Dial Up Motive
Presented by: Dan Hollinger, David Belson
Originally aired on July 16, 2023 @ 5:00 AM - 5:30 AM EDT
Human-interest segment asking Cloudflare employees what their first Internet experience was and how it informed them joining Cloudflare. Dial-up modems, bulletin boards, punch-cards, Twitch, Twitter and more.
This week's guest: David Belson, Head of Data Insight @ Cloudflare
English
Transcript (Beta)
Hello, hello. Welcome everyone to episode 26 of Dial Up Motive. I'm your host, Dan Hollinger, and with me today I have David Belson.
On this show, if you haven't caught it before, we like to dive into early technology and Internet experiences with Cloudflare employees.
So, David, thank you for joining me. Thank you for having me, Dan.
Awesome. Well, to kick things off for those catching either the live stream or one of the recordings, can you give us a taste of what work you're doing here at Cloudflare?
Absolutely. So, at Cloudflare, I joined at the beginning of January of this year, and I am the head of data insight on John Graham-Cumming' team.
What that means is my responsibility is helping to communicate what's happening on the Internet out to the general public through tools like Cloudflare Radar.
And ultimately, the larger goal is to make Cloudflare the trusted source of knowledge about Internet trends.
So, when something happens on the Internet, we want people to turn to Cloudflare and say, okay, what are they seeing?
What do they know? What can they tell us about what happened?
And what are some of the things that you think are unique to Cloudflare that help with some of those data insights?
Is it sheer volume traffic?
Are there any other things that allow us to hold our own against some of the other sources of insight for Internet traffic?
Yeah. So, I think the breadth of the portfolio, I think, is definitely unique.
There's a lot of different services throwing off data exhaust.
I think that, obviously, we have the CDN business, which I think provides us some good insight into what's going on from a traffic pattern perspective.
But the Quad One Resolver, I think, gives us some really unique perspectives on how people are using the Internet in aggregate fashion, of course.
And I think the rich set of security services, I think, also give us a really great perspective on what's happening from a cyber attack perspective, but at multiple different levels.
And then, yeah, you've had quite a fascinating career so far across both Internet and data.
Cloudflare, Fastly, Dyn, in your past, this is where I'm fascinated to learn more about some of those early days as you were starting to get involved with technology and the Internet.
And happy to transition back into one of those stories if you'd like to kick us off.
Sure. Sure. I mean, I moved up to... So, I live in the Boston area now, and I moved up here in 1994 to go to grad school at Northeastern, going for a master's in technical writing.
And as part of that, you ultimately had to do a master's project, prepare something.
And so, I was looking around for a job that spring that I could use to gain experience and probably build a master's project on, and found an ad for BBN.
So, this was the company that had built out the original ARPANET and had been very instrumental in building and managing the networks that ultimately became the foundation of the Internet.
And they were looking for somebody to support their emerging web hosting business.
So, this is back in the spring of 95 when web hosting was non -trivial, to be honest.
I mean, it was easy enough to set up your own NCSA server and put something up.
But if you were a business or an enterprise, you probably didn't have the skills or the connectivity or the ability to do it.
So, BBN set up a dedicated web hosting service. They were looking for sales engineers.
And they had listed... The job description of the four or five requirements.
And I looked at it. I said, well, I've done those few, and I've always wanted to learn Perl.
So, I applied, got the job. And I sort of had discovered that they were a T-stop, a train ride away, or a bike ride away.
So, in my head, that was like, wow, the company that really built the ARPANET, building the Internet, is so close to where I'm living.
So, I got the job. And I remember after the first couple of days of work, I came home, and I was all excited.
My roommate, they're paying me to talk about the web.
And at the time, I couldn't imagine a cooler job.
So, I did that with BBN, and then ultimately GTE, because they got acquired for four years.
When I was... The last few months, I was... Yeah, out of curiosity.
So, in a company in that position where they're trying to hire for someone, and really there's not that much technical depth of...
You can't basically ask for five years of experience when the technology has only been around for three.
How did you find the interview process during that time, if you remember?
Or once you were on the other side of it, what skill sets were you really trying to look for?
Was it a generalized coding and network, like hardware level networking? What things helped translate into knowing that that person would be successful with such a new technology?
So, I think translate is the right word there. And that's one of the things that I love about the sales engineer roles that I've had, has been that it's largely about translation.
It's about going to talk to a customer, and understanding what they want, and need, and then translating that to the product engineering folks.
And saying, hey, we don't have this thing. The customer wants to do this thing.
We need to build a feature that does that. And then having the engineering product folks saying, well, given the descriptions of what might take, or the timing, or whatever, and being able to translate that back to the customer to help them understand how we'll be able to help them their problems.
From the interview perspective, I don't, I mean, yes, to your point, there was no, hey, let's dig into five years of the web, and get to details there.
I do remember being interviewed by a couple of folks who had been at BBM for a while.
And they were, as we say around here, wicked smart. And they treated me as an equal.
And as a 22-year-old, 23-year -old, that made a really big impression on me.
You know, it said, hey, we're really smart, but we're not going to try to show that off.
And so, I don't remember what we talked about specifically. I think they had asked me, they were like, oh, we're working on this thing.
And you know, how would you do this?
And I thought about it, and I gave whatever answer I gave, and probably it was okay.
So, I think probably at the time, it was just more a question of like communication skills, you know, the ability to understand technical concepts.
But they weren't, yeah, they weren't going deep into, I mean, I guess even if they had gone deep into how web technology worked, there quite frankly wasn't a ton of it at that point.
So, I probably would have held my own fairly well.
And prior to that, you know, as you were getting into that job, were you, you know, one of the types of people that was already dabbling in code at that point, in kind of early Internet technologies?
Where was that, you know, aspect of your journey as you were kind of new enough to get that kind of job on your radar?
Yeah. So, I had actually spent the prior summer being, I guess, the webmaster, that's an archaic term, but being the webmaster for my alma mater.
Graduated and then got a job in the computer center that summer helping them set up their first real website.
So, got to, you know, learn about Linux and, you know, installing web server and configuring the web server and learning the importance of backups, because I screwed that up pretty badly at one point.
So, I had experience with, you know, web technology prior to moving up to Boston.
And, you know, as you jumped into that first job, what were some of the lessons you took into your next phase of your career?
Let's see. I think, well, I think, I mean, one of the things that, one thing that I have been impressed by that I think I was sort of exposed to at BBN was, that I've taken along with me, is the, not everybody needs to have the same background I have.
You know, not everybody needs to have a computer science degree to be successful in the industry.
When I was at BBN, I worked with folks that had theater degrees.
I had a coworker who was a geophysics major, but had had experience administering the student Unix system at Brown University.
So, he had hands-on experience there. So, it was just, I think one of the things that has been interesting has been that I get to work with people over the last 26 years that have these different varying backgrounds and that bring unique perspectives and different perspectives to the problems at hand.
So, it's not all about how do we solve this in a computer science way, but it's, you know, about how do we solve this from a medical perspective or a theater perspective or, you know, a mathematician perspective.
Yeah, I mean, I think that goes way back to kind of that concept of translating.
So, if you are coming from multiple levels of experience or venues, you know, you're gaining those analogies to be able to translate, you know, either technical to non-technical across multiple audiences.
And, you know, it's very fascinating to see that and particularly seeing that kind of broad base in, you know, Cloudflare employees and a lot of employees across tech.
Right. And it's also, I think, helpful from a conversational perspective where, you know, you need to be aware that you can't just go deep into the tech, that if you're in a room with folks who have mixed backgrounds, you know, it is important to be talking at that, not dumbing it down, but to be explaining at a level where it's more generally understandable and not just, you know, it's super in the weeds with the code.
And would you kind of advise those that are looking to get into tech to kind of find one of those side projects or find, you know, one of those volunteer roles, assuming they're still available and, hey, go build a website or go manage some hosting for someone?
Do you think that's still applicable?
Absolutely. Absolutely. I think the hands-on experience is amazingly important.
And I think when I look back, that's one of the things I wish I had more, to be honest.
I think that the undergraduate work that I did was very theoretical.
So we didn't have the opportunity, or I guess in the early 90s, maybe I didn't know about the opportunity, but, you know, at that time, I feel like I got out and didn't have the opportunity to have had administered systems and servers and configured networks and done all those things that I think the people I ultimately wound up working with had had some experience with.
So, yeah, these days, you know, it's so easy to do.
It's so easy to set up a home network.
It's so easy to build a website on, you know, Cloudflare Workers or, you know, any, you know, pick your platform, quite frankly, you know, write JavaScript.
There's so many things I wish I had done a better job of keeping up on over the years.
And speaking of things getting easier, you know, as you were starting your career in tech, was it very much in the realm of, you know, you had to stop by the library to pick up a networking book, you had to go to actual bookstore, you know, the accessibility of information or the ability to learn was still pretty constrained or was it at the phase where, yep, you could find an obscure website if you looked hard enough on Lycos or Ask Jeeves?
Where were you kind of in that phase?
I think the answer is yes. So, I think there was definitely, you know, I still have my, I'm trying to think which library they're in, but my, you know, the O'Reilly books were invaluable.
You know, you want to learn something, go find the book with the animal on the cover.
And, you know, you would learn how to program from that.
You'd learn how to set up a web server or a DNS server or whatever it was. But again, the great thing that I've found also throughout my career is the coworkers.
You know, you have a question, usually the hardest part is figuring out who do I ask?
Do I ask Dan? Do I ask Paul? Do I ask, you know, whoever, you know, but as soon as they have pointed at somebody, generally they're very willing to help.
They're very willing to help you figure out what you want to know.
So, I think a lot of times you go in with a question that is not the actual question you want answered.
So, yeah. So, there's the books. I think back then it was definitely, because we were all learning at the same time, quite frankly, you know, in the mid -90s.
So, there weren't a ton of books, but there were a few. Coworkers were great.
Occasionally you'd find the websites that had information. Usually, I think more often than not, I was looking at this earlier, it was more Usenet news and Usenet news groups.
So, you'd go in and you'd say, you know, and I did this a lot.
I was responsible for the web server for a Datamation magazine, which I don't think publishes anymore.
But they used a, it was called Waste Wide Area Information Search. So, sort of think of it as sort of an on -server Google, if you will.
But I was compiling it from source each time and configuring it a certain way and whatever.
And I would keep running to these problems.
So, I'd go post on Usenet news and the appropriate news groups and say, hey, I'm trying to do this.
I'm getting this error.
I've got to configure it this way. And, you know, what am I doing wrong? And inevitably somebody would come back and, you know, explain what was the right way to do it or explain what the problem was or go, oh, you found a bug or something like that.
I was looking back and realized that I probably have, you know, Usenet posts that are older than some of the folks I work with, which is a little bit weird right now.
But... That's fascinating to think through. Yes. Especially given how permanent many things on the Internet are to be able to say, oh, yeah, that post there is, you know, predates some of our newest employees.
Yes.
So, given that, you know, breadth of experience that you have, what advice might you have for some of our viewers that are in a similar spot in tech, you know, as you were starting your journey, what did you find was most helpful to grow in your career, both within a company or, you know, within some of your earliest companies and kind of across different skill sets cross-functionally?
I mean, I think I have found that I've always been sort of a mile wide and inch deep.
So I know enough to be dangerous.
I know enough to have a conversation and to help figure out what the right questions are.
So I'd say, you know, having that broad perspective, I think is valuable.
But I'd also, you know, honestly suggest, you know, people find an area of interest, an area of specialization to go down that road.
I think one of the other things I would certainly suggest would be finding a mentor or a, you know, somebody that can help guide you, not guide, I don't know if guide is the right word, but sort of help guide you through the tech world.
Whether it's, you know, whether it's within the given company or somebody say, hey, you know, here's this new technology that, you know, I just read about, go, you know, go check it out.
I, when I was, I haven't done a good job of doing that throughout my career.
I wish I had done a better job of finding those people, you know, finding the helpers as Mr.
Rogers used to say. But, you know, I think, I think those are, you know, finding the right people to work with and to, to look up to.
Yes.
I guess either emulate or kind of position as a role model of, oh, I want to be.
That's the right way to put it. I want to be solving those types of problems in five years, 10 years.
Like how do I, how do they, again, then learn how to accomplish that via, you know, that mentor?
Exactly. Yeah. I think that for me, I did have a little bit of that.
I think, you know, as I've come through my career where early on seeing folks in a CTO role, chief technology officer, and I've seen that the role sort of splits in the industry.
One is basically a VP of engineering.
If you will, the other is more of a sort of a polymath, you know, sort of, they understand the technology, they understand the business, they understand the impact on people.
And that was the direction I kind of have tried to go in is, is to, you know, not just be focused on one area, but to really try to understand something holistically.
So, so again, you kind of want to be like a Google search where you can search across plenty of fields, but you're only ever going to click the first two links anyway, and you're never going to click into the more obscure things.
And kind of given your career journey, at what point do you decide, you know, it's time to move on?
You know, I've, I've learned everything I could here.
You know, it's, you're, are you intrigued by a new technology and, you know, given your span, especially on the Internet, you know, what pushes you to kind of that next frontier?
Good question. I mean, it's funny because I think when I first saw the web, Other than the, other than the paychecks.
Paychecks are important.
No, when I, when I first saw the web, I was largely unimpressed, but this was also like 1992, I guess.
So it was like the certain line mode browser and, and sort of, it was very clunky, very hard to use, obviously all text.
And I kind of played with it for a little while.
I'm like, this is just going to go anywhere.
But as things evolved, as Mosaic came out, you know, I recognized the value in that and started playing with a lot more in, you know, 93, 94.
So I think that was what sort of pushed me in that direction.
You know, having exposure to the Internet at college, and then pushing me into to getting into the industry.
In terms of moving on to the next thing.
I think that, I mean, I've really been largely in the CDN space since 1999.
But I think having said that, it's evolved just amazingly since then.
You know, where it started out as, you know, making, you know, when everybody was on dial up in 99, helping to make websites faster, helping to scale, you know, expanding to streaming and security and so on.
But I think about now, let's say about 15 years ago or so, sort of realizing just the, the amount of data that was being generated, you know, as these services grew and expanded, and being able to take that and take that data and do interesting things with it.
I think so that's where over the last 15 years or so, my career has has moved in that direction.
And yeah, it's I mean, the entire over the past few years, the entire kind of data science industry of, you know, as we're capturing all this data from a log perspective, as we're moving all this traffic through and getting what kind of a meta analysis can we then perform, not to mention, throw into a machine learning algorithm and optimize on whatever criteria you're after, you know, it's a fascinating to see how that will continue to evolve.
And I know, in the Cloudflare, we use our own data in a few ways, you know, across our security products or plot management, and, you know, even our performance routing.
Right. Yeah, I think, you know, for me, I originally had gotten into it, looking at connection speeds.
So, you know, being able to calculate effective connection speeds, from the the CDN data exhaust, the log file, the aggregate log files, and realizing that that had a real world, it had a real world impact, to be able to publish information on that, and have people look at that and say, well, you know, we've been investing in improving, we've been investing in broadband, in this given geography.
And yet, you know, this report says that, you know, speeds haven't improved, or look, they've have improved, and this is an ROI.
So there was a lot of, I mean, that was super interesting to have the opportunity to build a report around that.
You know, it was also interesting to use that data to understand the how real world events were visible through web traffic data.
So looking at outages or shutdowns in a given country, I'm looking at how, when I think it was the first Obama inauguration, looking at aggregate traffic for media customers, and watching it climb through the through the inauguration, and at the same time watching commerce traffic drop.
So you can kind of see how people are shifting away from shopping online, they're watching the inauguration.
And then when it was over, you could see basically the opposite trend where media traffic dropped and ecommerce traffic picked right back up.
Yeah, I mean, that's almost fascinating is, you know, given more and more people coming online, the Internet becoming more globally accessible, being able to kind of notice that echo across the entire Internet of here's a large swath of activity, you know, around large political events, sporting events, that then alters their behavior of how they're using the Internet, you know, how quickly they're solving their wordle, like all of those things, you know, coming into play and casting off that exhaust.
Right, exactly. Yeah, exactly.
And then, and then being able to break it down also by, you know, not only geography, but also looking at, you know, mobile versus fix the trends, that kind of usage, trends in, in, you know, user agents, browsers, you can get some, you know, when you start slicing and dicing by some of the different other lenses, it becomes a really, you can start telling some really interesting stories.
So I'll use that as a wonderful segue.
So speaking of interesting stories, you know, what are you most excited about for the future of the Internet and the future of your work here at Cloudflare?
So I think there's, there's just a tremendous amount of data that, you know, I haven't even scratched the surface of trying to do something with yet.
So I'm looking forward to Cloudflare to working with, you know, everybody to figure out how do we take the data from all these different services that we have launched, all the different services we're planning to launch, and, and telling the stories with them.
And how do we, how do we get people to understand what's happening on the Internet from those perspectives?
In terms of the industry of the Internet, I think, I think the continued ubiquity of connectivity is really interesting.
You know, I look back to at a prior job when Wi-Fi was just starting to happen, and had to go get VP approval to get a Wi-Fi card for my laptop.
And, you know, justify because I was spending so much time in meetings or whatever.
Now, just like you can't buy anything without Wi-Fi in it. So I think looking at, you know, how things like Starlink will help bring connectivity to underserved areas, looking at, at least in the US, the ongoing investment into broadband through the NTIA, and their grant programs.
So I think, you know, trying to, I think, bring the Internet to the underserved areas, I think is continues to be exciting.
I spent some time at the Internet Society for Cloudflare.
And that is a big, big focus area of theirs is helping folks in underserved regions, bring connectivity to those areas.
Because if anything, what's fascinating is, you know, very much being a network of networks, the more participants you have on, you know, this grand experiment called the Internet, you know, the more stories you can share, the more businesses that are spun up, the more human needs are simplified or taken care of, you know, outside of the social media and shopping.
You know, it is fascinating to see, you know, more and more people connect via Starlink.
I think it was at Google when Project Loon was first announced. So that was fascinating to be like, oh, people can connect via balloons now.
The kind of leapfrogging straight to mobile for many communities and countries out there.
Right.
So it's fascinating to see, and especially from, you know, the front seat that is Cloudflare.
Yep, absolutely. And I think, and just the continued movement of people, I think, I mean, that makes it a little more challenging in many cases for us to do some of that aggregate analysis.
But looking at it, but, you know, I think that's another story to tell ultimately is how things are shifting.
And, you know, are we seeing, how are we seeing traffic trends moving from fixed networks to mobile networks or fixed networks to satellite networks?
You know, do users on satellite networks have different consumption patterns than users on mobile networks or fixed networks?
You know, are there different diurnal patterns and things like that?
Because they're potentially on the move. And kind of our last question to close things out, what advice would you give to anyone kind of looking to get into the data space today?
Wow. I think that there's, I think one would be deciding sort of where you want to play in it.
So I think that there's the side that I'm on, and I wouldn't say they're diametrically opposed sides, but I think there's the side that I'm on where I'm sort of sitting at a more meta level and looking at the aggregated data, the analyzed data and figuring out what are the stories that are hiding in it.
I think there's also the folks that I work very closely with and have a ton of respect for the data scientists who are the ones that are ultimately wrangling that data and, you know, helping make sense of, you know, the trillions of lines of, the trillions of log lines and, you know, helping tune the queries to get the data sets out that we can then look at and graph and tell stories with.
So I think, you know, it would, I'd say look at the different opportunities available in the data space and think about where you want to play.
Yeah.
I mean, because if anything, the opportunities will not be shrinking. We're only ever creating more data.
Just like the Internet proper, the current trajectory is there's only going to be more of it and having expertise to be able to, you know, work with it, process it is only going to become even more lucrative and rewarding.
Absolutely. And there's also sort of the data visualization side, which I didn't even mention, but, you know, that's another key piece of this is, you know, wrangling the data, telling the stories with it, but there's also the visualizations.
And that I think is core key to be able to tell an effective story is how can you take all of that information, all that data and make it understandable, make it visible, get people to really go look at it and go, okay, I get now that, you know, this is why that changes good or bad or meaningful.
Mm-hmm. And then just the time spent explaining, you know, correlation and causation to your audience.
Right. So, yeah, that's, yeah. So one is not always the other.
So with that, you know, we're nearing the end of time. David, I would love to thank you for taking the time with us to share your journey and your career story.
And for everyone watching us, thank you for taking the time as well, whether you're catching us live or one of the many reruns here on Cloudflare TV.
Thank you, Dan.
Thank you. Have a wonderful day. You too. Bye.