 Welcome to Amsterdam and KubeCon, CloudNativeCon 2023. Join John Furrier, Savannah Peterson, Rob Streche, and UPSCOT. As the Cube covers the largest conference on Kubernetes, CloudNative, and open source technologies together with developers, engineers, and IT leaders from around the globe. Live coverage of KubeCon, CloudNativeCon 2023 is made possible by the support of Red Hat, the CNCF, and its ecosystem partners. Good morning, favorite nerd crowd, and welcome back to Amsterdam. We are at KubeCon, CloudNativeCon EU, and I am joined by my fabulous co-host and analyst. I've got Rob on my right. I've got John on my left. I'm Savannah Peterson, and this morning we have a very special guest who's going to bring a lot of insights. He's a contributor and a total baller. Jason Bloomberg, welcome to the show. How are you doing? Very good. Thank you very much. Awesome. I'm so excited that you're here. You've been spending a lot of time. We love sitting here on set, but we don't get to interface with people who aren't guest very often. So I'm really excited. You mentioned you've been spending a lot of time in the startup zone. What's hot? Well, it's great to see that there's greater focus on day two, that is full production operations. One of the challenges Kubernetes is it takes a lot of ramping up. Planning at day zero, ramping up is day one. But day two is where the rubber hits the road, right? Companies who are in full production, have multiple clusters distributed globally, and thus require infrastructure that can support that in production in a live environment. There's increasing focus on that, both from the vendors as well as the attendees here. So we've been talking a lot about the tipping point and the momentum. It feels very different at this, KubeCon. Do you feel like the ecosystem has matured to that next level? Well, I think what's really different about the feeling here is that almost 60 percent of the people are first timers at KubeCon. So as a result, that 60 percent, they're not the vendors who always come to these shows. They are the enterprise end users who are either infrastructure people or increasingly the application people, application developers. And that is a new audience for KubeCon. For the most part, it's been all about infrastructure. So you have the infrastructure vendors and infrastructure engineers talking about infrastructure. But the whole point to the infrastructure is building running applications. So you can't really get to that until we have that application-centric audience. And this is the first KubeCon I've been to that has really a critical mass of application people at it. Jason, I want to get your thoughts on your discoveries and conversations around what their people are talking about. Because obviously it's that tale of two KubeCons right now. We were talking about this on our first day and second day. The submissions for speakers were as of November. And Chad GPT took the world by storm in January, February. So a lot of AI conversations in the hallways, but not so much on the tracks. What's some of the conversations you're seeing around the companies, around what they're focused on. Because this is like an infrastructure crowd. Obviously the practitioner side, day two, means more operationalizing it. But AI is now in there. I was just talking to Sajiv, who was on early, a former gardener analyst, about data. This is not a data-centric crowd. Is there a lot of data conversations? What are you hearing? What's some of the AI buzz, if any, or is it, was it muted? I actually think what's remarkable is that there isn't a lot of AI buzz. You would sort of expect generative AI to be the hot story. But that's not really what's going on here at KubeCon. If anything, there's less talk about AI or generative AI in particular than you might think. I've talked to a few vendors who rushed some sort of generative AI thing into production, but for the most part, this is a very, I guess, very down-to-earth realistic crowd. And this is something that's been true of KubeCon all along, is that these people are focused on technology that actually works, that can run today, that actually meets customer needs today, as opposed to a lot of sort of marketing hand-waving. So, even though- I mean, chatGPT isn't, I wouldn't say hand-waving, but there is a lot of hand-waving. I mean, the prompt, it does, it's pretty cool, but it does have hallucinations, Rana. So, these are infrastructure people. Rob, we've talked about security. I mean, you can't have a gap, like, oops. Well, I think what's been interesting is that I think over multiple guests and the few times I've been off the set to talk to people, it's been security is really still an open issue. And I think to your point about day two, especially how do you productionize apps in a secure way? And that's been another theme that I've seen outside of some of the big guys who've been here a lot, but even they are talking about everybody has their security bent to what they're doing as well. Are you seeing that with the others and the startups? Well, absolutely, but that's nothing new, right? Security has always been a top priority and it's part of every conversation to some extent. So that's important and it continues to be important, but it's not something that's really new. That's a change from previous years. Well, I wanted to bring up the AI thing, mainly because I want to get your reactions to what you're hearing. Because I can agree with you. Actually, we've been talking a lot about AI only as a frame, which is trying to understand where it will land, how it will meander around, what it will be focused on. But to your point about day two operations, this is not just testing and dev, it's like full ops. So there's no tolerance for BS and whatnot, right? Well, yeah, exactly. So, and it's one of the reasons why I like coming to KubeCon is that it's a very low BS show, right? You're not going to run it. Low code, low BS. Well, you're not going to run it. There's never been talk of blockchain here, which I love, right? Oh, bless. If there's ever a blockchain at a show, it's like, okay, that's a BS show. So this is nice to have a no BS show. And I think with Genitive AI, obviously- I feel a little bit of BS on the cube. Okay, that's fine. Genitive AI, obviously, has very impressive capabilities, but it also has sort of that dark side, right? The hallucination. I 100% agree with you. And we're just figuring out some of its negative implications. So yes, you could very well apply it to, you know, I don't know, day two operations. You know, how can you configure very broad, complex set of technologies and you might leverage AI to optimize that. But the last thing you want is for AI to go off in the wrong direction, right? And recommend something that doesn't make sense. So everybody would be very careful about doing that. I think that's a good point. I think that's one of the reasons why we took a kind of a frame approach, because I look at AI as more of the weather coming. We like the forecasters. We're sitting in the sunshine right now getting the work done. We know it's coming. The question is, where will it land? I always say security is one area, but like, if you connect forward day two operations and saying, okay, assume there's some augmentation. This community knows automation. They like automation, no policy. Where would AI fit if it had to fit anywhere? Like, what would it, like, what would you imagine that being if you had to kind of connect the dots? Assuming it rolls forward, everything's kind of going on, where would AI help this community? Well, AI has played in a very important part in anomaly detection as part of the AI ops story and that's been a story for a number of years. So that's well established, right? That you can, if you have a large quantity of operational data, you can pick out the anomalies and AI's very good at that. Machine learning in particular, you don't need generative AI for that. The generative AI story could easily help with the overall context of configuration in a complex environment. So you could, instead of having an engineer have to figure out what command lines to type over and over again, they could say, well, I want to set this up. You could say in English, just expressing themselves, I want to set up this environment to achieve these goals. Here's my data sovereignty goals. Here's my security goals. Here's my latency goals. And then hopefully the AI is smart enough to configure it properly. Or maybe it'll go off into left field to do something complete off the walls. You don't know, right? Well, so I kind of want to dig in here a little bit because building on your weather analogy, I mean, yeah, there's a forecast and we can see it coming, but we've had the NASA climate scientists on the show and I got to ask them quite literally, why is the weather so hard to predict? And the answer is the butterfly effect. Anything can happen. And I feel like we're very much in, you know, I'm a little more nervous about some of this, I think, than some of my co-hosts because I understand the dark side and the bias pretty well. I'm wondering what sort of safeguards we're going to end up putting in place and the governance that's going to be necessary to make sure that nothing nefarious happens with this. I mean, it's a very much a wild, wild west stage. No one is an expert yet right now on what's going on. It's all too brand new. Have you seen anything? And this doesn't have to be AI centric. Have you seen anything at the show that really surprised you this year? Well, not specific to AI, but there are some other trends that are new trends. As you talked about security, one of the important security trends is the software bill of materials or SBOM, right? Or just dealing with supply chain security in general. That's been a problem since the sonar type and the log for J reaches. But now with the US federal government mandating their cybersecurity program, that's coming up in June and then the year of his doing something similar. It's now urgent, right? That the SBOM, especially for regulated industries. So there are vendors here that are focusing on that and SBOM is slipping into a lot of the messages. So another key theme is platform engineering. It's a very- We've talked about it a bit. It's Rob's favorite topic. And a number of vendors are talking about that. And so it's interesting, on the one hand you have the platform folks saying, well, our platform is great for platform engineering. Use us or incorporate us into your platform. And then there are other vendors who are saying, wait a minute, we want to be on the golden path. And this is interesting. I haven't seen this before, right? The golden path is essentially the established set of best practices that a platform engineering approach will not mandate, but recommend, right? You can always diverge from it, but it's going to be easiest to follow that golden path. So now all the vendors who have a piece of the story, whether it's CICD or the day two operations are saying, you know, I want to be on the golden path so that any platform engineering approach that you take will include us in that best practice approach. And so that's interesting. So everybody wants to be one of the selected golden children of the platform engineering. Yeah, I think that one is pretty interesting in that there are, to your exact point, there are so many different pieces. And I think it also just tying it back to, because I think AI will help in this space in the platform engineering space in particular, especially with SRE work and some of the other stuff and tracing and observability, not to mention the security aspects of it. I think the one thing that I'm finding interesting from the people around is getting a big enough data set of their own data, so that becomes their IP. Because I think to the earlier problems with open, very large open models, I think it's how can you have your own data that gives you that? Have you been seeing different companies really pushing in on collecting that data because it's a GDPR and everything else has some very gray areas in some of that data? Well, the story around telemetry data is maturing quickly, right? Open telemetry is an open source project that is really ramping up very quickly. It's getting a lot of adoption and acceptance. And that now relates to operational telemetry. So logs and traces and events and metrics. But that's only part of the story, right? There are other kinds of data that fall outside of telemetry. And that's where it gets interesting, right? You could consider essentially all Kubernetes metadata as being data that you could feed your AI. So all of the YAML files, all of the Helm charts, all of the manifests, think of all of those as data. Now feed those into your generative AI and now you have essentially a large language model that will basically converse about how best to configure your environment, right? Maybe, or maybe it'll go off, or maybe even troubleshoot it. And I think that's- Well, actually, it has to be known configuration. It can't be like, do something for me. It's got to be the fundamental Lego building blocks, not so much. I want to ask you on the platform side, because you just brought that up, the platform engineering. On the keynote this morning, the co-chair, a partner who was on with you yesterday, she asked the question, how is Kubernetes doing with platform teams? There's a whole team discussion, what's the right make of a team? But the question was, is Kubernetes delivering on its promises and where is it working and where does it need improvement? How would you answer that? Well, Kubernetes is becoming, I guess, part of the fabric of enterprise IT infrastructure, right? So we're still talking about it as a thing, but I think we're getting to the point where we're going to be taking it more for granted. The same way we take TCPIP for granted, or now we're taking Linux for granted. We just assume that it's there. We could talk about it if we need to and everybody here knows about Linux, it's an open source crowd. I think Kubernetes is sort of following that pattern. We're just going to assume it's in place. And when we talk to some of these enterprise Kubernetes end-user companies, it's basically just part of how modern approach were for deploying and provisioning application capabilities generally, right? Why would you use any other approach, right? Now that it's still part of the infrastructure and you have to layer on a lot of other things. It's a layer cake, but you're absolutely right. But it's just, there's no reason not to use Kubernetes if you want to deploy enterprise software at scale. So, Rob, I want to get your thoughts because I'm looking at the Twitter feed right now. They actually have a graph and they have simplicity, scalability, extensibility, reliability, and portability. That's the rank of the who gets the most stars of gold stars as a performance. Simplicity gets the lowest grades, obviously. Decreasing complexity, number one complaint. Scalability, it's got three stars. And extensibility, five stars, and reliability, portability, five stars. So, you can see kind of the trend patterns. The simplicity question is huge. We've seen a lot of managed services come out of Kubernetes. We've heard teams want to focus on other, more critical DevOps or DevSecOps things and can someone else manage my Kubernetes for me? What do you guys see in terms of that simplicity piece? How does that get solved? Yeah, I think what's really interesting is that portability also got five stars where I don't look at portability as being solved yet. And I think going between different flavors of Kubernetes is still incredibly hard. I experienced that most recently and I can tell you it's not easy. So how many stars would you give it? I would probably give it two on the portability part of it. Scalability and I don't, was it serviceable? I don't know. Reliability, extensibility, scalability, simplicity. Yeah, so simplicity I think is still, it's when you're, I think it was said very well by one of our guests from Red Hat, the fact that we're still showing YAML files on a keynote is ridiculous. I mean, if you want to- Can I just say amen to that? This is one of those things where we very much agree. Yeah, we got to jump the shark here and like really get to a usable system where you don't have to write in YAML and upload it and then figure out what you have a typo in there and stuff like that. I think that to me, we still need to get there. I think there are some people on the floor who are bringing that. I think it's an evolution. We saw it with VMware, where everybody wanted to use CLI for years and years and oh, then we'll use Power CLI and then it moved into the sphere and into the UI and I think that's when the people aspect that they talked about in the keynote yesterday, more bigger tent type of discussions, people with different assets, platform engineering becoming the new IT. Do you agree on that portability piece? What do you think about those grades? I mean, they're self grading themselves so everyone always thinks they're excellent. Yeah, I think we ever came up with that on this. Yeah, it's a little bias, maybe a little bias. Yeah, I really wanted to make sure that Ildy was at the end of every word. And so they, limit, they... Little marketing sprinkled it. Hyper, a few other important things. For instance, dynamic capabilities, I think that is a critically important part, right? So the way I see Kubernetes is it delivers two things very well. It delivers scale and it delivers dynamic software capabilities, right? The ephemerality of the microservices enables organizations to deploy software much more dynamically and enables software to behave more dynamically, right? Because of the auto scaling built into Kubernetes. So, and that's something that delivers a whole different set of capabilities for software that it can scale up and scale down and it enables large organizations with multiple teams working in parallel to be able to deploy software in a consistent way without getting into each other's way. Now, that requires a lot of moving parts. It requires a get ops mentality. It requires the CI CD being placed. But all of these pieces have to fall into place. This is all part of the cloud native story. It's that once you have all these pieces in place, then you get scalability and dynamic capabilities at scale. And I think those are more important. Portability is a piece of the story, right? Workload portability is a priority for some organizations, but I don't see that as being high on the list compared to some of these other things. Well, one of the things I want to get your thoughts on before you go, I know you got to get out, but this show I just saw on Twitter, the data backup and recovery is being discussed. Now, Rob, when you hear a show talking about data protection, it's mainstream. I mean, this is like a vendor show now. This is not just Kubernetes conference. IBM's here, Microsoft's, Intel, Portworx. This is a trade show. I mean, this is like an industry show. This is like a big deal. I mean, it is quite literally an industry show. I mean, this is not just KubeCon for developers of open source. It's become VMworld almost. Well, it's a really big show, yeah. I mean, VMware is no longer a really big show. That last one, they don't even call it VMworld anymore. It's called VMWare Explore, they downsized it. Yeah, they downsized it, but this one's being upside. So this is to the trends of mainstream, so it's like, when I see stuff, when I start to see some of the stuff we see at these shows, it moves out of like, nerd developer configuring, but yet, YAML files are still on the keynote, yeah. Well, geez. You know, Rob, just a quick question for you, because you mentioned it, many different flavors of Kubernetes. What does Kubernetes taste like? It's sweet, not sour. I love it. Jason, last question for you. We do a very important segment here on theCUBE all about the swag on the show floor. Have you had, you've seen all the booths, especially in the startup section, standing out, differentiating, very important when you're trying to attract attention of these enterprise customers that are here. Do you have a favorite piece of swag from the show? Well, personally, I don't collect swag anymore and my kids are grown, so there's no need for it, but the one that stood out for me, there are these Lego kits, Lego Star Wars kits, and I saw the R2-D2 one, and it's for ages 18 and up. So it's like, here's a Lego kit that is designed, not- You've got to be a voting age? You can't be a child. You have to be a voting age, Jesus. So what's Lego thinking? If they're building stuff for adults that are only for adults, it's like, what's this world coming to? It's their nation. We need to liberate the Legos for the kids. Yeah, I was gonna say, we can head to their head, let's have Legos for the kids. Wow, I think kids would like this stuff. Yeah, well, I'm with you. Everything should be for the kids, for everyone of every age. There's no need for ageism. Jason, thank you so much for joining us. You were the perfect start to the day. John, Roff, always a pleasure. You just offended my inner child there with my Lego, you know. You're 18 plus, you should play with it. All right, we'll let you go and nurse your wounds. Thank you all for tuning in to day three of KubeCon EU coverage here in Amsterdam. My name is Savannah Peterson. You're watching theCUBE, the leading source for emerging tech news.