Patrick McFadin: When your developer advocates look into the community, they should see a reflection and vice versa. These are your community members that just happened to get a paycheck at your company. And that is the best program you can have. Let those people do what your community wants to do. Enable them, empower them, accelerate them Eric Anderson: This Contributor, a podcast telling the stories behind the best open source projects and the communities that make them. I'm Eric Anderson. Eric Anderson: All right, we're live today with Patrick McFadin, who is one of the people most closely associated with the Apache Cassandra project, long time DataStax and Apache Cassandra expert. Patrick, welcome to the show. Patrick McFadin: Thanks Eric. Yeah, long time. Long, long time. Eric Anderson: In fact, I'm starting to feel like I'm an old guy in the community and I remember seeing Patrick present at a Strata years ago, back when we met in person. Patrick McFadin: Ah, that was a long, long time ago, back when Strata existed. Eric Anderson: Yes. Existed and was cool. So Patrick, as is customary, introduce us to Cassandra. Although it's one of these projects that I'm sure everybody knows, but what is it exactly? Patrick McFadin: Do they though? Well, it is a very popular project and it has been around for a long time, but Apache Cassandra is a database, first and foremost, and it's a database of records. So you would use it to store many things. So it's a general purpose database. It would be classified as like a NoSQL database. So it doesn't fit into the same world as say, a MySQL or a Oracle. Patrick McFadin: The thing that really makes Cassandra interesting is that it was born and bred around the idea of scale and distributed, and the idea of cloud. So the computers that you use are somewhat disposable and can come and go, so your database should be able to stay online. So it really embodies everything around scale, always on, the elastic workloads that we put against databases these days. Patrick McFadin: And in the past 10, 11 years that I've been working on the project, it's picked up quite a following. I mean, you look at the hyperscalers that are doing really cool things on the internet from back in the day, like the Netflixes, the Facebooks, and Apple. I mean, they're all using Cassandra, because that's the database that works. Eric Anderson: Awesome. And let's see if we can recreate some of the history here together. So [inaudible 00:02:38] Cassandra came out of Facebook, is that right? Patrick McFadin: Yeah. It came out of Facebook and it's actually, it has a little more of an interesting route. So there was one of the engineers, Avinash, was working at Amazon and worked on this original paper called Dynamo, the Dynamo paper. And that should be enough of a legend right there. But Avinash left Amazon, went to go work at Facebook and was working on interesting problems at Facebook. And he teamed up with another guy, Prashant, and their job was to build something similar for the Facebook Messenger and the Facebook inbox service, which no longer exist. But they were trying to make this work. And Avinash was like, "Hey, I just wrote this really cool paper." And Prashant was like, "Oh, I love Bigtable". I think it was literally like a lunchroom conversation and Cassandra was born. Eric Anderson: And Bigtable, reminds us, this is Google technology, right? That kind of evolved out of MapReduce to do a similar thing. Patrick McFadin: Yeah. Another one of those, "Oh my God, we have so much data. How do we deal with it?" And it was more about sequential workloads and large scale like search results, which is Google. Yeah. But yeah, those things combined together turned into this Cassandra project at Facebook, and then they eventually open-sourced it through the Apache Software Foundation. Eric Anderson: And just so we get to understand where your story comes in, when did you first run into Cassandra? When did you and Cassandra start dating? Patrick McFadin: It's funny, because my mom thought it was my girlfriend for such a long time. I'm like, "Nope, it's a database, Mom." I don't want to explain what a database is. That's funny you should say that, Eric. Patrick McFadin: So there was a friend of mine. He and I were having lunch and a lot of people will know him. His name is Adrian Cockcroft, and he was Chief Architect at Netflix, and they were trying to move away from the DVD business and go into streaming. And this is like January, 2011. Patrick McFadin: And he and I were both at a conference. We were both speaking on a cloud computing panel, and we had lunch and we're just kind of exchanging war stories. And of course his is way cooler than mine. And he's like, "Yeah, we're using this database called Cassandra." And I'm like, "Oh." I kind of heard about it, because at the time NoSQL is exploding. So you had to keep up. And that was 0.7. Patrick McFadin: And I went home that night, installed it, and I'm like, "Wow, this is pretty legit." And I guess I didn't stop. So I just kept going. But that was it. That was really how I got started with it. Eric Anderson: And at that point, DataStax was the thing. Do you recall how DataStax kind of emerged? Patrick McFadin: It was originally it was Riptano. Is that pre-Eric? Eric Anderson: Yes, it is. Patrick McFadin: Yeah, so here we go, some real history. Yeah. Riptano was the original name of the company, and it was a cool name. It had a rhino as a logo. But it was when DataStax was going to get a B round of funding that's when there was like, okay, A round was great, B round, let's let's think about adultifying the company and thinking about a different name, and moved from Austin to the Silicon Valley. And DataStax was kind of born out of that process. That was when we first came out with an enterprise version of Cassandra, DataStax Enterprise. So yeah, all that kind of happened in like 2012. It was boom, things kind of blew up from there. Eric Anderson: Yeah. And it was kind of the darling of the Valley and had all the things you'd want to see in a burgeoning company, It was used by the who's who, the Facebook, you name it, and powering their biggest applications, and it had Patrick on staff. Patrick McFadin: Yeah. That was actually the key component. It's like, "Oh, you got that McFadin guy. All right. You guys must be legit. Yeah. Let me just give you more money." I think you're right, because during that time, like in 2012, NoSQL was a fun word, but everyone was a little cautious, especially if you were doing more enterprises. Like that sounds great if you're Facebook or Netflix, and you have a bunch of engineers that are ready to go build it. Patrick McFadin: But the enterprises were a little adverse to this because they're like, "Eh, it seems a little risky." But there was also like 400 flavors of NoSQL out there. I mean, there were so many databases, like Tokyo Cabinet and Voldemort. I don't know if you remember those, but those are out there. Patrick McFadin: And yeah, it was this explosion of NoSQL. And in that 2012, 2013 timeframe is when things were starting to whittle down. The clear winners were emerging, and that was part of DataStax strategy was to, like I said, adultify it. Like, "Okay, let's turn this into what an enterprise needs. They need support. They need certified releases. They need all the things that make it... Security." And it created a transition point where we have our open source world, which I worked in quite a bit, and then our commercial enterprise world. Eric Anderson: Take a moment. We'll deviate from the history a bit to solidify the value of NoSQL, how it works, why people use it. We've talked about scale. And that seems to be the reason people were first kind of drawn to NoSQL, right? The idea that I can't scale my SQL database. I need something- Patrick McFadin: That was it. I was working in infrastructure before I started doing the startup product game. And in that early 2000s, we were trying to make everything scale. And what we were taking is these 30 year old databases that were great when we had hundreds of people using them, and trying to make millions of people use them. And relational databases are amazing for what they do. They're very flexible. They're very good about how they work with data, but it fits a certain model that we didn't need. And that was this highly coordinated, transactional-type systems that every transaction in the system is expensive, because it wraps all this safety around it. And I hesitate to use the word safety, because it doesn't mean NoSQL is unsafe, but there's certain guarantees that relational databases offer that are just overhead. Patrick McFadin: You know, it's like the undercoating on your car. Do you really need that? I mean, maybe? But what NoSQL databases do is they get to the core of the problem. They solve the problems in different ways. And so for instance, they eliminate a lot of the coordination tasks that have to happen whenever you write data into a database. They create a distribution so that you can do things in parallel. And you know, those are different ways of doing data. And they're all valid, but when you're doing millions and millions of requests per second, you have to use the right tool for the right job. And NoSQL started emerging. Patrick McFadin: And there's different flavors of NoSQL in there too. Some are more flexible, like schema free, that's MongoDB that everyone's familiar with. Some are very specific, like key and value, and that's Redis. And then there's this kind of hybrid, which is Cassandra, which fits in between those. But again, the way that they work is the use case. That's what you use it for. Eric Anderson: Yep. And how about, I always wondered if part of the trade-off was like, you mentioned that you don't need these transactional guarantees, but do you also not need the expressiveness of SQL in some cases? A lot of these apps are just kind of putting data and getting data, and the thought of a full analytical, be able to aggregate and group and return sums and averages, isn't what you need out of the database. Patrick McFadin: Yeah. Not at all. I mean, if you think about the things that you do, like with a shopping cart. I'll use probably the most important thing that everybody wants to stay online, which is a shopping cart. You don't want to lose that. That's where your money is. But a shopping card is put the thing in the cart, get the cart out. There's not a lot of calculations going on or anything that needs complex queries. It's here's my widget. Put it in the cart. Store that, and make sure that when I ask for it, it's there. That's the use case. And putting all the other things around it doesn't make sense. Eric Anderson: Yeah. Makes sense. Back to the history. So NoSQL is taking off, DataStax is taking off, and they were early in the world of commercializing open source, and also ran into some of the first situation. I mean, what am I trying to say, Patrick? Patrick McFadin: Yeah, what are you trying to say, Eric? Eric Anderson: Today we're in kind of another commercializing open source... Crisis is probably too strong a word, but evolution, where open source companies are trying to figure out how to navigate their [inaudible 00:11:32] crowd providers. Then DataStax ran into situations with the Apache Foundation where they were also trying to kind of sort out how you commercialize open source. Patrick McFadin: I gave quite a few talks back at OSCON, another conference that I'll miss. But I used to give quite a few conference talks around just open source licenses. And so if I was to characterize the struggle, it's this push and pull with the license that you're using and the concept of this free as in beer. So there's free as in freedom, free as in beer in open source. Patrick McFadin: And this free as in beer, you can just use this code any way you want. Go ahead. No, you don't have to pay us. The licensing is probably... That's an MBA right there, if you can figure that one out. Probably a doctoral thesis at this point, because there's just so many nuances, potentially, in how you license your software. Patrick McFadin: And you're right. We're going through yet another phase of open source project, that's attached to a commercial company, and they're relicensing their open source software to something that... Like right now, I think the famous phrase is everybody but Amazon can use it. That that's the license. [crosstalk 00:12:45] Eric Anderson: There's a license we're all looking for. Patrick McFadin: Yeah, the Amazon can't use this license. Patrick McFadin: But before them, there was others. I mean, there is GPL was where Linux was created. That was a license for Linux. And it just had some really strong restrictions. It was a very restrictive license. And the one that surfaced to the top for open source projects was the Apache license. And you didn't have to use Apache Software Foundation, but you could use the Apache license. And the Apache license is super open source-y. It's like, do whatever you want. You can fork it. You can run it on your cloud. You can do whatever you want, and the only stipulations inside of it was that if you put code inside of an Apache licensed project, you lose the rights to your patents. And so that makes a lot of people hesitate for a second. [crosstalk 00:13:34]. Patrick McFadin: But yeah, it's a very permissive license. And when you have something that permissive, it can cause trouble in finance. Like wait a minute, what? Yeah. Eric Anderson: And then the other part I wanted to ask you about was your role in evangelizing Cassandra. Today the path to evangelizing open source project is well-trod, and there's a lot of that going on, and people see a lot of value in that activity. I feel like you were kind of a pioneer in community building. Patrick McFadin: Wow. I hadn't thought of myself that. It's interesting because I was talking to another co-founder who is trying to jumpstart a DevRel program, and I just offered him this advice. It was in this world of, I was doing it. I was a user. Actually I was a consultant at DataStax. That was what I did. I was a Cassandra consultant. But I was just doing it all the time. And I happened to be talking about it a lot, as well. I would do meetups, I would do conferences, things like that. So I was doing both jobs. Eric Anderson: And when you say doing it, you mean just building Cassandra and collaborating with people on using and consuming Cassandra, right? Patrick McFadin: Yeah. I mean, I can not tell public stories, but I'll tell you some of the biggest installations that are out there, I was involved in building. And that was really a cool time. And yes, I was an engineer working on the project. And a lot of it was we were just trying to figure it out. How do we run a thousand node cluster? Petabytes of data? Patrick McFadin: But when you get that experience and you talk about in public, I could kind of clean it up a little bit so I wouldn't use the protected names, but I could talk about the experience. And that turned into being very valuable for our community. And Billy, our CEO at the time, he's like, "I want you to just do this." And I said, "What?" And he's like, "Just go out and keep talking about how people can do this and help build this community." And it was just a really different thing. And I said, "Yeah." And it was funny because we had this discussion, "I'll do it for six months. And then I'll go back to being a consultant." Yeah. Eric Anderson: My sabbatical from engineering. Patrick McFadin: Right. Exactly. And I never went back, because it's really just been that rewarding. And I still help people build their things and make things. But I do it in this open way that we can build community. And so like, what's the secret sauce is finding those people doing things. You don't want to just go hire someone and say, "Figure it out." That's like hiring somebody who could do training or documentation. You know, that's a skill. This is a passion. Patrick McFadin: I want you to understand how to make this work. That's what it's about. And then you and I are just going to sit there and hash it out and figure it out. And then when it's good, you got something and I feel better. Like my OCD's over. Okay. Eric gets it. We're good. Eric Anderson: Right. So if you were advising, it sounds like you just recently did, an open-source founder, you'd tell them, don't go hire a Patrick like me, who's maybe a career evangelist. But go mine your user base. You help empower people to accomplish something. And you find the people that are just super passionate about your project there, and help them go do it for others. Patrick McFadin: Yeah. There's two people. The two people that I say go after are, go to one of your conferences or a conference where people are talking about whatever you have, and go talk to them. And say, "You want to do this full time?" I hired a lot of developer advocates by saying just those words. "Wow, that was a great presentation. Do you want to do this full time?" And then a lot of beers, discussions. Patrick McFadin: But then the second one is actually inside, like product and engineering teams. A lot of times engineers that are helping build the product, sometimes they're looking for something different, and sometimes they show up. Like they writing blogs or things like that. And you're like, "Wait a minute. You're good at communicating. Would you like to do this full-time?" Again, have that conversation. Patrick McFadin: But it's the passion that you're trying to capture. Someone who really just loves what this thing is, whatever you have, and that wants tell people about it. That's what you're trying to put in a bottle. Eric Anderson: I had this question ready, and now it feels silly, when you put it that way. But I was going to ask you what these people actually do, as if it was a formula. And you're going to tell me, no, you just let them loose because they're full of passion and they kind of would know what to do. Patrick McFadin: Yeah. And it's really funny because I do a lot of DevRel conferences now and that's one of the questions is like, "What's a good developer advocate look like? And what's a good developer relations program look like?" It looks a lot like your user base. When you look at your user or when your developer advocates look into the community, they should see a reflection and vice-versa. There shouldn't be any like, "Oh, that's the corporate side." No, that should be, these are your community members that just happen to get a paycheck at your company. And that is the best program you can have. Patrick McFadin: And you know, what are the formulas? Well, the formulas are let those people do what your community wants to do. Like enable them, empower them, accelerate them. And this is a virtual cycle that keeps happening. And there's lots of ways to do that. Workshops, blogs, example code. I mean, these are all the established things to do, but how you do it is the most important thing. And that's the passion. Eric Anderson: So one kind of contrarian question, and then we'll move off the DevRel topic. But if it's just people in your organization, engineers, or just people in your community today, do you need to convert them into full-time folks? Or can you just tell all your engineers to be good at writing blogs, and if they want to do it, just keep doing more of it, but also be an engineer. Patrick McFadin: I mean, that works too, and I mean, if you're looking for a breadth of content creation, for sure. I think there's a lot to be said for someone who's spending all day every day, thinking about that, and it's not just a part time job. And only because it is not something you want to do as a part-time effort. Patrick McFadin: If you have a product that you want people to be passionate about, then you make that a part of your portfolio, of passionate people that are ready to do that with others. And if it's a part-time job, then you're getting great blogs, I'm sure. But what's the outcome you're looking for? Are you getting that outcome with great blogs? Eric Anderson: I'm also realizing there's value to having a face to the project. And like people love Kelsey Hightower, and in some ways he brings... What am I trying to say? People get excited about Kubernetes and then they get excited about Kelsey, but also they get excited about Kelsey and he makes them excited about Kubernetes. And there's something about his kind of fame that I think concentrating it in him, as opposed to just a bunch of team, a hundred Google engineers that you'll lose track of, there's some value in that. Patrick McFadin: There's a personality, for sure. And yeah, that's one of the things that has really been funky in my... I mean, I'm an engineer. Like we don't get out much, right? But then now that I'm in an airport and someone recognizes me. Like I was in a train station in Melbourne, or no, in Sydney, sorry, Melbourne doesn't have a train. I was at the train station in Sydney and some guy came up to me and says, "I saw your data modeling talks? Where am I at? Patrick McFadin: But I mean, that's cool because what you mentioned, like with Kelsey, Kelsey's creating connection. Eric Anderson: Yeah. Patrick McFadin: And you know, when I talk to Kelsey, I feel like he's just another engineer doing cool stuff. But I mean, that creates a friendly human face on something that is kind of cold. Technology is kind of cold. And building connections with people is really the name of the game. Eric Anderson: Yes. That's an aspect of community I feel like I always keep forgetting. We had the creator of Chef, Adam, Jacob, and he just- Patrick McFadin: Another 10X personality. Eric Anderson: Yeah. I think I asked him some questions about community and he was like, "You're looking at this all the wrong way. You're talking like it's a business. You're supposed to like apply these processes to a group of people. It's like, these are my best friends for the last 10 years. That's the community. These are my people. I care just as much about announcing their kids as I do the new PR, because it's a group of friends." Patrick McFadin: And you also feel an obligation to a community of users like that, because it's your friends. It's like showing up at the barbecue without any food when you said you were going to. You just feel bad because these are the people that you want to be good with. Patrick McFadin: So that community creates this, like I said, a virtual cycle. And it also creates a ton of conflict, because when the business is making decisions that goes against your own values, oh, that can be some drama. And it's something to be aware of, try to avoid, of course, but it happens, because it goes beyond just, "I'm doing a job." Eric Anderson: I keep saying I want to move away from this topic but I really like it. I'm advising an open source company who felt like the chat room was inefficient at answering questions, and so they should move to kind of this ticketing forum system and they abandoned the chat. And to their credit, the answers are actually much easier to come by. But a group of kind of rogue members were like, "I want to talk to the people though. I want to just like, hang out with people who are also excited about this project." So they kind of formed an unofficial chat room because they're like, "It's not about the answers. You think that's why I'm here? That's not why I'm here. I'm here to hang out with my buds." Patrick McFadin: We have a saying. Actually, this is an old saying, it's like water finds a way. This is something that wants to happen. And you're either in its way or you're helping it happen. Patrick McFadin: And yeah, that's a really critical part of the community. Yeah. I mean, if you start looking at like KPIs around how many people are chatting, oh boy, you're going to be really disappointed. But if you're thinking about an overall engagement, like are we engaging people? Are people there? Are they engaged? Then yay, go team, because that's magic when that starts to happen. And good for that user community for figuring it out. Eric Anderson: So tell us about Cassandra today. I feel like Cassandra needs to be kind of reintroduced to the world. Like there was the Cassandra, remember, from the early DataStax days, but there's a lot going on right now. Patrick McFadin: Foreshadowing. I do have a blog that's supposed to be coming out soon around Cassandra 4.0. One of the things that I think is really interesting about the project, it's 10, 11 years old and I think it's emerging out of its adolescent phase. Patrick McFadin: If you think of a database lifespan, it's like 30 or 40 years. I think that's realistic. And a 10, 11 year old database, it's coming out of those years. And it's so funny because when you talk about a new project, you're like, "It needs to be mature immediately." Well, there's time in service. Patrick McFadin: I'm sorry. It's not you it's physics. But you know, Cassandra 4.0 is getting ready to ship and it's been a long time since we've shipped a major version. And part of that was this soul searching that happened inside the project about what is a database. And it was really cool because it was like a really great sit-com, where it was the growing up episode. It's like, oh, the database is growing up because the project community, the PMC, the management committee, was really adamant. Here's what's important, is stability and correctness of the database. And we're going to do everything we can to make sure that this is the most stable database on the planet. And so it isn't going to ship until some of these major users are using it in production on .0. That's unheard of. Patrick McFadin: So I think that's a really interesting milestone and say, "Wow, okay, you're right. That did grow up a bit." Because it's not just throwing code over and good luck, and let us know if you find any bugs, PRs accepted. No. The project has really gone into a different stance about quality of this code. We should not ship until it is perfect as we can make it. So that's, what's going on with Cassandra now. Eric Anderson: Is part of that maturity, is it a function of the governance? So I think we talked a lot about projects that are kind of new and interesting, and they're all trying to figure out kind of their governance story. And Cassandra's had that nailed for a while. I mean, the Apache Foundation kind of gives that to you out of the box. Patrick McFadin: Sort of. Yeah. Eric Anderson: Is it maturing both, on the governance side too? Patrick McFadin: So in the Apache model, the Apache Software Foundation has guidelines and bylaws, but then they really put the governance of the project in the project management committee, the PMC. And so when you hear people talk about the PMC these are the people that have binding votes. And a binding vote can be a minus one, which means no. Patrick McFadin: And so this is what creates the push and pull inside of a project. You have to have consensus and you have to move forward. But the Apache Cassandra project has gone through a really interesting renaissance with their own governance by posting how we will govern this project inside the CWiki, which is for the Apache Software Foundation, provides these wikis for each project. But it's posted online for everyone to see. There was a huge debate about it online and really matured quite a bit in it's governance, instead of just vote on something and maybe it's... No, this is how we propose things. This is how we vote on them. This is what a release means. And this is another maturing point, which is really important. And the early projects just don't get that. But I'm very happy with where it is today. Eric Anderson: And tell us for those kind of who maybe are getting excited about the project. How do folks get involved? Or where does the community hang out? Patrick McFadin: It's interesting because of pandemic, where do we hang out? Online. Eric Anderson: Yeah. Right. Patrick McFadin: The last major conference I went to was an Apache conference, as a matter of fact. But what we're doing, I mean, mostly we're on Slack. There's the Apache software Slack, and the mailing list, of course. Patrick McFadin: But you know, there's a lot of active user groups out there. Like we see the meetup groups are still doing their thing, but they're doing it all online now. So for instance, the London Cassandra group is still really active, still doing meetups all the time, but they're all online. And I mean, if you want to get involved, any one of those forums, the mailing list, or Slack. Of course, anyone can come find me. I'm all over the place, and I will help you with that. And I think it's just finding your people that you want to get involved with and how you want to get involved. Patrick McFadin: There's really two ways. If you want to help build a database, that's the contributor side of contributing to the actual code base. But a contribution to Cassandra project can also be things like doing a great presentation. They can help out by doing meetups, by blogs and things like that. I mean, that's the kind of thing that helps others. And sometimes if you spend a few hours working on a presentation and it helps somebody, that is just as an important contribution, as if someone added a feature to Cassandra, the database, because you're helping move the project forward. Eric Anderson: A hundred percent. Patrick, anything we didn't cover today, you wanted to discuss? Patrick McFadin: Boy, I don't know. Where are we today? Like, what's going on? I think we were talking about earlier. Eric Anderson: That's right. Yeah. Patrick McFadin: Yeah. The DataStax like, the DataStax still around? Of course we are. We're going strong. Well you know, what's interesting is right now, we're talking a lot about Kubernetes and that's like the big deal for, I think, data. So I've spoken at the last couple of KubeCons around running data on Kubernetes. And I'm really excited about that community and what that is going to do. Astra, our databases of service, runs completely on Kubernetes. So we believe in this as a thing. We have a project called K8ssandra, which is running Kubernetes and Cassandra together. Patrick McFadin: But this is why I'm excited about the next 10 years, because we're building out these cloud native data infrastructures now that are nowhere near like what we did 10 years ago, where we were pounding out bare metal and hoping for the best. And it's a lot of really cool energy in the project now around, wow, we could just make virtual data centers with Kubernetes and deploy them on any cloud we want. And I remember I told you before, I think open source is going to start eating clouds. Eric Anderson: Yes. Yes. Patrick McFadin: That's it. Kubernetes is going to do that, because it's reducing clouds down to the three things that they really offer. Compute, network, and storage. Eric Anderson: And now that all the clouds are homogenous, I mean, there used to be these proprietary interfaces that tripped us up. Patrick McFadin: Yeah. And it's called Kubernetes! Eric Anderson: Yeah. Patrick McFadin: I like how Amazon decided to do their, "Well, we have our own version of Kubernetes." Fine. That's great. So does Google. Everyone has their own version, but it still runs the same Kubernetes. Eric Anderson: Yeah. As long as you can all take my same YAML file, I'm on board. Patrick McFadin: Yeah. If I can [inaudible 00:31:20] Kube Cuddle, deploy, and I'm done, then great. But yeah, that's a pretty exciting future around data on Kubernetes. Eric Anderson: What I think's interesting about that future is, there was a time eight years ago, five years ago, when everyone said, "We're going a hundred percent in the cloud. That's the future. We all just need to get in the cloud." And then with cloud native, suddenly people were like, "Wait, I can have all the cloud goodness here at home." And a couple of folks were like, "Maybe I don't have to put everything on Amazon. Maybe I can just kind of put Kubernetes here in my data center." And I think a segment of the market has kind of put the brakes, slowed down on the cloud migration because they can cloud native-fy themselves. Patrick McFadin: Yeah. And you know, we, I think we went down that path a little bit with OpenStack. We were like, "Oh, I can build my own Amazon using OpenStack." And I mean, it required a lot of hardware and a lot of expertise in building out those things. And now I see this great equalizer. Open source has always done this great job of resetting markets. You know, things like open source, what it did with operating systems, reset. I mean, we used to have to pay for operating systems. I don't know if you remember that, but that was the thing. And then databases. What if Google had to pay Oracle full ride on an Oracle database to build out their infrastructure? Eric Anderson: Yeah. Patrick McFadin: Impossible. Yeah. It wouldn't work. Larry would have had more than one island. But I mean, I see that happening now, is this reset. Patrick McFadin: It's happening again with cloud economics. Cloud is selling all these doodads on top of compute, network and storage, like proprietary databases that only work in one cloud and their flavor of the month, CI/CD pipeline, and that sort of thing. But it's just to try to get you locked in, but open source is going, "Yeah, no." Company X can say, "We know what we need to run and we're smart enough to do it." And Kubernetes makes it work. Yeah. So let's get disruptive at the end. Eric Anderson: That's a really great kind of analogy timeline, that open source kind of first took on operating systems and then databases. And now it's the whole kit and caboodle. Patrick McFadin: Yeah. And then what's next? I don't know. AI probably. Eric Anderson: Yeah. Patrick McFadin: Yeah. But now we're just speculating. Now we're just writing science fiction. Yeah. Eric Anderson: Patrick, thank you so much for coming on the show. As mentioned at the beginning, you've always been kind of the developer relations evangelism person in my eyes. What you've done for the community at large, these kind of computing communities, is impressive, and particularly for Cassandra. Thanks for sharing the insights today. Patrick McFadin: Yeah, absolutely. I'm glad I could share. That's what I do. Eric Anderson: You can find today's show notes and past episodes at Contributor.fyi. Until next time, I'm Eric Anderson and this has been Contributor.