so the answer may not be as, or maybe you can't talk about it. but a customer that has implemented some, you know, third act of the internet type edge computing, where there is compute storage.
[00:47:56] Mehdi: We have one customer that has done that, but they have the, [00:48:00] the technical capabilities because they own a lot of real estate. And so they've built their own edge compute, kind of, solution. we are talking to many that are exploring various solutions. mean we're very early on. and the other thing that is, I think, causing a lot of delays and confusion is the whole 5g thing.
[00:48:20] Matt: Yeah, I want to talk about that.
[00:48:22] Let's talk about five G.
[00:48:23] Mehdi: I think that's creating a lot of, confusion between the different, protocols or standards and who has what and when, and, and differentiate image in the marketing fluff versus reality is a, is causing a lot of angst among people. But,
[00:48:40] Matt: Yeah. So let's, let's talk a little bit about, about, about mobility. So I imagine measuring user experience
[00:48:48] changes when a device is in motion and
[00:48:50] Mehdi: Correct?
[00:48:52] Matt: Tell me how Catchpoint, looks at this.
[00:48:54]Mehdi: So we have a wireless capabilities, so we have about 40 in the U [00:49:00] S locations, but they're fixed that they don't move 40 locations that are, for G. Yes, correct. So they are literally connected to a Verizon at and T T-Mobile sprint three G four G and whatnot.
[00:49:15]then we, because we have a real user monitoring capability, we also can collect data from every single user interaction with a website or mobile app. So we have, we have that perspective, but it's not synthetic, it's passive. It's after the fact we collect that data and, we we've been trying.
[00:49:34]for the past God, two, two years since five G has become quote unquote available, or people talking about it to work with some of these companies to give us access to 5g, capabilities. But. You know, it's not available yet. and when we talk to, let's say a rep from any of the major ISP or providers,
[00:50:00] [00:50:00] Matt: Yeah, I have, I have a Verizon five G phone and
[00:50:06] Mehdi: right, Yeah, no, it's, Yeah. Yeah, no, it's a it's. I mean, I have a bunch of them here, like a bunch of, hotspot systems
[00:50:15] from all, all three of them that are supposed to be five G I have a huge antenna on my garage, but still no 5g reception
[00:50:23] Matt: Yeah, well, and clearly that's going to change. It's just a matter of when. So how, how, I mean, w how do you see the world changing with five G and, you know, not necessarily directly related to your business, but maybe to the businesses of your customers and your need to monitor.
[00:50:37] Mehdi: Right again, this is what's going to require more low-latency stuff. You're going to open a fire hose of data capabilities on a mobile phone. I'm going to, I mean, I can see my kids streaming, like not just one Netflix. They get to be watching Netflix, Hulu. Disney and, and peacock at the same time. So the data consumption just keeps on, on, on, on [00:51:00] increasing, right?
[00:51:00] So again, the low latency stuff is going to be more important there. So that's how I see things. Again, more data, more consumption,
[00:51:15] Matt: Got it. And when you think about edge computing broadly,
[00:51:19] Mehdi: Yup.
[00:51:30]
[00:51:30]being able to put stuff under a tower, a cell tower. I think that is an interesting approach, I think, in very dense areas. Buildings, being able to stuff, compute systems in a closet somewhere. yeah, I, I think it's like how, I mean, if, if there was an ideal solution where I literally, I can ship a little, Catchpoint one of our Catchpoint, appliances and put it anywhere everywhere I want.
[00:51:57] I think that would be unbelievable. [00:52:00] Yeah, there is that there are bugs. no, no, it's a, it's an Intel it's based on Intel, not platform, but we're trying to shrink it to literally have one of these models. It's literally a tiny device, inch by two inch, where we want to, but that means very limited capabilities, right?
[00:52:20] The compute is so small. but I think we, we, we have different
[00:52:24] appliances.
[00:52:25] Matt: time and phone
[00:52:27] Mehdi: Exactly. Exactly. So it depends like if you want you to do like heavy duty browser stuff, you need, you need big processors, but you know, if you're doing API
[00:52:36] Matt: send, like, like Nielsen, we had like Nielsen families. You might
[00:52:39] Mehdi: Yeah.
[00:52:42] Right.
[00:52:43] Matt: download this app to your phone,
[00:52:45] stick this on your
[00:52:45] Mehdi: So it's funny you say that. so we have a network, they have about 150 of these running at people's houses. and we call that our last mile network, where we basically pay people to host one of our Catchpoint. no, it's not, [00:53:00] it's not bad. I it's a it's, but it's actually production.
[00:53:03] And what we do is we subsidize their internet connection for the year. So we
[00:53:07] Matt: Super cool.
[00:53:08] Mehdi: bucks a year to to have that. I mean, the craziest place we've put Catchpoint is on an airplane.
[00:53:15] Matt: Oh,
[00:53:16] Mehdi: and, that was an awesome thing.
[00:53:19] Matt: Last 10 times I flew.
[00:53:21] Mehdi: yes. but this one was, it was a big company's CEO of private jet.
[00:53:26]And, so we had like three Catchpoint nodes. And what was amazing is, you know, the internet. No, no, no, no, no, no, no. Jet, if I had the jet, I wouldn't worry about latency. I would have other things, but, but, this company has an interesting problem with their satellite communications. So we were running like trace route data as the plane flew from back and forth.
[00:53:50] It was fascinating. So again, can, is a, is a problem everywhere.
[00:53:55] Matt: the, the, the, you know, the internet of flying
[00:53:59] Mehdi: Right.
[00:54:00] [00:54:00] Well, I learned, I learned the hard way you conflict cats. You can't put the cat five on a, on a plane by so that we didn't know we to learn that the hard way you can't have it. Yeah. So it has to be wireless wifi, only those kinds of things.
[00:54:14] Yeah. Yeah. It's very interesting. But every time we shipped a new device where to get their favor approval, it was, and then the second craziest place is on the buses. So, so some of these companies, as you know, in California and whatnot, have buses to shuttle their employees. So we had, like, we had a fleet of, this company's buses and we had, Catchpoint making sure that the employees were getting a good experience on the wifi, working from San Francisco to wherever
[00:54:41] Matt: Do you have a sense of the scale, the amount of data you generate per second or hour or day? A lot. Well, so, so yes, I won't make you reveal that. but a lot. are you or any of your customers using machine learning too?
[00:54:59] Interesting. Can you [00:55:00] tell me a little bit about that?
[00:55:01] Mehdi: well so much for machine learning to really work, Matt, you, you need a lot of data and you need a lot of contextualized data, meaning that you need a lot of data sources. I have one data source. It's not, I mean, we, we do some machine learning on our end, but it's not very interesting.
[00:55:17] Right. But let's say, imagine you are a company and you get like, Telemetry from Catchpoint you have your network, telemetry, your server, telemetry, your IOT devices, telemetry, et cetera. This is where machine learning can become really fun because now you're absorbing a lot. You're learning, you're detecting patterns.
[00:55:35] And so we have a lot of customers they're doing that. They build their own things or they are using. You know, companies like Splunk and others, to be able to, to analyze that data at scale and try to build models. But
[00:55:50] Matt: It seems, it seems absolutely inevitable, right? Because you've got, you've got these real time devices. some of them are, you know, safety, critical life critical, or at least dollar [00:56:00] business critical. And because everything's happening in milliseconds, the detection and the remediation has to be a milliseconds,
[00:56:08] Mehdi: Oh, yeah.
[00:56:12] Yeah.
[00:56:13] Matt: Like a human could go figure out what's going
[00:56:15] on and you're not going to have that option.
[00:56:17] Mehdi: math. I have some customers, they deal with code blue stuff at the hospital. So the, the, the criticality of monitoring, some of these things is, is, is, is insane. I mean, if we screw up, you, you can see the consequences, right? So, so these guys are dealing in milliseconds and not dealing about the cart being abandoned.
[00:56:39] They're dealing about like somebody. Having a heart attack and dying out of it. So, so the monitoring and, and, becomes critical. I mean, again, critical if I'm the CEO of
[00:56:59] Matt: Yeah. Yeah. [00:57:00] so, so many as you, as you, as you look into the future. Right. and you know, we're on this precipice where I certainly feel it's and I sent you, you feel that, that, that, that edge computing in the way that we've just been talking about it, which is compute right at the edge of.
[00:57:15] The network supporting millions and billions of devices and low latency. that's just starting to become reality. You know, like my company's in four U S cities today and we're going to be in 36 and in 18 to 24. but if you look out in the future, let's say the next, next 18 months, there was one domino that you could personally topple.
[00:57:35] Like one thing would just,
[00:57:37] you know,
[00:57:38] get this whole industry. What would be the one thing that you would push on?
[00:57:48] Mehdi: I think, I think we're in the middle of it. Is this pandemic. I think this pandemic has, has shown. the need for better data, better network connectivity, better end user [00:58:00] experiences. Low-latency I mean, if we're going to do videos with 6,000 people with 5,000, with 200 people, I think, I think COVID is a digital accelerator that.
[00:58:13] No industry has ever seen. I mean, I wasn't born, I wasn't born in 1918, but I'm sure there was a lot of innovation that came out of that pandemic. I think we're in the middle of the, the biggest kick in the butt. We all need to make sure that we innovate and, and deal with this challenge. I mean, we are, I mean, I don't want you to be on the pessimistic side, but I don't know when we're going to have a vaccine or to deal with this. know when we're going to go back to whatever normal is going to look like.
[00:58:49] Matt: Yeah. You know, it's, you know, it's, it's, one example of this that I just have seen in the last few days. you know, virtual reality has like, like every year for the past 10 years has been [00:59:00] the year that virtual reality is suddenly going to become, you know, mass market.
[00:59:03] Mehdi: year. Now this is,
[00:59:04] Matt: so. I think so there there's a, there's a guy that's quote, unquote, driving himself across the country on
[00:59:10] Mehdi: Yeah. Right.
[00:59:13] But the education.
[00:59:14] Matt: throw on an Oculus and you go visit the, the, you know, Paris it's.
[00:59:19] Mehdi: but, but that's static. I'm talking about virtual reality with kids in a classroom. Because my kids are not going back to school anytime soon. Right. We just got noticed that until next year or 2021, but need that social interaction and, you know, yes, you're on zoom. It's kind of cute at the beginning, but that zoom is not the classroom.
[00:59:44] Matt: the whole deck on the enterprise.
[00:59:46] Mehdi: Yeah, so, but, so I think covered is a, is a kick in the butt for all of us to innovate and bring solutions. I mean, even like zoom you're seeing. So, on the video conferencing, so many companies that have been innovating the past [01:00:00] few weeks and few months like to bring different kinds of video, video conferencing solutions to.
[01:00:07] To deal with fatigues and absolutely. I mean, Microsoft, Microsoft is announcing this new thing in teams where you can see
[01:00:15] Matt: Like an, a, an audience. Yeah. Yeah. Very cool.
[01:00:18] Mehdi: So I think COVID is here to make us better. We just need to hurry up a little bit.
[01:00:23] Matt: That's great insight. So, Maddie, thank you so much for spending time with us this morning.
[01:00:28] how people can find you online and how they can find Catchpoint online?
[01:00:31] Mehdi: Yeah. Sure. So catchpoint.com. LinkedIn, Twitter Catchpoint or,
[01:00:38] M Dougie MDA O U D. I work at Catchpoint Catchpoint in one word looking forward to connecting. Yeah, absolutely. [01:01:00]