Open Source Adoption, DevRel, and FOSS: Learning from Apache Cassandra w/ Patrick McFadin - EP. 26

Patrick McFadin, VP of Developer Relations at DataStax and Chief Evangelist for Apache Cassandra, joins the Hacking Open Source Business Podcast on Episode 26 to deep dive into open source.  In this episode Patrick talks about:

- His time working in open source database community, including Apache Cassandra's journey and upcoming developments.
- The role of evangelism and contributors in driving adoption and getting people to try your project.
- The challenges and mistakes companies make when commercializing open source, with lessons he has learned from his time in the database community.
- How new features are chosen based on his experience with Cassandra highlighting features such as transactions and open-source tool Guardrails?
- Does open source innovation slow down as products mature?
- What is cloud-native anyways?  And what does it mean in the database context?
- Building a diverse and gloabl team by building trust.
- DevRel Best practices includeing, how do you measuring DevRel success.

- Patrick McFadin's LinkedIn profile: https://www.linkedin.com/in/patrick-mcfadin-53a8046/
- Learn more about Apache Cassandra: https://cassandra.apache.org/

Checkout our other interviews, clips, and videos: https://l.hosbp.com/Youtube

Don’t forget to visit the open-source business community at: https://opensourcebusiness.community/

Visit our primary sponsor, Scarf, for tools to help analyze your #opensource growth and adoption: https://about.scarf.sh/

Subscribe to the podcast on your favorite app:
Spotify: https://l.hosbp.com/Spotify
Apple: https://l.hosbp.com/Apple
Google: https://l.hosbp.com/Google
Buzzsprout: https://l.hosbp.com/Buzzsprout

Transcript

Matt: 
[00:00:00] Introduction
Matt: Alright. Welcome everyone to another Hacking Open Source business podcast.
I'm one of your hosts, Matt Yonkovit, once again, joined by co-host Avi Press. And today we are joined by the one, the only Patrick McFadin from Data Stacks and the Apache and the CNF CNCF community. Like he's everywhere. He's, he's at all places. He's like omnipresent in the open source space. Patrick, what's going 
Patrick: on?
I haven't found a foundation I didn't want to join. Yes, that's right. Oh, oh, we're gonna, that's a great jump point. You got a new question one for me? Oh yes. Oh 
Matt: yes. We have, we have some new ones to, to, for you to join. That's great. So, So Patrick, so it has been, um, like nine months since we last we're on a podcast together.
And so I've, I've, uh, you know, joined a Voice Forces with Avi on this one. But if you have not listened to any of our episodes, have you listened to any of our episodes? 
No, 
Patrick: not yet. You is kind of new. Yeah, I guess Link and I'm like, oh, perfect. Yeah. There's what Matt's doing, so I need to go catch up. 
Matt: There you go.
Okay. So what we do is we start with, um, 173 random questions, and so we just throw them at you and see whatever you answer. And we just go with it. Um, and then, then, then we'll about other 
Patrick: real stuff. This is not a training set. No, we're not doing this. Oh no. This we're training 
Matt: our AI to be, you know Yeah, yeah.
Chat Patrick. Yeah. 
Patrick: See this is why we should stop all research right now. That's, we 
Avi: got six month pause with a six month pause. Yeah, 
Matt: that's just, let me catch up. That's all. That six month pause is for, is to let me catch up. Yeah. Yeah. So, um, alright, so Patrick, we're gonna start with rapid fire questions.
Right. And so, um, rapid fire questions, it's whatever the hell pops into my head. So, um, we're gonna just start with your first Linux distribution. 
Patrick: Uh, boy, what was it? Um, I downloaded 24 floppy disks from, uh, from somewhere in Sweden or Finland in 1991. Ooh. So, Ooh, 
Matt: okay. It was, so that would've been. So, so it could have been, it could have been Red Hat because Red Hat definitely was, was around then.
But it could have been slack wear. Um, 
Patrick: no slack. Slack wear was my second distro and Slack wear wasn't really nice. This was, I, this was, so I was working in a university and um, well it was in 92 that I was working at university, and we just downloaded whatever Linus was talking about because we were sick and tired of paying $10,000 for a Unix license.
Ooh, which 
Matt: Unix license was it? Uh, well, 
Patrick: we had SCO of course, uh, and at and t Linux, uh, or Unix. At the time it was running on some ancient X 86 hardware, some 3 86 wow servers that we had running. Um, And of course, all our sun gear, you know, we paid for sun off and, but that, that came reinstalled. Well, 
Matt: so, so Avi, this might, you know, like I, I don't, I don't know if you've ever experienced this, but it used to be that they would charge you for that os right?
So, A I X, uh, HPO X. Um, bns, all these, it, it would be like, you know, 50,000, $70,000 
Patrick: for the os. Yeah. It's 
Avi: so 
Patrick: wild. That's so, but it came with support. So expensive. 
Matt: Well, well, a actually the joke was you really don't care about the cost of the US because the hardware is like, you know, you know, 800,000, $900,000.
What's an extra 50 K? Right? Like, you know, like, huh. Yeah. You know, um, there were some pretty insane. Spins back then, let's put it that way. 
Patrick: Yeah. Being in a university, we were like, Hey, we'll just build it ourselves. Some guys started, let's keep, let's go roll with that. Let's go, 
Matt: let's go. There you go. Right.
You know? Right. Okay. That, that's always the, the best way to do it. So, um, as you started to, to get involved in the, that Linux space, um, and started to use it, um, what was your favorite early tool that you, that you just kind of fell in love with? 
Patrick: Oh, let's, well, in, in Unix land. In Lennox Land. Yeah. Yeah. Make, make, Hey, you know, fair.
Hey, they're building stuff. You gotta do it. Oh my God. You gotta do it. I was, yeah. Well vi and make were like my mainstays for so long because every time we got, I mean, we had random. Network cards and none of them, of course, there were no kernel support, so you had to build it all yourself. So yeah, I spent probably the first two years using of Linux usage, hand editing C files to make our network cards work.
Wow. Yeah. 
Matt: That's awesome. But, but I'm sure you don't do that anymore, right? Like that's not part of your current. 
Patrick: Day to day makeup? No, no, no, not at all. Um, no. Okay. Fair enough. So Curl is awesome. I'll, I'll give Curl a little more mainstream. Curl is awesome. Curl is the best? 
Matt: Yes. Is Curl your favorite tool, like right now, like open source tool?
It's one 
Patrick: of my, well, it's one of my favorite open source tools of all time because, I mean, it just is so useful for everything I need to do. Um, and I'm, I'm calling it out cuz you've started me on this like command line Linux thing, so I'm gonna just like, let's go og, but still relevant today. It doesn't have to be ai, it doesn't have to be Kubernetes.
It. 
Matt: Curl, you know, that's funny because I just saw that, uh, Bloomberg put together like a technology, you know, like Grant for like, uh, the, their engineer's favorite tools and Curl was one of them. 
Avi: So, you know, yeah. It was on the list. It's interesting to 
Patrick: me that. Yes. 10, 10 grand. Yeah. 10 grand. 
Avi: Yeah. Um, but it's interesting that like there's other, you know, there have been successor tools that have been, or, you know, intending to be the curl successor that have been built.
And some of them are pretty good, but they just don't catch on just like Curl does. H e t pie comes to mind. It's like quite a good tool actually. But I still use Curl at the 
Patrick: end the day. Yeah, I agree. I mean, it's built into the Mac now. You know, you can just get open a Mac. Fresh out of the factory. It's got the new Cupertino Cupertino smell to it and type curl and it works.
Matt: It's there. Yeah. Yeah. Well, it's there. Okay. So has anybody done the Cupertino air freshener for your car? I wonder if that's a thing. Can we, apple, can we get that? I think we 
Avi: shouldn't start a side business of that 
Patrick: Cupertino. It's 
Matt: like, no, you don't want the smell that I want. It's from old Nintendo games.
They always had that plasticy smell, which was Oh, somewhere between new car and like 
Patrick: electronics. Yeah, it was, it was that, uh, injection molded plastic thing. Yeah. Yeah. So good. 
Matt: So good. Oh, good. Good old days. Good old days. So, you know, speaking of the open source space, you mentioned Linus. Um, you know, is there someone in that open source space that you look up to that you, you know, think is doing a good job or that you think is just awesome every time you hear, speak?
Uh, except for me, you, you, you. I, I know you don't have to bring me up, but that's cool. Um, anybody 
Patrick: else? Anybody else? Uh, well, I'm gonna just shout out to coworker Sam Ramsey. Um, you know, he, he's been. And it's like the unsung hero so many times for open source. You know, he worked at Microsoft for years and helped bring them into the Microsoft, into the OSS world, which.
Really seemed like this is being tasked. Like, really, you, you're gonna try to get Microsoft into open source? Good luck, bro. So yeah, he would, he did all that and you know, he's still pretty active in the Linux Foundation and, um, so yeah, Sam Ramsey and he has a great open source podcast out there that, um, talks about open source data.
He's still doing it after 20 plus years, you know, uh, I, yeah, I just wanna make sure that he gets acknowledged for that. That's cool. 
Matt: Right? Super cool. Yeah. Consistency and longevity does matter. 
Patrick: It does. Yeah. He, he, he was like, he was open source before, well, not before. Open source was cool, but he was open source and enterprises before it was cool or even acceptable, you know?
There you 
Matt: go. Yeah. There you go. So, Patrick, what, what's your favorite conference? This the, you know, like right now, like, is there a conference that you just have to go to every year? Like, you're like, ah, I gotta go to that conference. 
Patrick: Um, well it's funny cuz conferences are such a hot mess right now. I mean, that's true.
You know, we, we tried to do a Cassandra summit, uh, in March and, and right on the heels of everyone, you know, every layoffs everywhere. And of course nobody could travel, so we had to move it to December. So this is a bad time to ask that. But the one that I do consistently go to and I feel like it's a family reunion, um, And I'm, I'm gonna pour one out for all the O'Reilly conferences, cuz those were, those were them oscon back in the day.
Oh, the latest. The one that's still around day-to-day, Texas 
Matt: Day-to-day, Texas. All right. 
Patrick: Yeah, all I have not signed a day-to-day Texas. Have I? I don't think I have. Which is unusual now. You should totally go. I never You should go. Both of you, huh? Well, together. When is it? I don't, yeah, it's, it's, we'll, we'll get end of January.
Yeah. We'll, every year. Okay. Yeah. Okay. Okay. All right. Good to know. That's great. Yeah. Fair enough. 
Matt: So, so Patrick, if you're not working on open source, what are you working on nowadays? What are you doing? What are you doing with your free time? 
Patrick: Well, right now I'm, um, I'm, well it is from a technology standpoint, I've gotten headfirst into home assistant, which is like the automation software.
No, lemme just not corn. Corn. If you haven't done this, Take stock. If you have way more free time than you think it's a great idea, go for it. But man, it, it is a rabbit hole because you're just like, you start out, it's like the gateway drug. It's like, oh, I just want to have it so that the lights turn on when I walk into a room.
Yep. 
Matt: And then it just keeps on going from there. It's like, can I play ambient music? Can I set the temperature? Can I, 
Patrick: oh, it's, it's crazy. Like, um, my, my kids, here's, here's a use case example. My kids are horrible at, um, like leaving things in the washing machine. And so I set up this whole automation routine so that if the, the washing machine stops, it will bug them until they move it into the dryer.
It's amazing, amazing. I mean, I've, I've, my girlfriend doesn't hear this podcast. Wow. That's, that is spectacular. 
Matt: I, I wish that, I would've thought of that years 
Patrick: ago. Good things you wanna do as a software developer is solve problems that make me lazier, you know? That's true. 
Matt: That is true. Mm-hmm. We'll have to get back to that laziness topic later on as we dig into some more technical stuff.
But, uh, I love that. But we we're, we, we've got a long list of questions to go. I don't know if you do this, Avi, but um, Patrick and I share a similar interest in science 
Patrick: fiction books. Oh, yeah. Not know that. Yeah. 
Matt: In fact, I didn't wear my Skippy habits downstairs, but I do have my 
Patrick: Skippy The Magnificent.
Yeah. 
Matt: See, like, so my, my trust the awesomeness. Yeah, if that's right, you know. So, uh, Patrick, favorite book now? What, what, what, what are you reading and, uh, what, what's, what's keeping your interest? Well, you 
know, 
Patrick: I've gotten into the biography, rock biography thing lately. Um, especially on Audible because, uh, I just finished the Bono book and I'm reading or listening to the Dave Grohl book, you know, from fu Oh, that's cool.
So 
Matt: good. The Dave Gro book is so good. I 
Patrick: mean, I just like, I mean, I, you know, I, I had my coming of age in the eighties as well, just like him. So we're, we're almost the same age. So it's kind of fun to listen to his spin on it. But, you know, I, I went tack and he went music. I think he's doing better. But anyway, um, the no regrets.
It's no regrets at all. Um, but I just, yeah, I really, I, I love the fact that, you know, you're like, why would I want to hear Bono talk? Man, that, that, I'm gonna listen to that one again and again because it's just he, you find out that these are just people and he's way more down to earth than you expect.
And he even called it out. He's like, yeah, I can be a rock PR Madonna, but that's part of my shtick. But, you know, and, and Dave Grohl. Dave Grohl is probably the dude you could just hang out with and not even know he was a rockstar. Yeah. Seems awesome. Yeah. Yeah. 
Matt: He's just, he's just so down to earth.
Everything that I've seen read, you know, like, you know, it's just, yeah. He's that down to earth guy. Right? Yeah. Um, like if 
Patrick: you, if you didn't have any idea, you lived under a rock and you met him, you would never know. Yeah. 
Matt: Yeah. So many great stories. I don't wanna ruin it. You know, we, I, you know, but there's some, there's some really good stuff in that book, you know?
Uh, good choice. Good choice. Highly recommend it. Um, so speaking of rock stars, is there a celebrity that you, you know, who, who, who do people mistake you for celebrity wise? 
Patrick: Oh, yeah. Uh, who? Well, Tom Arnold, I've been mistaken for Tom Arnold. Okay. So many times. Really? That was kinda first class seat in Newark, New Jersey because the, the gate agent.
Was convinced I was Tom Arnold, and she was like, it's o I said, I'm not Tom Arnold. She's like, it's okay. I understand you went your privacy. And she walked me onto the plane and said, can we find a seat for Mr. Arnold? You should have said yes. It's amazing. It's amazing. Well, and they were like, oh, you know, we're all booked out in first class, but you know what, we can move you up into, uh, into like the, whatever it is, the, the expanded first or the, the great economy.
And they, they let me just sit down before they boarded the plane. I was like, wow, okay. It was just a weird 
Matt: situation. Okay, so welcome to the Hacking Open Source Business podcast. Our guest, Tom Arnold, um, John today.
Great. Yeah, so I, that's interesting. So, Uh, so, okay, so if, if you or Tom Arnold is at a conference and you know, you, your, your, your love of music, you, you, you have to pick the song to walk up to the stage on what song do you pick? 
Patrick: Oh, wow. Well, that's a tough one. Um, I, uh, one time I, I picked and it was just the total troll, but it was so much fun.
It was the theme from, uh, from, what was that movie? Um, oh, it wasn't, Shaft The shaft. Didn't you know that song? That is such a good answer. And it was only cause David d Coney doing the shaft theme in the X-files was like a transcendental experience because he would, he would like, he was walking down like this dark hall and it was super scary and he started singing the, the theme to shaft.
I'm like that, I want that. Yeah. Wow. Okay. I love that. Speechless fair. Yeah. 
Matt: I've, that doesn't happen very often. I've done, I have not done that. You know? Um, I, I've, I've been to, I think the wildest thing that I've done is, um, you know, doing, I was at a wedding that did karaoke as part of like their, you know, their thing.
And, you know, I, I decided to sing. It's the end of the world as you know it. Um, which I thought was apropo, you know, that's 
Patrick: a lot of words. Yeah, yeah, yeah. You gotta know all this, but that's okay. 
Matt: It's okay. Yeah. Um, yeah. Alright. So what, what, uh, getting a little bit back towards the technical side, what, what, what are you using to edit code now?
What, what's your code? Editor of choice, 
Patrick: uh, vs. Code like the rest of the planet. I guess intelligence, the other answer. Right. But, um, yeah, everybody's doing vs 
Matt: code. Everybody's, I'm doing B 
Patrick: is good code. Yeah. 
Matt: Avi's, not Avi's. An Eemax person. 
Patrick: Space. 
Avi: Yeah. Uh, space Max. I, I, I like them too. So don't, uh, before we get Oh, all up in arms at me.
Yes. It, it's, but I'm thinking of switching too. I feel like I'm gonna get left behind on all the AI tooling if I don't switch soon. Um, um, emax and Vem are not keeping up with that. 
Patrick: That's true so far. True. But emax, it's a, it's an operating system. No, it's an editor. No, it's both. It's all, it is. Both I, 
Avi: I'm so attached to org mode.
I don't know how I'm gonna live without org mode if I switch to, to VS. Code is my one problem. 
Patrick: Um, I feel like if you look into VS code marketplace, you'll probably find how to eax my VS code and all those things will be available. Yeah. 
Avi: I, I check, I check every couple months to look at what the org mud plugin is looking like.
It's still kinda abandoned. It's so sad. But I will switch one day 
Patrick: maybe, or maybe, you know, used Space Max forever. Yeah. I mean, yeah. My, my brother. Yes. You could be the last A you could, yeah, 
Avi: maybe I'll just take the project over. 
Matt: They'll write a movie about you called The 
Patrick: Lamp maybe of us. My brother who writes Fortran still uses eax by the way.
So just, oh, calm that out. Very good. Very. You go. Very good. 
Matt: So last question in the rapid fire mode here. Um, so the craziest thing that you've done to promote Cassandra. 
Patrick: Oh, this is a, is this a softball question? You know the answer to this, don't you? Um, I don't, so many of the listeners may not, may, maybe not the craziest thing, so, As many people may or may not know, we'll just say that.
We'll just assume nobody knows that one of the primary features of Cassandra is its multi data center capabilities where and shared nothing. So you should be able to destroy anything inside a Cassandra cluster and it will still be working. I took that literally. And uh, at one particular oscon, I think it was in 20 15, 20 14, um, in our booth we had a destroy a raspberry pie set up.
So I had a raspberry pie cluster of, uh, Cassandra cluster using ra raspberry pies as a note. And then I had a selection of tools, uh, hammers, drills, you name it. Um, I love cutters. Whatever you want. And I invited people from the audience to come up and do whatever they wanted to a raspberry pie, pick one and destroy it.
And it was, it was really out of control. As a matter of fact, we got in a bit of trouble with the staff at OSCON because they're just like, you guys are creating a scene. And I'm like, mission accomplished. But the thing that was really cool is we would have like an app running like a little laptop with the app running and then, you know, here is running out of this cluster you see in front of you and somebody would take a hand, you just get so violent.
We had like a little plex, uh, plex sand shield go around here. It was total myth busters thing. I had a lab coat and they would just destroy this raspberry pie. A couple times you had little near fire because you know electricity, but, And then after we'd still be running through this whole thing and it was, I bet you we probably got a lot of Cassandra users after that.
Yeah. What a great demo. Well, what a great demo 
Matt: demo. Yeah. Yeah. We need to figure out something like that, Avi. Right? Like break 
Patrick: something. Yeah. We 
Avi: need to, we need to follow in that. Yeah. 
Matt: Patrick survived rapid 
Patrick: fire. You gotta do the inner PT Barum. That's the thing. 
Matt: It's part of the dev role. Right. You know, it's rah rah, rah.
Patrick: So, 
Matt: um, alright, Patrick, you've survived the RapidFire questions, you know, congratulations. Although, although there, there is something that I don't know if Avi knew, um, you know, we, and, and, and Patrick, we talk a lot about businesses and businesses obviously, you know, sometimes they get bootstrapped, sometimes, you know, they're funded by VCs.
Um, I, I don't know if you knew this Avi, but Patrick's first computer was bootstrapped. 
Patrick: What does that mean? 
Matt: So Patrick, why don't you tell us how you paid for your first computer? 
Patrick: I'm, this isn't even a funny story anymore. It's just now you're just making it weird, Matt. Yeah. 
Matt: I am not making it weird. 
Patrick: Avi doesn't even know.
Matt always makes it weird. I think we can count on that, huh, Avi?
So, uh, yeah, no, I, I didn't have the money. This is like 1979. Uh, so computers were, as my dad called it, Hey, Fad. Um, so I, wow. I was kind of on my own. He was. And I lived on a farm. I lived out in the middle of nowhere way northern California. Um, if you drive north of San Francisco and you get into the scariest part of California, that's where I lived.
Um, and the. The only choice I had was to figure it out. And the way I did it was I got into a pig breeding operation and started selling pigs and, um, made a lot of money off of that. Yeah. And I bought it Atari 400 from that experience, 
Avi: how many pigs, uh, was one Atari 
Patrick: worth? Oh, well that's the theme.
That's a valid question. It's a valid question. Um, if you, when you do pig breeding, I mean this is 19 80, 79, so not much. But at Atari I actually still have my original Atari and I still have my original receipt from that. And the whole thing costs $500, including a cassette to wow reader 16 K 400 and, um, with a basic cartridge.
So, I mean, that was quite a few. A lot. Yeah, but, well, there you go. Yeah. That's amazing. 
Avi: That is 
Matt: amazing. See? See, that's a great bootstrap story for those looking to bootstrap your business. Why don't you just have, you know, a side business of selling pigs? I think 
Avi: it's super fun. There's really, I love the stories like this, you know, there's the Airbnb story of like the founder selling cereal to, to kind of get them over the hump of the cash, uh, that they needed to just like, make it another month.
They were just selling, like it wasn't something like that. No, it was great. I, I love those stories. Yeah. 
Patrick: You gotta do what you gotta do. Fair enough. Fair enough. Alright, why 
[00:22:06] About the Apache Cassandra project
Matt: don't we move on to something a little more technical, a little bit more, yeah. Um, related to Patrick's shirt today, which is of course the, uh, patchy Cassandra project.
Um, so what has been going on in the project lately, Patrick? Tell gi give us, give us a little rundown on what's new and exciting. 
Patrick: Uh, let's see, what's new and exciting. Well, I mean, we we're getting ready to. Release Cassandra five and which is a pretty big milestone, you know, as, as the as story goes with open source.
There's definitely eras and we've gone through the eras of, um, Just the early days where it's just a few people contributing and then a whole bunch of people contributed. And then there was this breakdown of trust and love and it was the adolescent years. And I, and I think as a project, we've moved beyond that.
With Cassandra four, it was clear that, all right, we're ma now mainstream database, you know, there's, we have some rigor, et cetera. Cassandra five is really cool cuz it, uh, Apache Con. Last year it was most of the committers. We were all there together and there was, uh, some other folks, it was mostly like a project co collection of project people.
It wasn't like just base users, so people who contributed. But what was cool, it was like Apple, Bloomberg, Netflix, of course data stacks. Um, a few other smaller, I like Microsoft. Small companies like that. Um, that. You know, the, we were all in this same room. There's about 40 of us, and it was like, we were just all working at the same company.
There was no, there was no lines like us and them, but our, the, the discussion that we had was like, wow, we've got something really special here. Um, and it's because we never shipped a product that was based on a marketing deadline. We were shipping a product that solved our little individual problems that we agreed to, and we built something that actually solves a problem instead of looks awesome on a glossy.
And you know, it, there's other things that have suffered along the way. Like there's absolutely no user interface. It's like you're on a command line, so you know there's that. But um, but what do you do with that? And. It's, it's, that's where we're at and that's, I think we're getting into the more exciting phase of Cassandra because we have asset transactions and completely revamping the way we do distribution.
Uh, it's just gonna be really cool. 
Matt: Now, it's interesting you mentioned some, you mentioned a lot of things, but one I want to jump to first is you talked a little bit about those kind of like growing pains, the adolescent years. You know, you go through the different phases. And a lot of open source projects do that where they start off and there's a lot of buzz, there's a lot of interest, there's a lot of contributors.
And then, you know, certain things change. They evolve over time. And then that kind of dwindles as those contributors who are really in it for the cool, you know, tech or you know, they have other things. They move on to other projects and then you are left with a core group of people and then you'll attract more people as the project.
Get stable, who have an interest in the productionization or the long-term longevity of that particular product. That's a really difficult span for people, especially when people are looking at the number of contributors or the community as a metric on how fast their project is growing or how they might be able to, you know, um, you know, gain interest in the market, you know, place.
Um, so as you're going through that, It was, was that like eye-opening to be like, you know, where did all of our contributors go the first time You saw that? Or was it something that like, you, you, hey, just everybody calm down. This is normal. Were were, were you going through that phase where, um, there were some people who were more concerned than others?
Patrick: Yeah. It wasn't, you know, this is the lifeblood of any open source project contributors. Like, you don't want to just sit there and see, look in the GI, and see that their last commit was a year ago. And you're right there is a core, but it goes to. I think it goes to a couple of things, especially around the fact that there are companies that are willing to participate, like the company itself.
Like, oh, you know, apple and Netflix are great examples, huge companies, tons of engineering talent, but they're willing to let their engineering talent spend time on an open source project. Huge, right? If, if, You know, if Apple or Netflix have one day said, oh, no, no, you can't work on that project. Well these are people who are employed at that uh, place and they couldn't work on the project anymore.
So straight up acknowledgement like that is huge. Bloomberg is really good about contributing to open source and letting engineers spend time there. Um, we were just talking about Curl, you know, they just contributed like $10,000, so yay. Yay. That. Um, but. The contributor thing is something that we always have to keep in mind though, and it, and I think that becomes harder when you have a, a longer term mainstream project because over, because there's such a runway to get ac or get productive with a huge code base.
So I this, this is the downside of being around for a while. It's like when a new person comes through the project, they're like, Ooh. It's like claiming Mount Everest. Yeah. There's a 
Matt: lot of stuff there. Yeah. Yeah. 
[00:27:34] Contributors and their role in the adoption at their companies
Avi: Did you find that early on, the contributors who were. Who were coming to Cassandra were driving the adoption at their companies?
Or did that require some amount of, um, you know, marketing or selling on your behalf to, you know, actually get people to take a bet on a new database? Um, I, I imagine that even if companies are struggling with a particular problem that like your database solves, you have this question of like, okay, well who else is using it?
Can we really trust this open source database that we don't know a whole lot about? How did you build that early momentum? Um, You know, with, with 
Patrick: corporate users? Yeah. There was a, there's a couple of things there. First is, is the early adopters, you know, are they able to first talk about it? And, and that actually fueled a lot of the initial growth with Cassandra cuz uh, the classic fomo, the fear of missing out.
Um, this was like 2011, 2012, um, that there was a lot of choices out there. There was a lot of NoSQL databases. Um, and there. In just in like the Cassandra adjacencies, there was Voldemort, there was React, you know, there's Tokyo Cabinet. There's a lot of, there are a lot of choices in there that you could make.
And you're right. Like do am I gonna make the wrong choice? But when you see the ones emerge, it's the, oh look, Netflix is talking about using Cassandra all the time. So the fear of missing out, I was like, I wanna do that. But to your point about marketing, um, yeah, and I, you know, I was working at Data Stacks, uh, doing consulter.
I was con a consultant working on Cassandra implementations. That's all I did. And. My, my boss, uh, Billy, who was our CEO at DataStax, um, you know, he's like, I think, I think that we need to put a different kind of energy out there and I want you to lead this effort. And that was when I created an evangelist team for Cassandra and that was it.
Um, and we had a lot of fun and it was just going out and doing meetups, doing. Conferences and talking about it in helping people understand it and evangelism eventually now is, is what we call developer advocates. But that was, that was his thing. And so data stacks put money towards that function for good reason, right.
But I think that we saw the most growth there cuz we funded the Cassandra Summit and stuff like that. 
[00:30:03] How FOMO shapes adoption in certain ways
Matt: You know what's funny is, uh, Patrick is, you know, that fear of missing out, you know, to another point, to Avi's, you know, question there is a lot of times that fear of missing out makes people adopt it when they really shouldn't.
For certain use cases that you're like, why are you using this product for this? Like, what the heck are you doing? Um, you know, I'm sure, especially as a consultant, you might have gotten dropped into a few of those places where you're like, huh, never would've thought anybody would've done this with 
Patrick: that.
Yeah. Well, and that was, that was what was really cool about, uh, just a, a taking a, a turn at open source is you don't have to do the marketing slide. Like, this will solve all your wildest dreams and, uh, use this and, you know, all your problems will go away. Some of my most popular talks that I would do were the ones where, Where were, don't do this, like what not to do and um, five ways you'll fail.
You know, it was like, these are really little negative and little down. But man, they were always packed because people wanted to hear the reality. Like, if you're running Hadoop, don't think that you're gonna use Cassandra cuz they're different, you know, or, you know, it's like that kind of stuff. And I think there was an appreciation that, you know, thanks for the real talk, you know, like, because.
There were people trying and they were failing. So, um, I, I feel like that's, that was really important. That's, that's where truthiness in an open source project is really helpful. Hmm. Yeah. 
Avi: I, uh, Matt does a, a talk on destroying open source communities. That is very good. And I think that 
Patrick: it's, yeah, it's 
Avi: a great theme for talks, uh, just to how to do the negative thing that's got the hat and everything.
Patrick: Got the hat for the talk. Of course you do. Um, 
[00:31:51] The don'ts of commercialization of Open Source databases
Avi: so I'm curious, like we, you know, we see a lot of startups nowadays that are com that are trying to build new databases in commercializing, uh, open source, uh, yeah, just. There's so many problems that can be tackled in the database space, and I think that it's a really great example of, you know, what other companies could and should not do in trying to commercial, uh, commercialize an open source database.
I'm curious, do you have any things that you see commonly a lot of these companies doing wrong or, or doing specifically? Right. That, um, You know that, that you've seen from, from working on Cassandra and at Data 
Patrick: Stacks? Oh God. There's just a list of sins. It's probably in the, in Matt's talk about how to destroy an open source community.
I mean, I, it, I think this is any company, data Stacks is not immune to any of this. No, no company is, I've never seen a company nail this because it, it's, it's just not human nature, but it's like, Wow, everyone's using it quick. Let's commercialize the hell out of this thing. And, and then pure panic, oh my God, no one's gonna pay us money.
It's just gonna be free software out there, like a floppy disc in the mail, like the AOL disc back in the day. Like that's, we're, we're gonna fail. And then, so the result of that panic is usually like some really bad choices for. Open source licenses, open source ish, like bsl, like, you know, or I mean, yeah.
Yeah. I mean there's, I remember side sidebar on, on bsl, um, Monty. Uh, viness from, uh, Maria b. He, he and I used to talk all the time at oscon and he was dreaming up bsl and he was trying to, he's like, Hey, you know, the Cassandra project, you ought to think about this. And he and I were debated it quite a bit and I just like, you know what?
It's so complicated if I need a lawyer. To understand my open source license then no. And so we just agreed to disagree and, you know, but that's Monty's thing. He's, he's very good at getting, he's going out and debating and that's, I always enjoyed debating with Monty, but he, he did it. And so, you know, that exists.
But then there's other, you know, the bigger controversies, like, you know, what happened with Elastic? Like, what happened? Well, Wilson Hammond's, Amazon was like open search and. It didn't work out the way you kept it. Um, we've gone through this with the Apache Cassandra project and Apache license is if it doesn't hurt or it doesn't make you second guessed, and you're probably not doing apa not doing open source right.
[00:34:35] Follow-up Question
Matt: So Patrick, I do have a follow up question On your Monty discussions, how much vodka did you drink and how often did you end up under the table? I, 
Patrick: I didn't, I made it a point to never drink with Mar. I already, I have my tea. 
Matt: Oh. How did you get away with that? He made everybody drink that black vodka. Oh.
Patrick: Because we always did it in Oscon. So I know I was actually 
Matt: at a conference talk and I was at a panel with him and he brought the vodka out on the panel and made all the panelists drink. 
Patrick: Yeah. Uh, boy, you know, my liver said no, but I, yeah, I know that the Monty story with his black is, is world famous.
Yeah. Yeah. Um, fair enough. Yeah, yeah, yeah. No. 
[00:35:18] Cassandra's new features
Matt: Yeah. Yeah. Um, so, okay, let, let me ask you this. I know like you mentioned another thing about Cassandra that's coming out as transactions, I know like, um, a big feature recently was, um, I think it's called Guardrails, right? Which is Yeah. A, a way to help, um, automate and keep your configuration.
Uh, properly set up in tune to avoid common mistakes. Is that correct if I got that 
Patrick: right? Yeah. This was a classic, Hey, we're solving a problem that sucks for us and we're gonna open source it. Yeah. Guardrails was that. It was, we actually built it for when we were building Astra, our Cassandras of service.
Cuz we're finding out quickly that when you give anybody access to a cloud version of Cassandra and. Just let 'em go for it. There's a whole lot of YOLO going on that you can't support. So, um, we, we implemented this, we started implementing this idea of guardrails, which is, it just puts these reasonable boundaries around things where, and that it's been kind of a thing for Cassandra.
It's like you can do anything you want and just run it into the ground. Um, it will. Hang on to the very last second and do whatever you ask and just destroy itself. So it's, and it's, uh, it, it actually is, it turned out to be like one of those things since we've opened, sourced, you know, put all the code in.
Like Netflix just did a big talk about it. Like basically like how this thing Oh really our lives. Yeah. Like we are, all the people who are on pager rotation are like loving this feature. It's their favorite thing. And this is Netflix. 
Matt: Oh, that's cool. They have like top engineers, so I hear, 
Patrick: but they still suffer the same problems, you know, so it's really cool to see that kind of thing happen.
Yeah. Well, so 
[00:37:02] Building new things and the argument of loss of innovation
Matt: that, that's an interesting question because, so one of the things that I have heard, and I've talked to you and I debated with some different folks, um, in the last six months about, is a lot of the technology that we're building, especially in the open source space, is either derivative works or it's designed to solve, augment, automate existing projects, make them easier, which is all great and everything, but there's an argument that.
You know, we're having to do this because either, you know, there's not enough talent or we're just all getting dumber from, you know, whatever, you know, mechanisms happening or you know, the fact that there's just so much to do and we can't keep up with it. But are we losing some of that innovation?
Because a lot of the innovation, a lot of the new things, even in the database space, it tends to be. The quote unquote build on the shoulders of the giants who came before us as opposed to net new things. And there, there's a lot of people who say, you're an idiot if you ever build your own database nowadays, because there's so many great databases out there, you might as well just start with, fill in the blank and then augment it until it's something completely different.
Patrick: Yeah. It seems, I, I think the, the. Timeframes I've heard is like, it seems like every 10 years we just reinvent everything and it's like, oh yeah, we need a new thing except for curl, apparently curl, it's like everlasting, right? So, good job. Um, but, you know, uh, in my Kubernetes work on the Kubernetes side of my brain, um, I've been involved in quite a few.
Well talks and then other folks that are inside the Kubernetes community where I think the, the, the really strong message is I think we can stop inventing things now. Um, and the reason I, I say that is of course we need to reinvent. We always need to invent things, but stop reinventing things and start thinking about how we assemble them.
Like we're moving into a different era. Do we need. 17 different choices for a database, or do we need a better way to deploy it in a cloud native way? And now it's like instead of reinventing. The, the basic, let's rein, let's invent better architectures and deployment methodologies. 
Matt: Yeah, so it's inventing new things, just not reinventing the things that are already there.
So just reuse what's, what's there, but look for new ways, new opportunities to maybe do something completely different. 
Patrick: It's, it's an invitation to try something different instead of just reinventing things. But, um, and part of that is innovation cycle Inside it, it will spur more innovation, um, in the MySQL space.
Uh, look at what, um, you know, there, there's projects like Planet Scale is doing a lot of innovation, uh, with. Uh, my SQL making it cloud native, um, Cassandra project, we're doing a lot with Cassandra, making it cloud native, um, T I D B, you know, they're doing a lot of innovation work around making databases cloud native, and then they're actually, I, I love Ed and you should talk to Ed if you have it.
Um, ed w um, at T I D B. But, oh yeah, yeah. Uh, he's embodied this idea of assembly and created like an assembly of good things in that are open sourced into a single open source project. And it's really this cool derivative thing, uh, where. I feel like he and I, when he and I talk, he's always like, oh, I just wanna bring in the best, you know, the greatest hits create like a mixtape.
Uh, I, that's, there's still plenty of innovation and a lot of code that needs to be written, but it doesn't mean you have to reinvent a consensus algorithm. 
[00:40:43] "Cloud-native" meaning
Avi: Okay. Here's, here's maybe a dumb question, but when we, when we say, uh, we want to, you know, make, make this tool, this database, this product, more cloud native, What do we mean by that?
I assume, you know, a lot of people running Cassandra, other things are already not, you know, not on-prem. Maybe they are, but a lot of times they're not. What does it mean that we're gonna make 
Patrick: it more cloud native? That's a great question. Almost another softball. 
Matt: Um, this was not 
Patrick: planned. Not planned? No.
Yeah. Funny. I wrote a whole book about that. Mm. Um, Availability, Amazon and book sellers everywhere. Um, there you go. Now the, the i the idea of, of cloud native database especially is, you know, moving away from, Hey, um, my, my product, my thing can run on Kubernetes. Um, but what if it ran in Kubernetes? Which means that now we can rethink things like, oh, there's probably a lot of duplication of effort going on here.
Like, A control plane in Kubernetes can do a lot of things. Do I need to have two control planes or can I reduce the amount of code that I'm running? And it comes down to really, one thing is separating compute network and storage into its components and letting those scale elastically any way you need.
And traditional database installation was monolithic. When you install a database, it's the database and you manage it by yourself. You have DVAs and you do all this stuff well. Moving. If you don't need a dba, then what's the choice in a cloud native world sre you. You just need to have SREs that manage the whole thing top down.
And they don't need to be specialists with databases. They need to be specialists in deploying cloud native. And so how do we rebuild parts of the database to make their lives easier here?
That's a great answer. 
Matt: Avi was buying your book there. Yeah, 
Patrick: I know. Yeah, yeah, yeah. 
[00:42:48] Patrick's role at the CNCF
Avi: An order. An order will come in, in a sec. I just feel like that that's one of those phrases that people throw around a lot and they mean a lot of different things when they say that, but that I think, um, yeah, the duplication of work, um, pushing more of that stuff into Kubernetes, I think that makes a lot of sense.
Um, and maybe is a good. Segue into talking about, um, your role within the cncf, um, within the Apache Foundation. Like what, what are you working on in the, what, what, what is your primary focus within of those foundations? And I'm curious, like, what are, what are the biggest kind of challenges that you're working 
Patrick: on there?
Uh, well, the sf I mean, as, as. Everyone probably knows ASF is a, is a fdo, right? It's the, the real power and, and importance is in the project. You know, so each project is an independent thing that, um, but they're, the overall ASF is just a, a governing body that makes sure that, uh, the bylaws and everything are handled, but the, the real.
Product or project governance is inside the project management committee. Um, I'm a committer on Cassandra, so I get to contribute, um, things in that way. But right now my particular effort, and this is kind of funny, it's I'm running the marketing and promotion group inside of the Cassandra project, so like working together with other.
Companies that support Cassandra to like work together. And so we do a monthly, uh, well, we do two, two times a month, but it's monthly for different time zones, uh, or just a marketing working group that. Um, we lay down our arms and all get together and talk about like what's best for the project. So that's that.
Do a lot of that. Um, and we, we have a planet cassandra.org, which is our, kind of our centralized hub for, uh, nonpartisan information for Cassandra and, um, you know, so that, just trying to really hard to make sure that works, um, on the CNCF side. Uh, Well, not even a c and CF mainstream project, but thing that I, I do a lot of work with is, uh, um, data on Kubernetes, um, community.
And, uh, I just help finish a white paper that's gonna be presented, uh, next week at Coon EU on running data on Kubernetes. So, yeah, I feel like that topic is a really important one for. Future people like my kids are winning computer anything. Are they doing any engineering work? I'd really hate for them to install.
Software. Yeah. So I'm working towards Interesting, 
Matt: interesting that, that that's good quality character building activities. Patrick, I mean, don't you want your kid, it's, it's like, it's like having him go shovel, you know, like, you know, snow outside or you know, like, you know, uh, clean up the leaves in the yard.
It's build's 
Patrick: character. Yeah, sure. Oh yeah. But, um, Yeah, they get plenty of character trying to figure out how to write a JavaScript out. Yeah, yeah, that's true. Well, they're all just using cha G p T anyway, so who cares? Uh, well, yeah. That's a whole nother thing, right? You know? Did you write JavaScript by hand?
Avi: Wow. You must have so much character. 
Patrick: I told my son, who's 15, I was like, Here's a new, here's a new job for you. Prompt engineer. 
[00:46:16] The challenges of working in a commercial entity
Matt: Yeah. So let, let, let me ask you, you know, specifically on like, you know, pick your poise, your experience, you know, you, you've worked across a lot of different foundations projects.
Working from a commercial entity and also working as a project that deals with commercial entities. Um, what are some of the challenges, you know, that kind of overlap, those two? Because obviously there's sometimes competing priorities between the two. Um, how, how do you, as a member, um, but also as, as someone who, you know, works at a company, h how do you navigate that and make sure that everything works, you know, well together?
Patrick: First thing you have to be okay with is somebody's gonna be pissed off at you. There you would, and at no point in your life will you be like in this utopia of happiness because someone's gonna be pissed off because you biased something in one direction, either commercially or open source. Um, you gotta be, it's like being, working in a call center.
Just be ready for abuse. Um, the, the second thing, which I think is a little more positive, is this is open source software. Is based, is really the foundation, is the, the human connection. And, you know, building up those human connections and having a really good network of friends. And it's, it's always this thing of, you know, we, we can build a trust relationship between a group of humans first, and that makes those other conversations easier.
So I don't get side to eye when I say, Hey, we should go do something like this. Like I there, there's a trust relationship. Now maybe they're like, well, that seems a little commercial, or maybe you're biasing too much on the open source side. Great, we can have that conversation without throwing knives at each other.
But boy, that that has been a cornerstone of all my work. So I 
[00:48:18] Building trust in asynchronous teams
Avi: think Well, that's a really great point. Um, I'm curious in, you know, in an open source context where you just probably have a lot of people that are in a lot of different places and a lot of asynchronous communication is happening. I think that, um, getting people together to, you know, go to a talk or have a beer or do all these things in person can be a really, really effective way to Yeah.
Build that kind of, You know, trust, uh, get everyone aligned. But how do you do that in, how do you do that effectively in a context where that's not always feasible to do? Um, you know, especially pandemic times, especially in open source where this is just kind of default the case. Um, how do you think about building that kind of very authentic trust in your team early without flying people all over the world to get together?
Patrick: Yeah, I think that. One of the most important aspects of trust relationships is time. Um, you don't build a trust relationship with somebody in a week, you know, or one conference. You may get a good connection with somebody like, oh, we were able to hang out. Cool. But, um, it's, it's, you have to be willing to put in the time in consistency for a relationship like, you know, a trust relationship to.
Uh, to build. And I remember, I don't remember who said it, but I, this is, I'm not the originator, which is trust is currency. You know, you trade that currency of trust and with an open source project. I, there's a lot of people on the Cassandra project. I've never met in person and, but I, I really have a good, strong trust relationship because we've had 10 years of working on the same stuff and I, I, I know their motivations and.
Uh, can trust those just based on a long history of working together. Um, so, but there are, you know, there are ways to make things work. I've done Zoom is your friend. Uh, I know it's hard to work with, but, um, it's amazing that I, it's just so cool that I can get on, like wake up in the morning, have a quick conversation with somebody in Finland, and then later in the day have a conversation with somebody in New Zealand face to face.
I. What a time to be alive, you know? Yeah. Um, yeah, we take it for granted. Sometimes we do. We ought not to. But yeah. H hp there are a lot of ways to build. Um, I think a long time ago it was, I think it was in a, I was in a book that I helped write on dev. I said, you know, some of the strongest, some of the most important relationships I've had in, you know, my career.
We're started on a keyboard.
Matt: That's the society we live in now though. Right. You know, I mean it's mm-hmm. It's just natural. It's more natural for my daughter to be, you know, disconnected in virtual than it is to be in person. Yeah. You know, it's 
Patrick: weird. Yeah. We can do it. It's doable. Yeah. But you know, that, that's another thing too, is that, um, you know, social media is a really good way to break trust relationships because it, it doesn't really create those kind of relationships.
It creates everything. Everything in social media is about short bites and getting attention and, um, Twitter used to have kind of this underground vibe, but it was still that, you know, it was still like people like throwing out hot takes and, you know, doing all that stuff. Um, and it, it was never a place where I, I can't think of any social media where I'm like, oh, I've built some really great relationships with people here, never.
Never. Um, so I, I think what's, we admit that, you know, even LinkedIn, LinkedIn is a great place to get a connection, but it's not like you're building any relationships there. Yeah. Yeah. I would agree with that. 
Avi: Um, I guess we are running a little low on time now, but I think, Matt, you had 
Patrick: one 
[00:52:18] Patrick's DevRel framework/focus
Matt: question to wrap up.
Yeah. So, you know, the, the other hat that you wear, Obviously Patrick is, you know, the Derell hat and so, you know, obviously before I delve into my, you know, kind of point here, how, how does, do you and, and does your team define DevRel? Because, you know, that is like one of those words that. Can take on a lot of meetings depending on the department and the company that you work for.
Mm-hmm. Um, so I'm curious, like from a DevRel perspective, what's, uh, what's, what, what's your take on, you know, your framework or what areas you focus on? 
Patrick: Yeah, Debra, I mean, sure, I can talk about this for another hour if you want, but, uh, I'll keep it concise. Um, you know, dev Rail is a cross-purposes. Uh, practice where you're looking at various aspects.
So the three aspects that I'm primarily focus on, there are usually four in Debra, which include product, but um, primarily what I focus on is outreach, community, and education. Um, those are. Three core areas of Derell that are important for what I do. Um, and the people that work inside those can cross over it a lot of times and usually they do.
And that's why we try to put it into one little burrito of Dere. Um, cuz the education people are involved in outreach, they're involved in community. Um, people who do outreach are involved in community. So, I mean, it's not. You're not just playing one game, you're playing multiple games that cross over.
But a lot of focus is put on each area. So outreach would be events, you know, education is like online learning or YouTube. And then, um, community is about, you know, community management and getting to know people, building trust, relationships. So that's how I feel Dev fits into my world. Yeah. 
[00:53:57] How DevRel is measured
Matt: So, you know, outta curiosity, because I am wearing the official metrics hat.
Yes. How do you measure 
Patrick: dev? Uh, a couple of, well, I have a lot of measurements that I run, 
Matt: but what are your top five? Do you gimme Yeah, tell, tell me what, uh, what, what, what, what you look at on a regular basis or what are the most meaningful, 
Patrick: the, the most meaningful are? Uh, mostly around engagement metrics.
Like how many people show up to an event, how many people engage with your content. I mean, cuz you, you're building content for people to use. You should be measure that, are they using it? And usually that helps a lot like, Oh, this didn't quite hit. This one did. Okay, so this is what the community's interested in.
Okay. We can make a choice there. Um, other metrics that I think are important are around the community, which is like how many contributed community contributed articles are there. I mean, that, that is, without that your community is gonna die. Um, it's just gonna wither. And so we look at how many contributed articles there are, how many posted use cases, that sort of thing.
Um, engagement levels there. So those are my two easy hot buckets. I mean, there's a lot of metrics inside those, but those are two that I think, um, without getting overwhelming with this data dump, I think those are really two that you should always consider. So 
[00:55:14] Advice for those getting started
Matt: with starting out, um, and you know, a lot of the audience that we have is startups.
They might be smaller companies, you know, they're, they're looking to build that community. From the ground up. Um, a anything that they should focus on early on, you know, because they might not have the audience, so they might be talking to 50 people. Right. You know, online, right. So, oh look, 50 people read this article.
This is, you know, twice as many as the last time. Um, you know, a, any sort of advice for those folks who are just getting started to building out their community and trying to make a go of, you know, some sort of project. 
Patrick: Yes. Uh, always start out with the idea that. Everything you provide is without a transaction.
So you're not asking here, if I give you some education, you'll give me something. Never start with a transaction. You will get value from your community later on. Don't lose sight of that. You know, this is, and Jono, if you're listening, I'm using your words. Um, this is the thing. It's like, um, When you, when you have your, you know, you have your outreach and your awareness, but when you get that new user, don't think that once you have a user that now, now this is like a source for all new content.
No, that will evolve. But just getting out of this idea that every community member is a transaction is such a good thing. Cuz you'll run people off so fast. Um, when you're like, oh, I saw you click the link. Would you like to do six use cases and fill out a blog and do all this? Like, ah, get away from me.
Um, Just think about that. And um, the number two thing is get out there and make friends. You know, go to conferences, go to meetups. You don't even have to create meetups. Go to ones that are already created. Make some friends. 
Matt: Yeah. Networking is so critically important. 
Avi: It really is. And so underrated.
Continue. Like highly rated and still underrated. Yeah. Consistently. 
Matt: Yeah. I agree. Yeah. Well, Patrick, thank you for showing up and chatting with us today. We really do appreciate your time. Um, thanks so much. Always great to catch up and, uh, we hope to see you at an upcoming conference. 
Patrick: Yeah. Where we can network.
Matt: There we go. There we go. All right. Alright, go. Next time don't forget to like and subscribe everyone. Thanks much, please. Yep. 
Patrick: All right.