Dial Up Motive

Hello, hello, everyone. Welcome to Dial Up Motive, episode six. On this show, we explore some of the early Internet experiences with Cloudflare employees and how it shaped their career today. So if you're tuning in live, thank you for watching. If you're catching a recording, we appreciate your time. I'm here today with Usman Muzaffar, who happens to be the Senior VP of Engineering here at Cloudflare. So unlike previous episodes where I was talking with millennials that were growing up on gaming, here I have a much more seasoned Internet veteran. You're an old guy now. I have an old guy. I didn't want to say it. You can say it. I often joke that part of my credentials at Cloudflare include the gray beard. And as recently as the first company I worked for with our CTO, we called certain employees gray beards. Like, someone's been around for a while, that person's a gray beard. And so I'm happy to be the old guy on your show, Dan. All right, sounds good. And would you mind introducing yourself? Yes, absolutely. My name is Usman Muzaffar. I run engineering at Cloudflare. My job is to make sure that all the people who are building the product and who run the platform that the product is built on work together to ship all the amazing stuff that helps make the Internet safer, faster, and more reliable. Awesome. And what kind of technology does that entail? How are we working with the Cloudflare stack or supporting the Cloudflare stack? Yes. So the Cloudflare stack runs on Linux, which is an operating system, an open source operating system, meaning it's free and you can download it. And that's actually a big part of my story with computers because that became, because it was open, you could learn, you could, oh, you can, you can take the covers off and try to understand how things work. So the Cloudflare Edge runs on Linux. And then there's a lot of, a lot of other technologies. You might imagine there's a lot of databases, there's front end UIs, there's what are called reverse proxies, which receive things as lots of web servers of various sorts that receive packets on behalf of our customers and then proxy them, which means that we, we stand in the way of the traffics because that's the ideal position for us to block it or to make sure that it gets to its destination in the right way. So there's a lot of different kinds of technology that the Cloudflare stack is made up. And in some way or another, I wind up talking about them every day and working with them almost never. So part of the, part of the actual job of being an engineer is that I'm enough levels removed from the very talented people who are actually writing the code, developing it, and discussing that my job mostly is just to go around asking questions, hopefully questions that aren't too stupid, that help people say, wait a minute, are we sure we're doing this the right way? Or have we thought of this other thing? Or isn't this kind of related to what that other team is doing? Shouldn't we try to get the conversation and have those people in a room together? And that's, that's mostly what, if you ask me the question, what is the technology work with the technology that's on my mind is that the other, the other image that I think is worth mentioning and had the honor of, of keynoting one of their conferences is Wireshark. So Wireshark is a, is a, you can think of it as a microscope. It's the thing, if you point Wireshark at any, any two community, any two computers that are talking to each other, it will show you exactly what is going over the wire. What is the conversation that is going over the wire? And I don't pull up Wireshark a lot at Cloudflare, maybe did it once, just teach myself something, but that conceptual model of there's a lot of convert, a lot of detail in how two computers are talking to each other. That is, that's also on my mind a lot. And so all, all of those, all of those, those operating systems and those web servers and those protocol analyzers that I messed with in the nineties, they're still very much sort of front of mind here at Cloudflare, even though I never actually, I rarely crack an IDE. If I do crack an IDE, if I do actually do some development on my own, it's almost always some scripting in Python or something like that. Usually to cook up a little report, to help me, help me do my job or help, help me show the team that I want a report like this to help me understand how things are going. But yeah, I still, still find myself playing with computers all the time and really enjoy it. So hopefully that's a not too rigmarole of an answer to your question. Not at all. And what would you say is your favorite part kind of of a given day or your, your common work with people? The people by far, I feel so honored to be able to work with. The challenges, you know, they're always ebbing and flowing. We, we, even, even if we work really hard and we, we ship a great new product or we solve a problem, or we work with the customer-facing team, like folks like yourself to help make sure that we solve something. There's another, there's another problem right around the corner. It's really the people who I really love working with and learning from. And in some cases, getting to share my experiences and let people integrate those into their, their career as they're growing up. So I, I'm again, really honored to be here working with such talented people at so many levels. Awesome. So with that in mind, it sounds like you began life, you know, as a developer. Is that true? What, what kind of languages were you working with? Yeah. So certainly my, my career in software started in tech started as a developer. The first program that I got paid to write, I was messing around with, I was probably starting with computers as a kid and, you know, wrote in Pascal and Apple SoftBase, had an Apple II in the mid eighties, messed around with a little bit of assembly language. So I was a pretty well-versed technical kid. In those days, the biggest challenge was finding information. You couldn't get it. And it was a big deal. If your mom or dad could find you a book, like, I remember I had this book on 6502 assembly language. It's still on my shelf behind me because I treasured it so much. Like this thing knew so much about how the inside of the computer worked. But the first real job I had was in writing in the C programming language. I was hired by a university professor to write some code for a whole bunch of different experiments that he was running. And that was a lot of fun. And so I wound up, I wound up learning the C programming language that way. Actually has a direct line from that job to Cloudflare because the next language I learned was called TCL, the tool command language. And that programming language I learned because I wanted to create a user interface on Unix, which at the time was really hard to do. And then all of a sudden this TCL programming language shows up and you could knock something together in an afternoon. And it was sort of game changing. And, and as a result of that, I wanted to work, well, what's the company that's behind TCL? And the company that was behind TCL was called Scriptix and its head of engineer, VP of engineering was John Graham-Cumming, our CTO. And so that's, that is, that is the connection. So yes, which picking the language that you pick early on, can have an impact decades later. It's very interesting. And that's interesting. And maybe that, that explains more of my career is as I never could settle down. Like I started with PHP and web development that led me to Perl, which led me to Python. I used all those languages. Yeah. And that, so it all just kind of skirted the edge of, you know, hard, hardcore programming and just, I mixed and matched as, as I needed. What's interesting to hear is, the reliance on books and the reliance on, finding way, finding your way to information. Because from some of the previous interviews, what's been fascinating is for those that grew up, with the Internet, with the Packard Bells, a little bit more, you find an obscure topic, be it programming, a game or show, and you have the Internet then available to you to find that information. When, when was your first, I guess, pivot to using the worldwide web to either find information or develop yourself? Yeah. So before the web, I was on email and I didn't understand, to me, one of the most like important moments in my technical education was the day I understood what the Internet was. Because I had heard the word Internet and I had heard and used email and I didn't know what the relationship between email and Internet was. And nobody told me. I wasn't, at this time, I was studying to be a physician. So I was only around doctors. I wasn't around any engineers at all. And I remember being in a Borders bookstore. Remember Borders? Borders bookstore was like the library. That's because it had these increasingly large sections on computers. And there was a book and the title of the book was just Internet. And I was like, okay, let's crack this open while, and the first chapter was email. And the second chapter was Usenet. And the third chapter was worldwide web. And the fourth chapter was news groups. And I was like, okay, so email is one component of this thing. And, and to this day, I find if you ever crack open a book and the first third of the table of contents makes sense to you and the second two thirds don't, buy that book. That is, that means you have a book that is grounded in a foundation you understand, and it will launch you. If you buy a book and the whole thing is, is great to you, then, you know, that's, that's harder. But that was, I had this epiphany was like, wait a minute, I know the first part of this. I don't know the rest. I should figure out what the rest of this is about. So that would have been, you know, the first time I had an email address was probably 1991 in college and university. And, and I remember buying this book around 93. And that's when it all clicked. That's when I was like, oh, any computer can be on the Internet. Because up until that point, I thought of the Internet and the PCs and remember PCs were exploding like the, the PC that my parents bought me in middle school. And the one just five years later was so dramatically different. I feel bad that the younger generation never saw anything as explosive and extraordinary as that. The difference between the computer that I had in 95 versus the one in 85. It was just unbelievable to see what the, what the change in power was. That's not true. The difference between the computer you have in 2010 and the one you had in 2020, they're not that wildly different. But in the middle of the 85 to 95 thing, this Internet thing shows up, right? And it's like, what is that? And it's up to, in the university, there were these VT100 terminals on the, on the library. They didn't look anything like the PC I had, the Mac SE30 and the 2i, 2ci that my neighbor had with the colors and the graphics and the mouse. Like these things looked out of the previous century, out of the previous decade, the 70s. And here's this cool new whiz-bang stuff I have. And it took me a while to connect that, oh, the Internet is orthogonal to this. Any of these computers can get on the Internet. And once you connect a modern PC to the Internet, you can have a point and click interface to email. That was mind boggling. You can attach a document in an email. That was mind boggling. I remember the first time someone emailed me a document, I was like, how did you do this? There's no mouse on the terminal. How did you get a file onto an email attachment? And so that was a key part of it. Fascinating. Yeah. I think that the closest thing we might have to it is the development of cell phones and how quickly they've gone from a brick with numbers on it to a full screen, you know, smart computer, you know, so many tabs and tell you what the weather is like. Yeah. I can multitask. I don't have to memorize T9 words. That's right. Exactly. So, I mean, what's fascinating is, yeah, this discovery, I mean, were you still in kind of the mainframe mindset that, yeah, mainframes talk to each other. That's ARPANET's a thing. That's something I can picture. But then to expand your mind into, no, anyone can connect. Anyone with a device. Yes. So the other important part of the story was, at the time, home for me was still Islamabad, Pakistan. So my secondary schooling was in Pakistan and I came to the United States for college. And keeping in touch with family at home was a huge problem. It was really hard. As a college student, I would get those blue airmail letters. I don't know if anyone in the audience remembers them, but there were these very thin letters that, and I would write to my parents, try to write every week. And it was an asynchronous communication channel, right? So I would write every week, they would write every week, but only on the third week could you respond to a question that came. And so the idea that my closest channel of communication with my parents as an 18 year old was basically every two weeks with a handwritten email. I mean, that's the state, that's what the state of the art was. That was the most economical way to stay in touch. Phone call would cost $3 a minute. It was frightfully expensive. So it would only reserve that to maybe a couple of times a semester. And so along comes email and this idea that, wait a minute, anybody can have email. So I remember my parents or some other friends in Pakistan saying, email has come to Islamabad. And I was like, well, in whatever fashion or shape or form email has shown up, we have got to sign up. And it turns out it was from the United Nations. The United Nations has a wing called the United Nations Sustainable Development Program. And they were experimenting with basically the equivalent of a.org. They'd set up one and they had used really old school dial-up technology. So UUCP, if you remember, which is Unix to Unix copy and the email addresses, which had the bangs and the percents in them. And you would batch upload your email to this server in Islamabad that would hold onto them for about 24 hours. And then every night make one long distance call to London and upload everything. And the UN would foot the bill, which meant was I could have email turnaround in 24 hours, which was game-changing. That was absolutely game because it meant I could write my mom that morning and get a response the next day. And it was such a hit that they changed the frequency from 24 hours to 12 hours and 12 hours to four hours. And basically when I went back home to college, the summer is of 94, 95, I could stay in touch over email with my friends, which just seemed completely unbelievable. And what was going on on my PC was literally just, I was running temp terminal programs. They were saving the email locally to my, at that time, I didn't even think it was a hard drive or a very small hard drive, maybe 10 megabyte hard drive. Then using a 300 baud modem to very slowly upload it. And from there, but that was really important. Like it was an important part of my family that we were able to have email to talk to each other. And that was at the same time that I was learning about the Internet. So it very quickly became clear to me that this is not just a toy. This is going to be something that really, really helps people stay in touch and is important to how people talk. And that's one thing I've always found amazing is just the Internet's ability to connect people. So sometimes small town, Ohio, living on a farm felt very much like a different country and at least connecting to the world or connecting to others. And the Internet was kind of my personal gateway to, you know, information to people and to see that and to see how it developed. So even as something as simple as email via command line to your family across the globe, you know, it's fascinating and especially the way it's changed and evolved and will continue to evolve. You know, now you can have a video call, you know, with anyone across the globe and it's, no one blinks an eye. It's no longer amazing. Yeah, you just take it for granted. And, well, actually, I should say that not everyone takes it for granted. When I talk to my father, he always breaks out into laughter whenever he sees me. I'm never quite expecting you to show up on the screen. So that's great. Or just that, do you have the joke where the parents can never set the camera the right way? Of course, everyone should. We should be careful about making fun of our parents. At least mine are probably watching, so I'm going to get an earful later on. Like you did what? Well, with that in mind, so, you know, you were growing up in a time where you still had to find a book that ultimately had to teach you what you need to know. Information was about the Internet and about coding and development was still, you know, in its infancy. Even your first experience on the Internet was, you know, just email and email with stipulations. So batched work, you still had, this was not instant communication, but it was better than the alternative. So what do you see or what are you excited about kind of moving forward and what the Internet will become or what Cloudflare is building specifically? Oh, the future is rich. The future is very rich. It's because we are reaching inflection points and how powerful these things are getting and how ubiquitous the coverage is. It's still the case that if I go hiking and I'm in Silicon Valley, it doesn't take much for me to get out of range. You know, if I head into the hills and I think that will eventually go away and when that goes away in a completely reliable way, you know, there will be a change in the architecture of software. I think that right now, there's still a huge headache that everyone is constantly trying to manage multiple caches. And what I mean by that is, we're, as software engineers, we're always, we can't rely that the connection is there, is always there. We can't rely that the latency is low. And there's an analogy here, Dan, you know, like when, at the same time that I was, we were worrying about networks, we were also worrying about space and memory. And when I started my career, there was no such thing as a 32 -bit computer. Forget about 64-bit. We were just going from 8 to 16. But even the 8 to 16 transition was very fraught with a lot of manual stuff that programmers had to do. And so you had more than, you know, a couple, than 256k of RAM, but you had to be aware of how you were using it. You couldn't just take it for granted. You were aware that you were paging things in and out of the upper memory. It was a huge drain on engineering productivity and a huge limitation to how the software worked. If you zoom all the way out, I think we are still struggling with that at the level of the whole Internet. Everything we're doing, we're still worried, is it freaky? Is it latency? We're still have so much engineering goes into automatically reducing the bandwidth when the connection is flaky. So much engineering goes into wondering, is it in the L2 cache? Is it in the local cache? So much engineering goes into wondering, should I run this at the edge? Should I run this at the origin? And that, I dream of a world where companies like Plotler basically take that problem off the table. You are guaranteed that whatever language you write in, it is operating at essentially zero latency. It has access to everything and it has infinite memory. And once we get into that world, I think that there'll be a whole new, like, explosion of innovation. Because right now you still have to worry about all this stuff. And if you don't worry about it, you have a substandard user experience. And you can see the companies that worry about it are the ones that have been able to build great user experiences or they're not able to worry about things being seamless. So much of that as well is baked into our assumptions of what makes a good product. When really it should just be about the innovation and we should be able to take that for granted in the same way that we take power for granted. We just assume that there's electricity coming to all the devices. We don't judge software on how good it is with power, except for power when it's on battery. But we don't judge it on the service. And that's something I think I'd like to see changed. And I think software plays a big role in that. Have you run that by marketing? Take your network for granted? Take your network for granted. Yeah. I mean, that is a fascinating analogy, though, to position a network as the computer and us as a company worrying about where everything's stored. Is it at the top line memory? Will it be performant or can we move it down? And we'll worry about that as opposed to the end developer, the end user. Yeah, that's right. That's right. And there will still be, oftentimes, I think there's a fallacy that people make in the middle of their career. When they've done something about 10 years, they notice that they've learned enough about something. And they're watching people new to their area of expertise struggle with complexity. So, let's say you were a JavaScript jockey in the late 90s, and now you're overwhelmed and you've grown up with the JavaScript ecosystem. So, you've learned all these frameworks and all these tools and all these things. And you're watching someone who's starting their career. Maybe you're teaching your child how to use JavaScript and you're like, gosh, it's so complicated. It's too bad that these kids don't get to see the simple version that I saw, because that was so easy for us to internalize how things really work. And you're struggling to try to find the fun part, because you're teaching them all this complexity and like, where's the fun? The fun's buried in there somewhere, but I'm having to teach them all this stuff. And I've seen this pattern now multiple times where you're under the illusion that because there's so much complexity, because there's so many layers, there's nothing fun anymore. It's buried deep in there, but that's not true. The fun and the innovation and the creativity is just at a higher level in the stack. Or the green field is gone. It's a solved problem. You just don't see it. But the younger generation, the people who are coming at it fresh, they will see it because they don't have the same models in there. From my perspective, as a 13-year -old, my biggest goal in life was to build a video game. I thought that was it. If I could have something animated across the screen, I think the resolution of the Apple 2 was a 280 by 192. If I could have a little spaceship animated across that screen and shoot another little spaceship, I thought, wow, that's something else. And that's never going to be something they did. But on the other hand, creating a rich 3D world where you can spin around and explore something would be something more like what someone today starting out might fantasize about building. And so I think you should never underestimate the power of imagination and especially the power of the curiosity of young people to be like, wow, I wonder what would happen if you build X. Yeah. I mean, along those lines, I've had thoughts or I've heard the theory of when raising my daughters of introducing them to games and actually starting with the earliest platformers, like almost Atari level, because then they can understand the mechanics. And if we go through each console generation or game generation a year, they can appreciate, oh my gosh, these graphics are so much better. You get better and better. That's great. Yeah. It's almost like a history of computing through video games. So in a similar way and starting programming, and this is where I always enjoy connecting dots and understanding where things began, finding ways to teach the next generation, like, yeah, you can take this, this, and this for granted, but if you start here and kind of really understand where things came from, that will help you, you know, ultimately farther down the line, you'll, you'll have more models and more analogies available to you in your career, which will help you more than just kind of starting from, from the latest way you can start at anywhere in the stack and go up and down in either order. And you shouldn't feel you have to start with a, you know, with a first, I got to learn how to resistor works before I'm allowed to build my first web page. Like, I think that's, I think that's a mistake. And I think that overwhelms people. I think people think that I have to understand all this stuff, but you don't, you all, you can start at anywhere. Can we, can we quote you? Yes. You do not have to understand this. Yes. Correct. You do not, but you know, I think it's important to know I'm head of engineering cloud for, I don't have a CS degree. I took some CS classes, but I don't have a CS degree. I was, like I mentioned, I was studying to be studying, studying medicine in my college years, but I was, I was hooked, man. This stuff was just way too cool. And so I switched. So with that in mind, you know, coming without the, you know, traditional degree, have you felt like you've been able to bring in any of, you know, do you bring in anatomy metaphors, medical metaphors that have been helpful in your career at some stage that you're like, Hey, I can see this from a different angle. I do think occasionally that happens. I think it probably, we probably shouldn't stretch too hard for the analogies. I did make use of my medical background when I was talking about Wireshark, which I mentioned at the top of your show which is I think there is a, there's a very interesting analogy between Wireshark and microscopes in both sense. They're both tools that were built by a field to help them understand something. And then the development of the tool became a field in and of itself. And and in some ways, in the same way that learning about microscopes in informed drug design, like once, once people understood how molecules work, they're like, wait a minute, why don't we build drugs that way? In other words, design, the tool influenced design, rather than the other way around. You know, we had an example of that in Cloudflare. Once we, our engineering team spent so much time looking at at Wireshark traces, when it was time to build a really next generation firewall, what model monumental model did they reach for, like, you know what, it should work like Wireshark. And literally, it is the Wireshark language semantics that is embedded into the Cloudflare firewall, so that something as easy as typing a Wireshark filter is something that you could you could type at the run of the plug for edge to filter your traffic. And so yeah, I think there's some analogies with the life science studies. But mostly, the extraordinary difference between life science and computer science is that we built this, we invented this whole thing. And that's another great analogy that I think people forget, which is you don't discover a programming language, you invent it. And that, you know, all of this was someone's decision. And that's why it is very rewarding to keep asking why, because there's a reason why, whereas in life science, it's not always clear why or like the reasons are so opaque and go deep into evolutionary biology, or maybe not even there. And who knows, you know, why this enzyme works that way. But if you want to understand why the create Unix call is missing an E, you can look it up. There's an explanation for why you can you can possibly email the creator, you might be able to email the person behind it and be like, Why did you misspell it? And he'll tell you, I'll tell you why. Yeah, you can you can make the change on the next version of the programming language. Put on GitHub, you know, fork and edit and evolve. And yeah, the other thing that really hit me was, you know, in when I was studying life science, I had a, I met a grad student who was overjoyed one day, because the crystal that he was growing, so he spent a year growing a crystal, so he could shoot an x ray at it to take a picture. And so at the end of this whole year, like this is a whole year of his life, he's going to get a picture. And then he'll decide what that picture implies. And I just thought, I can't live on a feedback loop that operates at this, like this level, I need something which is more like, okay, I had an idea this morning, can I try it out this afternoon. And that I think is the most fun part about computers is the speed with which we can explore our ideas and iterate is, is so important. No, I can understand that I worked for the federal government for a while. That was a slower feedback loop. And I'm glad to be in the Bay Area. So so with that in mind, do you think that having a comp sci degree is necessary in this day and age? Or you can just stop right there? No, I don't think it's necessary. It's a huge, it's a huge advantage. I don't want to dismiss it. It's, it's incredibly important. Because again, as an example of someone who has the lower levels of the stack, as well, but no, I don't think it's a requirement. I not at all. And I think that if you look around, at people who are very successful industries, so many of them don't have computer science backgrounds, they came at engineering sideways, I do think you need a lot of curiosity, you need to be the kind and you need a certain kind of fearlessness, which is cannot be overwhelmed, just because something is here, you have to find that book where you recognize chapter one, but chapters two through 26 are foreign to you, and dive into it. But no, you I don't think you need to have studied it. And I certainly don't think there's some kind of time cut off. That if I didn't learn if I wasn't a hacker, when I was 11, I'll never be able to compete. I don't think that's accurate either. But but that that doesn't mean there isn't a lot to learn and, and a lot to experiment with. But yeah, absolutely. I don't think that's, I think that's really important, actually, for us to draw from people with other disciplines and other backgrounds. Alrighty, and with that in mind, I know we're coming to the top of the hour, I want to definitely give you thanks, Usman, for taking the time and walking us through your, your, you know, early Internet journey and early career journey. Definitely fascinating to learn about, you know, the early stipulations of email and global email. And, you know, to your point that over time, we'll just continue to take more and more of this for granted, people will grow up not having to worry about certain problems. And open up room for them to dream about bigger problems. That's right. That's right. Dan, thank you so much. This was a fantastic way to start the week. All right. Thank you, everyone joining. Have a great week and we'll catch you next next time on dial up motive. Thank you. Transcribed by https://otter.ai