Today, I thought it would be good, you know, there's not a whole lot of fascinating news going on, as we were talking about earlier. But I think what that means is, is we have we have the chance to talk about a couple of broad topics, weighing in with our various expertise in our our surveying the landscape constantly. The what's what's the name of the people in the military? It's not your it's the ordinance survey. I feel I feel like I feel like that was a big, unfortunately, you know, so many evolutions and civilization come from military needs. But I feel like the concept of an ordinance survey was a huge evolution and civilization where like, you know, what we need is accurate maps. That would be cool that don't have like dragons on them. And yeah, it's fascinating. You know, I worked for the ordinance survey in the UK for for a period. Wow. What are the chances? I know. I know. It is fascinating work as well. The way that they capture all the data and the way that they process it, we were building an event driven system with lots of rabbit MQ in there and microservices and stuff. It was pretty cool. It was good. So hello, ordinance survey. Are you still there? Did you get did you get to go to some like, like giant, like Victorian style rooms with like big drawers that you could pull maps out of and people be like, Oh, yeah, that's the map that that you know, Drake or someone used to fight the Spanish. It's right there. Please don't touch it. Yeah, I think I think maybe there was sort of a heritage room. But the ordinance survey in the UK is a fabulous brand new building down near Southampton. Very modern. Lots of glass, lots of concrete structure set in parkland. It's beautiful. So really, really nice. If you appreciate architecture, you'll appreciate that building. Does it get to get a little marker on the map? Yeah, it's not there. It's not on the map. Yeah, it's not on the map anymore, is it actually? My dad was into geography. He was like a town planner as a job. And all I remember growing up as a kid was we had a bookshelf in our corridor full of all the ordinance surveys maps, you know, nicely numbered left to right, etc.
2:12which are just shelf upon shelf of every map across the whole UK you can take out. And now, of course, you just pull out your phone and do Google Maps and it tells you turn left and that's what you'll see. that's fun. That's a better hobby than memorizing the train tables, probably. But that sounds like fun. I also spent time working for the naval mapping people who have a completely different building, which is near where I live in the southwest. So hello to the National Hydrographic if you're out there as well. Have you ever worked for Simon Wardley mapping? Because you could complete the whole thing. Career theme, you'd be done. Or space mapping, maybe with Space Force or something, you know. Yeah, the mapper. That'd be pretty cool. Yeah, well, it's good to know that they have a modern building. I got taken on a personal tour of Parliament once and it wasn't quite this bad. But every now and then it was basically like, oh, yes, the last time this room was painted was before the fire of 18 whatever. They still have chamber pots there, don't they? Maybe. No, I joke, but it was a great tour. It was like, I think the highlight of it is we walked by some, I guess, yard and like courtyard and my friend was like, oh, yeah, that's the original Scotland Yard right there. And now it just had like, you know, a little hut in it, but whatever. It's always fun to see stuff like that. Well, there's your tourism section of the episode. And speaking of space cartography, if that's even what you would call it, is cartography a two-dimensional concern or is that three-dimensional? Well, I remember when I was there that, you know, there is a three-dimensional aspect to it. Sure, sure. When they do mapping, map imaging, it's rarely, they do use satellite, but, you know, when you sort of zoom right down, that's mostly still aeroplanes, or it was at the time.
4:13Yeah, yeah, yeah. It goes from two different angles so that they can measure the heights of things. Yeah, now they have LIDAR and stuff, so you can have all sorts of fun. But, I mean, like, I feel like if you're mapping space, that is, like, obviously such a big, as far as we know, infinite thing, that there must be something beyond cartography that is the science of mapping infinite space. I don't know. Or maybe not. And every time you see, like, an astrophotography image or, you know, a JWST image or something, you see all those sort of stars and galaxies in there, there's no appreciation of three-dimensional depth. So, for a layperson, you've got no idea whether the thing you're looking at is really close or really far away or really big or really small. You've just got, it's just a flat image. Well, they're all far away, just relatively speaking. They're all far, yeah, that's true. But I find that fascinating, you know. So, it's like, I like this, I love this image, but I could really do with knowing the depths of stuff, you know. So, some sort of model would be fantastic. And they do do a lot of modeling, but not everything, obviously. Yeah. That's great. I'm going to have a lot of map. We'll have a special holiday episode this year called Map Talk with Ben. And we'll figure, we'll see what the relationship between event-driven architectures and maps are and how that works out. Whether you're on sea or land, which I think would be handy. Well, so, the first thing I thought would be fun to talk about. Now, there's been a lot of, I don't know. I mean, tell me if I'm right here, Ed, because you scan the airwaves and the internet waves, the intrawebs, as we used to call it a lot. But I feel like there's been a lot of discussion about not only cloud costs, how much does public cloud costs. No one really talks about private cloud costs, which is fine. Because nobody knows what it costs them. But then also, I think this has evolved maybe.
6:14I couldn't tell you what the chain of events that set this off was, but it's evolved into a cloud repatriation talk. Maybe I could just go search for the phrase cloud repatriation in Google Trends and see when that started to spike. But I feel like there's a big uptake in this idea of what we, the people, you know, the people using cloud are going to be interested in going forward for the next year. It's figuring out, not only controlling our costs, but if it makes sense to bring workloads back to, I don't know, back to where, but out of public cloud. I think probably what they mean is kind of like managed hosting somewhere. And, you know, and in fact, there is just a poll from one recent article. The most recent one was, it was kind of a roundup of a bunch of CIOs being interviewed and talking. And they were just saying how this is a priority for them. And then, you know, as with all these kind of predictions, who knows what will be happening in 2024. I feel like in the past four to five years, we've learned that predicting more than 12 months into the future is just a fool's errand, I think. But let's say nothing crazy happens like has never happened in the past three year periods going back since the dawn of time. By 2024, IDC says that, like, 70% of companies are going to have spent a lot of time worrying about costs. Now, they worded it a lot more baroquially than that, to make another architectural reference. But it's basically like that's going to be a big deal that people are doing. So what have you been seeing out there, Ed, about cloud cost, cloud repatriation? What's happening there? Yeah, much the same kind of stuff, right? Huge interest in FinOps. I mean, that's been bubbling along for quite a while. And I tend to see stuff and just think, yeah, more FinOps. It's a bit like more platform engineering with people talking about the same things. It doesn't mean anyone's kind of got it under control.
8:15But it's obviously just the pullback in the economy, isn't it? And suddenly everyone is seeing these bills, which have been going up for a while. And then they're suddenly like, OK, we need to cut them and then realizing whether they can and how to do that. And I think there's a bit of a knee jerk about the whole thing, right? Well, we could just put it back on prem when it didn't have these escalating costs because you didn't know what it cost. And you'd pay up front mostly. So until you could pack it full or you could just do, whereas cloud makes it more transparent. And then suddenly you're like, oh, that is a lot of cost, isn't it? And suddenly people worrying about it just because, you know, first strings are tighter than they used to be. You know, you mentioned a thing I'd forgotten of that comes up frequently. And that is if you've if you've signed these like two, three year agreements about basically you're signing up for a floor of capacity, the minimum amount of capacity that you're going to use that you pay for up front. I guess I mean, it's probably the case that a lot of those like two to three year contracts are like now midway through and companies are kind of be if they haven't moved, use that capacity. They're kind of like, oh, we better do something, which is, you know, a little bit of like good business case career management and a little bit of fallacy of sunk costs that you're kind of like looking at there. But whatever, you know, a very good, a very good example of that. I don't know if any of you saw this kind of what I think of the TwitPist right back in the day when you stuff, these little arguments on Twitter, the odd one caught everyone's attention. It was a great one that I saw earlier this week, which was exactly around this. So it was around Twitter as a company not paying AWS's bills as they sort of see, you know, they've decided to stop paying a load of bills for various offices and everything else. But one of these is they decided they weren't going to pay Amazon the 70 million they owed them or something. So Amazon in turn were then going to retaliate and not do any pay for any of the advertising that they were paying for. But the article had gone on to say that Twitter had a contract, had a three or four year contract, I think, with AWS, which they were not using, essentially, because they planned to move their kind of main timeline onto AWS and then decided not to.
10:20And actually committed a much bigger contract with Google Cloud and a plan to move everything to Google Cloud. And I think, you know, who knows, right? You could guess and speculate about why they're not paying their bills. But it was an interesting one of, again, right, as a big company, you've committed to one cloud platform, then committed to another. Sort of renegging on the first. You've got some capacity there. You've got everything paid for, but it's not as simple as, right, just stick it all over on that cloud, is it? And then, obviously, the whole just not paying your bills, watching your service collapse. But, yeah, definitely one of those FinOps comes to bite when times get tougher. And, obviously, Twitter, more than most, a year ago, we're in a very different place to sign those kind of contracts. Now, Ben, as someone who worked on applying event-driven architectures to naval wave patterns or something, I forgot the fancy word that you used there. Like, when you were doing that, and just in general, like, with the developer mindset, like, how often do you think about, like, the cost of the architectures you're choosing and, like, the stuff you're doing? Like, does that, do you think, do you think that comes up for developers? Well, I, you know, my experience in enterprises, it's rarely something that you would ask developers really to worry about. They've got enough to worry about already. So, worrying about cost is probably a stretch too far. But as an architect working in those places, you know, an enterprise architect, a cloud architect or whatever, then, yeah, very much so. You know, and cost can be a big driver for whole initiatives. You know, new projects that you want to deliver could all be about reducing cost. And that's always been the case. So, I don't think that's changed. It's just the cost center is changing. The cost center is now the cloud, and it used to be on-prem. So, you just got this sort of shift in appreciation for how much stuff costs, yeah.
12:22Yeah, yeah. You know, it seems, as I've observed all the FinOps cloud cost stuff from the side, right, it seems like, as when it comes to so much money-related stuff, a lot of it just has to do with, like, to use a fancy word, getting visibility. And just, like, understanding where the spend is. And, like, you know, I'm also thinking when I was a developer long, long ago before any of this, there was another way that you ensured developers didn't spend a lot of money, and that was just because it was physically impossible, too. It was just, like, you have this server. That's it. And so, there was no elasticity or burstability that you could have, whereas, you know, I would imagine, again, it's back to, like, the visibility and controls. Just as you were saying, Ben, being a different model of operating, that if you don't have those, you could say the word quota, but if you don't have those quotas or controls in place, it's just, like, it's easy to, like, go crazy with what you're doing. It's kind of, like, you know, one of my favorite transatlantic, would you call this, I think this is too silly to be a dichotomy, but, like, in America, you can have all the napkins that you want, and over here in continental Europe, you get one, if that, right? And so, like, so you can imagine that we, and this is the same generally in toilets or WCs or bathrooms, as we would all say. There's always a very, I feel like, I feel like there's a shortage of napkins over here in continental Europe. But that means that, like, of course, because you have unlimited supplies of napkins that are freely available, people just go hog wild with the napkins back in the States. But over here, very conservative, once you, once you control the napkin usage. What a unique observation. So, I think, you know, I, I feel like, you know, and going back to, you know, to close it out a little bit,
14:22like, well, there's two things I wanted to mention. One, you were making me think, Ed, that, like, and, and, and tell me if you experienced this, but I think maybe a lot of cloud cost stuff, it's easy to, like, get really, like, worked up about it, but it's just, like, a normal, somewhat boring thing. It's just, like, yeah, you should, this isn't a new thing that we've discovered. You've just got to, like, keep track of what you're spending. That's, that's probably what you should do. There's not a new phenomena, which. I'm sure that would be Corey's advice as well. Corey Quinn. Yeah. It's just, like, it's, it's kind of. But what I like is that it's also created a new market, hasn't it? There's a brand new market, an evolving market for cloud cost tools and products, you know, and they're becoming extremely popular because no one understands their cloud bill. It's just massive and sprawling and covers all sorts of stuff. I mean, I get, I work for VMware, obviously, and I get cloud bills because I use different cloud providers. So, um, I can see my cost and I can see the cost that I'm racking up. Some of it, I don't even remember doing it. It's like, what? I don't even remember doing that. I definitely don't need that. Why won't it delete? And there's always another bucket somewhere or a kind of, well, these costs are part of EC2, which bit of it, where is it? Yeah. Yeah. Backing down that little individual bit, even at a small scale. Exactly. So tooling like, um, VMware ARIA cost, for example, that's based on cloud health and stuff. They're very, very popular and it's a growing, um, you know, it's a growing market. So I like the fact that it's created a new market and that folks can start to get that visibility that they need on their cloud bills and have it wrangled into some sort of shape where you can actually understand it and start to control it. That's good. That innovation in that area is really helpful because it wasn't great from the cloud providers, I would have said. I don't think so.
16:22Not looking at my Amazon bills and stuff, I wouldn't have said it was that great. So I think it's good that there's been that innovation there. It's been very helpful. Yeah. And you're asking the cloud providers to step up now, aren't you? A little. I think, you know, they see that obviously people are tight in their belts and are a bit more concerned. I think it was Microsoft joined the FinOps Foundation this week. But even in their quarterly statements, they've all been saying they see belt tightening and they are sort of stepping up to help people with bills because they know if they don't, they'll just get hurt even more by people cutting back. And, you know, I'll add one, you know, I think a major motivator for paying close attention to this, right, is like if every now and then I try to track like how much, how many workloads are in cloud versus not cloud and what kind of like the rate of migrating stuff is. And I think there's a survey from O'Reilly from last year, their cloud adoption survey from 2021. They haven't published a 2022 one. So who knows how this has changed. But like there was basically like I think about half of the people who responded said they that in 2022, the year that just passed, they were planning on moving 50 percent of their workloads to public cloud. Which, first of all, unless they had already moved about 45 percent or 48 percent of their workloads, there's an impossible goal. So they're just setting themselves up for like, you know, picking up the iron triangle and jamming the pointy end straight into their forehead, like is what's going to happen there. But there's obviously and, you know, and I'm sure you all have these conversations too, right? But I have conversations with I think I've had let's see if this phrasing works. I've had more conversations in the past year than I think I ever have with just like mainstream organizations who are like, yep, this is the year we have to move everything or, you know, we're actually moving all the stuff instead of just like whereas the conversations I used to have is like we want to be ready to move things like we want that flexibility. But I think people are actually doing it now. And so when you have that huge like movement of things, now's a really good time to make
18:26sure that you you have your costs under control because I think maybe that's also what a little bit of what drives it as you were kind of saying, Ed, as people are finding this is I think there's been a lot of talk in the overall like mainstream of moving to public cloud, but not that many workloads have actually moved. So now it's time to figure that out. So we'll see. Hopefully when O'Reilly, we'll see if they do their 2022 and it'd be fun to revisit if people moved all their workloads or not. That would be that would be great. Now, the other thing I wanted to go over just I haven't, you know, true to form, I haven't actually gone to read the CD Foundation survey yet. I don't know if you two have, but I've read summaries of it. That's my favorite way to read things is to read things about things summaries, I think is handy. But, you know, another thing I'm always interested in tracking is how much build automation or continuous integration and continuous delivery people have in place. And, you know, I, I, I've been tracking, not for all 13 or 14 years, but every now and then I go in, and I find the used to be called like state of agile survey or something. And I think collab net made it now they've got some new name for their company, but they've been doing a survey about agile use for a long time. And so they had a question about using CICD. And basically it's been, it's kind of leveled off in the past three or 40 years at about 45% or so. And it's kind of like stayed there and you see this in other surveys. And then I came across recently another survey from the continuous delivery foundation that basically said it was, it was phrased optimistically, which I'm trying to be as, as I go through life, be more optimistic. Uh, so the optimistic phrasing was that 40% of developers use either CI or deployment, uh, and essentially, uh, and one in five uses, uh, both of them.
20:26Right. So I can't do the math in the air of what that, that means, but to put it the cynical way, 53% of developers don't use CICD, right. Which seems like a little bonkers that they're a, I guess there's an or in there. So maybe they're doing CI, but let's just say that anywhere between 40 and 60% of people writing software in the world are not automating their builds fully. Now, doesn't that seem a little bonkers? Yeah. I think, I think, I'm sorry, I should have probably prepped more on that, but I, I'd love to know more about that survey because yeah, my, my gut tells me it's entirely wrong. That, that, that idea that there's only 40% using either CI or CD, not in my experience. No, I can't think of an, a single environment that I ever walked into as a consultant that didn't have at least CI in place. Uh, so, so in my experience, it's like a hundred percent. They always have at least CI. CD is a different story. Having that good sort of path to production, uh, ready mapped out as a decent supply chain and having control over that supply chain. That's, I would say that was much rarer, um, you know, much, much less likelihood that you'd walk into, uh, you know, a bank or a, or a car company or whatever, and they would necessarily have all that stuff already in place. Um, but CI definitely, there's absolutely no one I can think of wasn't, wasn't building, uh, building software using CI. Unless the, the survey is particularly broad and covers maybe things that don't get CI'd a lot. Like you, if you went onto GitHub and looked at lots of software that's written in open source, for example, you know, little packages for this, that, and the other, maybe they're not CI'd the whole time. Maybe they're just built on, you know, built locally.
22:28Uh, and that's fine because the people who need it know how to build it locally. They do that and off they go. And you see that a lot in Linux too. Yeah. Yeah. You know, you, you raise a good point that like, I think the most interesting way to slice this would be by project size, right? Like, like, like, because, you know, a small thing maybe doesn't need that much CICD, let alone like, and, and you bring it, bring up another good point that if you're doing like a framework or a library, you don't really deploy that. Like, like, but also like generally, I, I, I would think that the larger your app, the more build automation you should have, and you know, at a, at a, at a pretty low level. And so like, it would be interesting to see for like large, significant applications, like those that were not automated, there's probably something wrong going there or trying to be optimistic. There's probably some room for improvement, uh, that, that you could bring in. I think the answer, I think that, I think that the breakdown is, is even more simple than that. It's, are you working in a team or not? If you're working in a team, you've probably got CI. If you, if you're working on something that's, you know, a passion project, an open source project, maybe something that there's a very small number of people working on, or that is not necessarily, um, something that you need to deliver into an enterprise environment, then, you know, maybe it doesn't matter so much, you know, that's fine. But, um, if you're working in a, in a team, in an enterprise, I would think it would be almost a hundred percent. I can't, I can't think of scenarios where I haven't seen that. Yeah. Yeah. Yeah. I'll have to go back and look at the, the state of agile survey, because I think, I think the definition they used, uh, is like that old, uh, the old, like Martin Fowler, Jez Humble, whatever that book is. Right. And there's like three or five criteria for each of the CIs and CDs that they go over.
24:28But it is always like with any kind of survey, especially surveys about a software development practice you use, you're, you're putting a lot of trust in the people responding on the survey to have the same definition as you, right? Like, like there, you know, just saying that you automate your build, you might mean something different than if you have continuous integration versus, or not like all sorts of things like that, but it is, uh, it's, it's a, it's a good thing to continuously look at. But what I do like is where, you know, CD is maturing a lot and, and there's a lot of talk now around secure software supply chains. And that's something that I'm particularly passionate about and that I, I see a lot of value coming from because it sort of, it really sort of formalizes that path to production more and adds more security into that path to production so that you've got more confidence about the code that you're running. You know, it's all about having confidence now that the code that you're running in production should be there and has a right to be there because it's met all the criteria you want it to meet. A lot of enterprises don't necessarily have that confidence about every workload and the problems that can arise from that can be significant. Yeah. Yeah. And maybe, you know, if you pushed CICD closer to what you're talking about there, then I could see that the numbers were, were a lot less, right. If you get these, uh, these higher level benefits and then, well, okay. So then finally, let, let me, let me, let me try one of my little jokey things. Cause you're bringing up something like I've been, I've been, uh, uh, watching a lot of that we we've, I've been consuming a lot of videos, uh, that, that our friends and coworkers have been making about supply chain stuff recently. And you brought up something that I haven't really like put into words too much. And that is, it seems like some stuff going on in the supply chain world is all about like kind of separating out the CI from the CD and basically saying like, we're going to
26:29establish a supply chain and like, whatever y'all developers do to get your little bit into this small part of the supply chain. Like we don't really care about how you do it. We just want to make sure you did the right thing. And so in effect, these things are very separate and here, here you go. Here's my little funny thing. Instead of it being a slash between CICD, we're going to convert it to a pipe. We're going to erect a wall there so that there's more separation between the, uh, the two things so that we, uh, we separate them. There's, there's a little something to think about through the week. Can we, uh, can we work on that? Cause I'm envisioning we're going to have a slide where the slash slowly animates to turn into a pipe and, and then, and then we'll just like do that, that clap out thing. And then the lights go out and the, uh, the house band comes back out to, uh, play the rest of their keynote at 9 30 AM in the morning. Well, I, I think that the two conversations that we've had today are related. So this, you know, supply chain stuff where we're worried about code confidence, it came up in the survey about cost that for the first time cost has actually become more important to some people than cyber security. Ah, so it's, it's like, okay, so it's cost must be pretty, it must, they must be serious about costs now because cyber security is always the thing. That's why the whole sort of supply chain movement is moving in that direction of adding extra security. It's a big, big problem. So to have costs come up and say, you know, actually we think it's slightly more important this year. That's an interesting fact. Yeah, I think you're right. That we'll have to, uh, we'll have to start following that more closely. And by we, I mean, Ed, I will find some facts and figures. They will be entirely unreliable, but I will find some, those are the best facts and figures. The, the reliable ones, they tend to be, uh, you know, 20 years after the, uh, the fact
28:31and totally uninteresting, just like, you know, yeah, yeah. All right. Well, I think that's a, that's a good check-in on, uh, things going on this week. As always, if you want to, uh, uh, there'll be some show notes linking to the stuff we've talked about. You can go to tonsutalk.com to check those out. And also the, uh, the SpringOne conference CFP is now open. I think it's going to close sometime in March, uh, hopefully for my sake, uh, later in March rather than early, so I can delay doing anything as long as possible. Uh, cause, cause as you two know, I'm always late, uh, to things. Yeah. And don't forget VMware Explorer is on at the same time. So their CFP is also open, I believe. So you could put in a paper for either or both or. Yeah, it'll, it'll be great. And then we have our ongoing, uh, golden path series, I think once or twice a week happening as well, that you can check out if you go to SpringOne.io, uh, and I think that's about all the, uh, the promotion we have, uh, at the end here. And with that, we'll see everyone next time. Bye. Bye. Bye.