Cloud Foundry Day 2023

Episode 267 · Fri, 23 Jun 2023 16:50:32 +0200 · 22:15 · SRT · archive.org · all Tanzu Talk

Well, it's very hot here at the Cloud Foundry Day here in Heidelberg, but we just concluded. Why don't you introduce yourself? Yeah. Hi, I'm Nick Kuhn. I'm a tech marketing architect at VMware Tanzu, and I do things for Tanzu Application Service and Cloud Foundry alike. And yeah, you are correct here. The temperature here in Germany is much hotter than I expected. I miss my United States air conditioning a little bit here. The weather's hot, but the water's cold. Now, I don't know if this happened to you, but what'll happen is I'll get hot and I'll want a refreshing, crisp, clear glass of water. And then I'm like, I don't know the difference between a bottle that's blue and red. So I'm going to go with the blue one because blue is water, right? Yeah, exactly. And then you open up the blue water and you take a big swig and you're like, ah! It's like I've drunk some water gravel because you got the sparkling one. Yeah. Which I guess there's people who find that refreshing. It's very popular over here. Yeah. They had a lot of good refreshments here. It was just kind of a guessing game, picking what you're going to get there for sure. Very well run as always. Yes. The foundation and the analytics foundation are absolute professionals at running events here. But what's the, so last time, when, where was the last Cloud Foundry Day? Detroit. In Detroit, right. Now we, we, we had been on, on the Tanzu Talk episode at the time, but we had you report about, as I think I called it, the vibe of Cloud Foundry Day. What do you, what do you, what do you think the vibe was here? Oh, wow. I think it was a great vibe today. So we, we sold the event out. So in-person capacity was, you know, hit the maximum and definitely from the climate control for sure. So that was great to see, you know, you know, entirely, entirely packed full room of, you know, Cloud Foundry users, you know, contributors and that type of thing. So it was a very good vibe today. Lots of really interesting talks. We get some like low level talks, high level talks, so kind of talks for everybody around the community and really kind of good to see a great turnout, you know, kind of, I think

we're really kind of getting past the post COVID, you know, conferencing and people are starting to get back out and, and really get, you know, attendance back up. For sure. And I want to see everybody in person and chat and, you know, do that kind of peer to peer, you know, get together and, and I guess, you know, marketing per se and not to, you know, see, see what's going on with their peers and community management, community management. Yeah. Yeah, exactly. Yeah. You know, I, I think it is, uh, I don't go to as many, well, I don't know. I don't actually watch as many conference talks as, as I used to just, just cause I flit around doing whatever, but it was fun to see like technical talks. And then, and then the thing that I always like to see, well, like to, that's encouraging to see in a community is there's actually a lot of, uh, detailed Q and a about like, what about this? And have you considered that thing? So every now and then someone would go over a, uh, a new idea or feature and there would be a, uh, uh, a fun discussion about like, oh, well, what's the plan for that? Like, how, how do we introduce that? Or how do we deprecate this other thing? And like, so it's, it's, it gives you that sense of like, there's a, there's a lot of like actual work being done. Uh, which, which is very nice. Yeah. So we definitely, you get that sense of, uh, the scale that some of these, these companies are working at. And then those questions back and forth about, well, how would we actually implement it with, you know, X, Y, Z or like, well, you know, that question's rooted in like, you know, serious either production use cases or, you know, they've been doing this for however many years and like, you know, the very kind of detailed and specific questions. So it was great to see that kind of back and forth between the presenters and, uh, you know, the, uh, community members and, you know, people would rotate and then, you know, they'd be asking each other questions back and forth on, you know, as different speakers would go up and that type of thing. So, so it was great, great to see that and that community engagement for sure. So have you, have you had a chance to catch up with, uh, like, like your people at SAP and 89s and other areas? Like, do y'all, do y'all have like a secret, like cabals where you get together? Oh, I don't, I don't know about that, but I, I would think this is probably the, uh, the,

the event of the year for sure. Uh, um, but yeah, I, you know, I, I flew in yesterday, so I got a few hours of sleep, uh, you know, coming from the States. So I'm hoping to catch up with some more folks tonight at the reception. Uh, hopefully it's a, there's some climate controller in a little cooler setting than today, but, uh, yeah, at least some ice cold beverages. Yes. Yes. I think would be in store. Plenty of those. So in, in your talk, I mean, you, uh, you were highlighting just, I mean, in addition to a few other things, you only had five minutes. Yes. Yeah. But, uh, you, you went over a lot of good momentum and usage, uh, sort of things and, and, and, uh, like go over, I mean, I tell me if I'm wrong, but I think you were kind of proposing a type of metric to use. Yeah. To, to, to gauge efficiency. So what, what was it? Yeah. So I guess we were coining this platform engineering efficiency or, and really that's, uh, equating to how many developers can a single platform engineer support. And really within, you know, the cloud founder community, we've been doing this thing called platform engineering for quite some time. Uh, you know, platform as a product was really big, you know, kind of in the pivotal days and, you know, running a platform team, that type of thing. Uh, so it's really good to see that. And we've started kind of, I don't know, I guess, calculating and interviewing our customers to start to track this metric and, and see what on average of like the Tanzu application service customer base, how efficient are our platform teams and how many, you know, actual developers they could support with, you know, X number of platform engineers. And on average, uh, I think we see, um, about 190 developers, uh, tied to one platform engineer. And then, um, um, from a team perspective, we'll see about six platform engineers supporting about 1200 developers. Um, and even in some of our, you know, really robust customers, we've seen, um, that, that ratio go from, you know, a single platform engineer supporting or assigned to 500 or more developers within organization.

And I think that's a kind of a critical metric to track for really any platform. But as we talk about, um, how efficient your platform is, how, you know, you know, a cost effective, um, that could be. Because if you talk about looking at potentially other alternative technologies, are you going to be able to achieve the same amount of work and value with, let's say a cloud founding platform versus maybe a do it yourself, Kubernetes style, that type of thing. And, you know, I've seen firsthand and we've seen some customers chart chart to go down that path and that efficiency ratio drastically changes and they end up wiring, you know, we'll get to a situation where, uh, more platform engineers are needed to support similar workloads on, you know, alternate technologies and that type of thing. So that's, that's kind of what I'm keying up on about, uh, the platform engineering efficiency, um, within Cloud Foundry. Um, and that's, you know, that's kind of highlighted my talk. Uh, I did, did, uh, release an article on the new stack about this as well, where we kind of dig into, uh, more detail about that too. So I'm sure we could put that in the show notes or whatever to link to, but, um, that's definitely worth a read if you want to kind of dig into more of that. Uh, and definitely thanks for the assistance and, and, uh, and helping and helping get that publish and, and, and coaching on that as well. Oh yeah. Yeah. Yeah. Well, it's always fun to get, uh, something written out there. That, that was a good piece. Cause you know, it's, it, it's, it's, uh, the written form of a lot of what you just said, right? Like, you know, here's, here's a lot of the usage that the, the Cloud Foundry, uh, community has like, you know, various, uh, distros of it, so to speak. And then, and then also introducing that idea of, of a metric to track, which, you know, I, I, I just gave, uh, I gave, I gave the last talk and that, you know, you were making me, you're kind of what made me think of be clear about like showing value in the metrics that you have. Right. Cause I think if you do come up with, I don't know, I mean, historically it seems like you want to have, it's always an odd number. You either want to have one, three or five, seven is a little too show offy, but you want

to have like one, three or five metrics that you can kind of track across, um, organizations to come and, and across different types of platform stacks to kind of like you're saying benchmark and compare and contrast what they are. And it seems like if your focus is running applications, which is typically done by application developers, maybe in the future, it'll be robots that do it right now. We've got actual application developers. But, uh, yeah, I mean, that is the efficiency, the cost, um, cost is such a weird word, but the, the cost versus value that, that you want to track is how much you're putting into a developer, an application, or how much, not that you're putting into it, I'm wording it wrong, how much you, uh, how much it costs to support each, each developer and the applications that they run. Yeah. Which do you, do you think, do you think we ever like think about developers too much? Like, like maybe one day we should invite a developer over and to kind of like de-specialize this mythological like thing. Cause they're always like, Oh, the developers, we've got to build all the stuff around them and make sure their pillows are fluffed at night. Well, yeah, so that's an interesting point because I think it, uh, you know, the developer experience and getting happy developers, you know, you know, letting them, you know, ship code faster, focus on business logic rather than infrastructure, that type of thing. I think that's, that's definitely, uh, uh, very, uh, merited and valued, but I think there was a point in time where, um, you know, DevOps kind of, you know, became the thing and the model was, all right, well, you know, we're slowing things down from maybe a centralized IT perspective. We're just going to hand out keys to all the clouds and developers. You can do everything like you, anything you want to do, you can do, you know, build that, but Oh, by the way, you're going to own it. You have all the operational burden now that a traditional team might have taken over. So I think that's what, you know, what we see like the, you know, DevOps is dead versus the platform engineering, that kind of methodology. I think we're kind of getting to a point where, um, you know, we need to still take care of developers and, and try to deliver things faster for them, um, as like platform teams, right.

That that's still a value. Um, and not just, you know, I think a lot of times two developers will even find coming out of potentially like, all right, well, I was a developer and I was given, you know, uh, keys to, to Azure and I own, you know, subscription and do whatever I wanted. But then I found out I had to do all these things. Like how do I get, you know, certificates or, you know, get my workload certified for certain security requirements or what's load balancing? What about DR plans and backups? Like all kinds of traditional IT type things that, you know, platform engineering teams would typically take care of that. I think that's kind of where we're coming from. You know, platform teams are good. You need to make your developers happy. Um, but you know, you know, things like cloud foundry and other platforms really kind of provide a good foundation to, to do that. Yeah. Yeah. Yeah. You know, you know, as, as, as you're, as I'm listening to you, it's making me realize that like, well, two things. One, I've never really good. When I think about this stuff, I never really fairly balance out like operations people and developers. I mean, like, I mean, you know, I used to be an application developer, so I'm developer biased, but you know, it's, it's almost like there's a certain, and I guess, I guess this is kind of like where DevOps has arrived is DevOps exists for the operators, right? Like, like it's a way of making people, whatever you want to call them, sysadmins, operators, infrastructure people. I think Forrester has a charming, I know, like, like it's some, uh, you know, monopoly railroad or something. Uh, but like DevOps really caters to and takes care of operations people, right? Like it's sort of like gives them tools that are better, automates things, makes what they deliver work better, but also like has a huge focus on, you know, like not burning out and like having, you know, uh, a good culture in place. And I don't know. I, I almost just sort of like, don't focus on that at all. Right. And, and, and I more kind of focus on the, the like valuable application developer and just sort of like almost like getting out of their way. But, but there, there's, there's something, it'll be interesting to see if, if the, the

like new platform engineering community, like kind of figures out how to finally have like dev and DevOps a lot more than ops and DevOps, which all my old DevOps friends always roll their eyes when I position it that way. But like, you know, it, I don't know if we've ever really figured out how to treat each parties the same in one community. Like they typically bifurcate into their own little communities to get taken care of. Right. But I don't know. I lost track of my thoughts. No, I think, uh, I think you can maybe point that to where like the platform engineering community now, uh, those teams running those platforms can subscribe to DevOps methodology and tools, but you don't necessarily want your business or your application developers getting stuck in that world of, all right, I got to figure out, you know, Terraform or Bosch or all these different, you know, tooling. It's like, that's not their core focus. So you really have this like, you know, platform or infrastructure engineers building things for develop, figuring out kind of all the hard things, all, you know, how to deploy cloud services, how to wire that up with their security practices, you know, you know, all those, you know, requirements, um, uh, for, you know, uh, uh, well, I'm blanking here on the identity management, things of that nature, just trying to get that all taken care of for developers, uh, so they don't have to worry about that. And then, and that also provides some control for an organization because if you, if you let that be on each team, they're going to build the wheel over and over again, it's going to look different. And then from a compliance perspective, it's going to be harder to, uh, you know, offer any kind of guidance or advice if there's something that's efficient. Yes. Or it just works. Yes. Yeah. So if you can get something that's efficient and works and provide that as the foundation, I think that's, that's, that's, that's the key here. So I, I, uh, uh, so, so, so have you had any conversations about like the, the, how do you say it? The, the, the, the Corfee, the running, the running, the Corfee, there you go. Corfee sounds like a terrible type of European yogurt.

There's, there's a few types we don't have back in the States. You should try to check out what we're here while you're here. It's a little, little interesting. Okay. But, uh, like what, what, uh, there were a couple of sessions on that. Like, like, like, but what's your, what's, what's Nick's overview of like, what's the deal with, uh, Cloud Foundry and Kubernetes at the moment? So definitely the, uh, you know, current iteration is Cloud Foundry Karifi where, uh, you know, it, it's, it's taking a lot of the core components of the Kubernetes ecosystem and integrating that directly instead of, you know, previous attempts kind of attempted to shim parts of, uh, Cloud Foundry into Kubernetes and that didn't quite work out well. So I think on, from a architectural and engineering perspective, uh, that's a good, a good approach to take. And I think we're starting to see some traction with, with Karifi. So that, that's good. Uh, we obviously have, um, within VMware Tanzu, we call this the application service adapter for Tanzu application platform or sometimes, yeah, sometimes we just call it the TAS adapter for short, uh, to really kind of get that CF push experience still, uh, on a Kubernetes type platform. So I think there's definitely interest there as well, but I think we're also seeing as, uh, you know, value in just the core Cloud Foundry for VMs and how efficient and how widely deployed that is that maybe, um, the enthusiasm for, you know, just running things on Kubernetes because it's Kubernetes is starting to just die out a little bit in the industry. Really? Uh, maybe, maybe that's just my own personal opinion, but I don't know. Like, uh, I think, well, and then we got things like chat GPT coming in here too. So it's like, man, Kubernetes, let's, let's talk about generative AI and doing AI things about with everything, right? We need Kubernetes GPT. Yes. Yeah. I'm sure no one's ever made that joke. Yeah. So I think we're definitely making good progress there. Um, and, uh, I think, you know, and the Paquetto build pack, uh, project within the Cloud Foundry Foundation has, has, has got a lot, a lot of, um, uh, growth and we're seeing, you know, that, that build pack that, you know, that kind of core foundation of some of the Cloud Foundry tech, uh, from, you know, years ago kind of being re, uh, rebuilt and, uh, re-presented

in a way that's, uh, for, you know, pretty much any use case. So, um, we even saw IBM has picked up Paquetto as part of, um, uh, I forgot the name. They had an IBM cloud product that they're, you know, they're starting to use Paquetto, uh, build packs within their product. And, you know, you can use it on your laptop. You can use it within Kubernetes. Uh, the Cloud Foundry, Karifa uses that, uh, to kind of bridge that gap of, uh, build pack support. And we're even rolling some of that back into the, um, Cloud Foundry proper for VMs. Uh, so that, you know, the OG Cloud Foundry, as you say. Um, it's getting some of that cloud-native build pack goodness as well. So I think that's good to see, um, that going forward. Yeah, I guess, I guess, I mean, this, this, this role doesn't necessarily exist in any open source role, but if I was doing some kind of like corporate strategy for Cloud Foundry, it would seem like one of the top initiatives would be like, let's make build pack, cloud-native build packs, whatever I'll say, build packs. But like, let's, let's make the build pack idea. And also whether it's cloud-native build packs, I mean, I guess it would be like, let's make that big in Kubernetes, right? And so like, even though that's not exactly Cloud Foundry, it's like a Cloud Foundry thing. And let's make sure that it's like, let's do all we can to make that a major way of configuring, deploying and packaging, do, doing the thing that it does to kind of like introduce that into that world and have this kind of, uh, not kind of, but have this similarity between the two of them. Right. Exactly. And it seems like it would be a good, uh, a good by, you know, it'd be useful for, it's useful for any application, uh, deployment thing. And it also be, it'd be useful to like be involved in the Kubernetes community to see what's going on there. And then the other way, it'd be useful to kind of like take some of those lessons and the little, like, uh, you know, positive burrs that come attached to it. Yeah. Into that world. Yeah. So we're, and two on that, that front, we're definitely, um, starting to consume, um, some of this, you know, newer tech back within the core kind of Cloud Foundry for VMs or teams of application service as well. Uh, so within TAS 4.0 released that TAS portal.

So that's kind of an example of where we're taking some of that, you know, that work in the open state on a different, completely different open source project. That's very popular within the Kubernetes ecosystem, but then we made it, you know, Cloud Foundry compatible as well. And, and we're offering that, you know, that internal developer platform on top of Cloud Foundry to kind of help you align even more with some of these, you know, new definitions of platform engineering. So we're, we're seeing that, that kind of, you know, pull back within, you know, uh, or, you know, take, uh, good tech and integrate that back into, uh, Cloud Foundry as well. So it's exciting to see, uh, those new features and, and how customers can use that, uh, to help, you know, even, even further, um, enhance their current platforms today. So I think, I think maybe there's another one, but I think the next big event, at least you and I probably have is like a VMware Explorer. Yes, for sure. So, uh, you know, what, what's, what's going to be going on like Cloud Foundry wise there? Oh yeah. Well, I believe, I think the number is still correct. I have five talks at Las Vegas Explorer on different, uh, Tainz of application service and Cloud Foundry. You, you do, or there are? No, I, I personally, yeah. Wow. And then, uh, you know, we, uh, we, we did our best to try to get, I think, uh, as many slots and, and, and airtime as we can for, for TAS and Cloud Foundry. Um, because I think it's a great product and we, we need to talk about it more. Just, uh, just, you know, just because it's not Kubernetes doesn't mean it's not awesome and we don't have a bunch of, uh, awesome customers and use cases to talk about. So we, we've got a couple of panel sessions lined up, um, where some of our customers, uh, I've got a talk within the kind of SpringOne talk, uh, what SpringOne track, uh, where how, uh, you know, meet the expert sessions that you can come, if you have questions and we can, you know, you know, open, you know, open mic type style, you know, uh, ask any question you want. So there's going to be a lot of, uh, definitely a lot of Cloud Foundry and Tain's application service coverage in both, uh, Explorer of Las Vegas and Barcelona this year as well. And then the Barcelona call for papers is still open for, as we record this, June 21st, for

three more days. Yeah. So I, I, I am hoping we, we, we should get good covers there. Yeah. Yeah. But definitely, I think, I think especially for the, a lot of people who are here, right? Like if, if they get a chance before the 24th of June. Oh yeah. I think, I think there's, uh, like a lot of, I mean, I don't know, maybe even some of the talks that were given here would be a nice, uh, to have there, but definitely there's, there's always room for, uh, more, more Tanzu application service or PCF or TAS or Cloud Foundry talks. Yeah. Uh, to have at those events in, in lovely Barcelona in November. So it won't be crushingly hot. Yeah. It'll be mild. Mild. And slightly breezy. Yes. But still, still a short time. Well, uh, any, any other reflections from the day before we wrap up? Uh, no, I was just, it was a great day. Great to see, uh, everyone turn out. I got to a lot of meet, meet a lot of people. Uh, I haven't met yet before besides, you know, just Zoom, including, you know. Sure. Meeting you yourself in real life. So that's, that's, that's cool too. Um, but yeah, it's just, it's good. It's good to see, you know, the community chatting and I enjoyed to see, you know, uh, we had a, uh, uh, Tanzu customer event last night, kind of pre-reception to see all that, you know, our different customers kind of chatting with each other. So that's always good to see and having that opportunity to kind of connect directly one on one. And I'm sure we'll do the same with, uh, both, uh, both VMware Explorers to get, you know, kind of detailed one, you know, uh, TAS dedicated, uh, customer events as well. Sure. Yeah. Well, as always, this has been a episode of Tanzu Talk. No video for this episode, which, which is, which is nice. I can, you know, uh, I've, as, as Nick has had to suffer through, I've, uh, picked my nose and gotten, gotten meat out of my teeth several times. Uh, things like that kind of, uh, cleared out my ear. I could go on, but I won't. Uh, but if you want to get the, the show notes like that new stack article we mentioned, see how to subscribe, uh, things like that. If you go to TanzuTalk.com, uh, you can find all of that and we'll see everyone next time. Bye-bye. Bye. Bye.

Tanzu Talk · 🍕 Get more pizza.