Click a paragraph to jump to that point in the audio.
Gareth: [00:00:00] I can force myself to think about it from your side. And if you’re trying to win an argument on the internet, then that’s not super important.
Bridget: It’s time for Arrested DevOps, the podcast where we help you achieve understanding, develop good practices, and operate your team and organization for maximum DevOps awesomeness. I’m Bridget Kromhout, and show notes for today’s episode can be found at arresteddevops.com/devops-weekly. Before I intro today’s special guest, a word from our sponsors. ChefConf will be held May 23rd through 26th in Chicago. Chef has been a longtime supporter of the DevOps movement and of this podcast. ChefConf will have talks on infrastructure automation with Chef, compliance automation with Inspect, application automation with Habitat, and a ton of other relevant content. Register with discount code ADO2018 to to save 10%. Visit chefconf.com for all the details. And remember, code ADO2018 gets you 10% off the ticket price at chefconf.com.
[00:01:09] Your application sits on layers of dynamic infrastructure and supporting services. Datadog brings you visibility into every part of your infrastructure, plus APM for monitoring your application’s performance. Dashboarding, collaboration tools, and alerts let you develop your own workflow for observability and incident response. Datadog integrates seamlessly with all of your apps and systems, from Slack to Amazon Web Services, so you can get visibility in minutes. Go to arresteddevops.com/datadog to get started with Datadog and get a free t-shirt. With full observability, distributed tracing, and customizable visualizations, Datadog is loved and trusted by thousands of enterprises, including Salesforce, PagerDuty, and Zendesk. If you haven’t tried Datadog at your company or on your side project, go to arresteddevops.com/datadog to get a free t-shirt and support Arrested DevOps. Today’s guest has shaped your perceptions of DevOps for years, whether you know it or not. So I’m so happy to welcome to the show Gareth Rushgrove. Gareth, can you tell our listeners a little about yourself?
Gareth: [00:02:18] Yeah, sure. Sort of somewhat overwhelming introduction. I’m Gareth Rushgrove. I’m basically Gareth R. In most places on the internet. I’m currently a product manager at Docker, previously worked at Puppet, and before that, I was doing a bunch of sort of operations and development stuff at the UK government. I also curate the DevOps Weekly newsletter, which I think some people are keen on as well.
Bridget: Some people, yeah, let’s, let’s talk about that. So we’re using the newsletter as a framing device. So like, let’s start there. I’m curious if you can tell our listeners, how on earth did you start emailing loads of people every Sunday, and Why do you keep doing it? Like, for how many years is this?
Gareth: Yeah, so, and at this point it’s 7+ years, like 360 issues or some, some large number.
Bridget: And how many people are you emailing every week?
Gareth: About 25,000, 26,000 people, like subscribers. Um, it’s definitely one of those things, all side projects that got out of hand more than anything else.
Bridget: [00:03:23] Um, cautionary tale.
Gareth: And it goes back to, I guess, certain moment in time. I’m actually like, yeah, like 7, 8 years ago, like sort of Patrick ran DevOps Days in Belgium. I wasn’t there, but one of my, uh, sort of colleagues at the time, uh, Dean Wilson, um, uh, went along to it, came back and said like, you would have really liked that. That sort of got me like the word and a bunch of people. I was at the second DevOps Days Uh, in sort of Hamburg. Um, met a bunch of people through that. At the same time, I was doing a bunch of Ruby programming. Um, Peter Cooper had started out with the Ruby Weekly email, and that was novel. This was like, we were going back to like someone was sending an email newsletter and it was curated once a week. And this was quite nice. And I would, and just the confluence of those things of being interested I’d been doing bits of operations stuff, bits of automation, bits of software development, that was my background. But we had this banner of DevOps suddenly, and a group of people who were also super enthusiastic about something I was interested in. I was like, well, I’ll just start collecting a bunch of things I’m already reading and putting them together for other people to read.
Bridget: [00:04:42] That’s how it starts.
Gareth: To start with, if you go back, to like, to like the first issue, which I think was issue 0. Um, of course there’s like, at that point, like it wasn’t on Sundays, it wasn’t actually weekly because I did— I had no, like, I didn’t always have enough content. I didn’t have like a process of actually doing it. It’s like there’d be gaps. Um, but a bunch of people like subscribed And I just kept sending it. Um, and over time, I think in order to keep myself sort of able to keep doing it, I, I sort of said, well, actually, I need a day that I do it. And I tried a few different things and Sunday stuck. And it’s been Sunday for 6 to 8, 6-odd years. Um, and again, I like getting into a regular habit. Like made it easier for me to do, but also I think people stuck with it more because it was, it was literally then every week.
Bridget: [00:05:46] Yeah, yeah, people, people like consistency. Like Stratton is always telling me we need to put the podcast out on a more consistent schedule, and I’m like, yeah, it sounds disappointing and difficult.
Gareth: I think it’s definitely one of those things where like it helps me do it, so it helps me curate it in there. Like I, I, I spend my time like a fixed time on Sunday morning and it goes out. And a bunch of people, the sort of feedback I get often is like, oh yeah, I have it with my coffee. Or if they’re in like, they have it with their breakfast if they’re in like New York. Um, or they read it on, they read it first thing Monday morning if they’re in Australia. And like, I get all these like different variants of like, yeah, it fits into my schedule. So that’s how I consume it. And it’s interesting.
Bridget: Oh, that’s awesome. But I remember, um, at DevOps Days in Ghent in 2014 that you were talking about how you should should never put the name weekly in the title of anything. Do you still feel that way, or is that just a joke?
Gareth: It’s both. I think the moment you put the cadence, weekly, whatever, you’re stuck. You’ve set up an expectation that ultimately you can never really change. So actually, if I wanted to do the newsletter every fortnight, I would sort of need to change the name and that would like, actually then people would be like, what’s this DevOps Fortnightly thing? And it’d be like, no, not that.
Bridget: [00:07:11] And you’d have to explain to the Americans what a fortnight is and it would just be a whole thing.
Gareth: Well, and are you, are you DevOps Biweekly? And then you go like, is that like, that means different things in different places. Like including the cadence of something in the name is not a great idea. Having said that, actually it being there is sort of a social pressure for me to actually make it weekly. Um, and that consistency I think helps, like, has it helped it grow? I like, all of the growth is totally organic. It’s just, I, I mentioned it on the internet. Um, I’ve not done, like, I’ve not really done advertising. Um, uh, I get embarrassed at conferences sometimes by like John Allspaw where he’s like a big fanboy. Uh, or, so like you can see occasionally like someone will mention it at a conference and I’ll get a bunch of subscribers. Um, but otherwise it just sort of ticks up. But obviously, if that happens over 7 years and every week, then it ends up as a larger number.
Bridget: Yeah, absolutely. And you mentioned conferences, and of course, you go to a number of them. And I think that you have kind of an interesting perspective, because as you mentioned, when you joined here on the podcast, that you were at the UK Government Digital Service for a while, then you spent a few years at Puppet, now you’ve recently joined Docker. I think maybe our listeners would appreciate if you talk a little about how these gigs that you’ve been in, and maybe other ones too, shape your perspective in terms of how you see the community? Because you mentioned this is a curated newsletter. Obviously, it’s opinionated, it’s your opinions. And I’m curious how the environments you’ve been in have shaped those opinions.
Gareth: [00:08:46] Yeah, good question. I think, so, like, for good or ill, like, one of the things that I’ve ended up doing with, across a number of jobs, like, has been, they’ve been quite different. So, I’ve worked for a company where I was the first employee of, like, 2 founders. Um, I’ve worked in like sort of more like, uh, sort of marketing agency type, design agency type places. I’ve worked as an in-house developer at a big radio like sort of station. I’ve worked for the government, which was obviously A, weird, and B, huge. Um, I’ve more recently worked for like 2 more similar companies in terms of like a sort of like, uh, sort of growth stage, like, technology vendors. Um, but I think that, yeah, I’m just interested in loads of things, and I think that’s come out in, like, me having worked in different types of organizations.
Bridget: I—
Gareth: the downside of that is I, I, I’m not going to have as deep an understanding of someone who’s, like, always worked for technology vendors or always worked for, like, large technology companies or always worked in government, but I’ve sort of like hopefully more than dabbled, um, like passed through a bunch of these things. And I think that perspective is interesting.
Bridget: [00:10:02] Yeah, I mean, I think that that helps probably. It helps you see what, um, your readers from a lot of different backgrounds might find interesting. Yeah.
Gareth: And I said, I mean, I made a sort of an explicit like decision to go work like in technology companies, but that wasn’t sort of my formative experience. And actually a lot of my A lot of people I know, whether that’s through the .NET community or whether that’s through previous jobs or other communities I’ve been part of, have invariably been more on the actual user side, the consumer. I think that’s useful and interesting. I did that on purpose, but I think I didn’t realize how interesting I would find doing so.
Bridget: Yeah, absolutely. When you’re curating this news, whatever you’re finding notable, newsworthy to share, I’m kind of curious if we have to consider a little bit of an observer effect here, because you’re not just reporting on the news, you’re also sometimes part of it, or you’re reporting on things that your friends and colleagues are doing. Um, like, I’m curious, uh, how being a co-creator to what you’re reporting on influences the way you report, or vice versa.
Gareth: [00:11:17] Um, yeah, totally. I like— and I don’t claim not to be biased. I think it’s part of it. Like, like, you’re gonna link to your stuff, you’re like I’m the newsletter creator.
Bridget: I’m going to link to my stuff when I have a thing.
Gareth: But also I think that like the, some of the biases of like things I like and don’t like, having said that, I will happily include things that I disagree with, sometimes with a bit of editorial, sometimes not. Um, like if they’re interesting, I think that that’s the, my, my sort of, and I read a lot more stuff than goes into the newsletter. It is like, I do curate it to that, to the extent that like a lot of things that I read or get sent don’t go in. Um, my barometer is basically, is it interesting? And there is a— there’s, there’s a definitely observing thing where like, is it interesting? Like, actually, is it interesting to me? Um, having said that, that’s sort of what like the, the feedback from people actually subscribing, sticking with it, like giving me positive feedback about it, is sort of like what other people find those things interesting as well. I’m not claiming to like cover everything. Or include everything. Like, turns out the internet is just too big for a— what, what is a side project of, of like, for one person? Um, but there’s a bunch of other things I do as well to sort of avoid, I guess, more agrarian, agrarian, like, problems. So I don’t really cover news exactly. I, I like, oh look, there’s a press release for new product has been launched. Probably don’t care.
Bridget: [00:12:51] Or these people acquired these people.
Gareth: Oh yeah. Or this person’s joined this company. Or like the sort of the industry news side of things is not what the newsletter does at all. If you actually look, and only a few people have ever actually like told me they’ve spotted this, it’s like the editorial bits never mention people.
Bridget: I noticed that like when you’re linking to an article, you don’t say so-and-so of such and such place. Which I think is kind of nice. It gives it like, it stands on its own.
Gareth: Yeah, it’s about is the co— is the content interesting irrespective of where, like, who it— like, it’s not to say that that might not be like, something might not be interesting because it’s from Google. On the other hand, something can be interesting without being from Google or Netflix. So absolutely occasionally companies get, um, mentioned because that’s why that thing is interesting, but actually the majority of the time it doesn’t mention like the actual like sort of why you should read this blurb, um, that I write doesn’t mention people or companies. And that’s not to downplay the importance of people or companies.
Bridget: [00:13:53] Sure, sure.
Gareth: Like, the thing is the interesting part, not the personality.
Bridget: Absolutely. Um, and that’s— and that’s— I think so. I was wondering if that was deliberate, and I’m— I— it’s interesting to hear that, yeah, you’re deliberately highlighting the ideas. Yeah.
Gareth: And I think that that also makes it easier like from, like, from where I sit in terms of, well, yeah, there are some things I’m more familiar with, there are some things that I’m closer to, um, like to include like things from a broader set of like outside my immediate, like, I’m doing this thing at the moment. Having said that, I bet if you, if you sort of mine all of the content from all the news outlets, you can sort of find, you can sort of discover what I was like doing at the time because there’ll be a slight uptick in like those things being covered.
Bridget: More container content or whatever.
Gareth: Yeah. Well, I think over time, it’s like, I think when the newsletter won’t work is basically when I’m not continuing to sort of learn about various different values of Next. So, like, you see sort of the container stuff crop up really early. You can see sort of the emergence of serverless. You can see the emergence of Kubernetes, like, just in terms of volume of content. And yeah, there’s totally a bias there where they might be things I was interested in. Um, so I’m like, I’ve never really been a database person, or I’ve never really been a network person. Like, having said that, I know people who are and try it. They like tend to sort of feed me interesting content, and I can appreciate reading those things without actually being— having the same network of people to gather it. Yeah.
Bridget: [00:15:36] And so I think you have a a broad enough range of interests. Speaking editorializing myself, I would say, I think you have a broad enough range of interests that you create and curate interesting content just by saying, hey, look at this interesting stuff. As it turns out, that interesting stuff as a service is something people want.
Gareth: Like, DevOps as a banner has been purposefully broad from the start. And I think that, like, for some people, they want that, like, concrete definition that will never change, versus actually the conversations we were having, like, at like the first few DevOps Days, like whether it was the like Gent one or Hamburg or like San Francisco or like the super early things versus the conversations that those people are having now or the, uh, or are happening at those events now, things have moved on. And that’s good, that’s a good thing. Like not only is the community learning, like other things are happening at the same time. And I think I like— so the newsletter now is totally different Partly because I am, partly because the conversation is.
Bridget: [00:16:40] Yeah. And that kind of brings us to, like, I’m thinking you speak at a lot of tech conferences, and that kind of gives— and not just talking about your talk, but also talking about the conversations you’re having there and the articles that come out of and the talks that come out of conferences. Obviously, that gives you a lot of material for the newsletter. I’m just kind of curious, like, is there a cause and effect, chicken and egg here? Like, how did you start speaking at conferences? And like, how do you decide what to talk about? And how do does that tie in with the newsletter?
Gareth: Um, good question. I guess so. I was, I, I was doing bits of conference stuff before I was doing the newsletter. Um, not a, not a huge amount, but I, I think I sort of fell into doing like the odd sort of like meetup. Um, and actually it’s nearly like sort of like first or second job type territory. Um, when I lived up in Newcastle, um, Partly I knew a few people out, like, sort of in Brighton. I remember, uh, sending an email to Andy Budd, who at the time ran the, uh, Brighton New Media mailing list. Uh, he now runs, like, Clear Left, basically a design agency in the UK that’s sort of, like, pretty well known. Uh, and I said, hey, this thing’s really good, but I don’t live in Brighton. But I know some people who do, but like, how did you get it going in Brighton? And he said, oh, we just sort of did it. And I was like, ah, well, I live in Newcastle and set up something similar. Um, I wrote my own mailing list program, which was the— like, this was back when I was more stupid than I am now. Uh, and ran a mailing, like, ran the local mailing list for a little while. Like, um, I used to organize gigs, like, sort of before uni, at uni. Um, and getting more into tech, it was sort of, it was an easy step for me to go like, oh, well, I can organize like sort of local meetups or go along to like smaller meetups and things. Um, and if you’re doing that and you’re trying to bootstrap that, like being able to talk turns out to be useful because you always need a speaker.
Bridget: [00:18:46] You don’t always have a speaker.
Gareth: Exactly. Um, and so I think some of it was just sort of falling into it. Some of it was then Uh, I, I think I appreciated the sort of the, the ability to travel, um, like in a way that I couldn’t do otherwise, um, either in terms of like time or affording money. Like that, that was not within my reach or budget to start with, um, and that like flexibility came from actually, well, actually, if I could, if I can do a talk, someone like, that’s easier to do. And it’s, it just sort of spiraled really, I think. Um, I found I’ve met lots of interesting people that gives me more things to think about. Um, yeah, when it comes to the newsletter, like, there’s definitely a, a theme of content. I like the DevOps community as a whole has been amazing at sharing. I mean, like, the, if you look through the, like, the, the sort of corpus of content that’s even just around DevOps Days over the past, like, 7 or 8 years. It’s amazing. It’s an incredible resource. And actually having that one banner has put all of that in one place. Like, the fact that actually the model of DevOps Days has meant that most of those things can be recorded, the, like, the videos and the sharing and the slides, is sort of pretty— like, it’s easy to find something interesting there. And that, I think, has also spread into other events near, like, next to those, because then If it’s set up as a model, like, of course you publish your content, then that’s—
Bridget: [00:20:19] it kind of influences conferences to keep doing that and not putting it behind a paywall or anything.
Gareth: Yeah. And I think then, like, more recently, and certainly, like, well, what I’m doing now to a degree, obviously new, but like the— and at like from there, and you’ll know, sort of moving to a software vendor, there’s, there’s a, like, having A, marketing content that actually reinforces your product, reinforces your brand is great. Having that come from people who are actually practitioners, who are actually already trusted by that community.
Bridget: And end users who don’t work for the vendor.
Gareth: Yeah, it’s so much more powerful than the pure marketing sales thing for practitioner communities. And before that, the government was— The work we were doing there, a lot of it was about making government more open, making it more accessible, actually pointing out we had really interesting technology challenges and wouldn’t you want to come and work for your government? Us being super visible was strategy, not tactics.
Bridget: [00:21:27] Being super visible is so valuable for people who are trying to hire.
Gareth: And so I’ve never done it as like, quote unquote, part of my job, um, but I found it useful for whatever other things I’m doing, often actually as a design tool. I often like use, uh, the ability to talk to an audience and get feedback from that audience to talk about something that I’m thinking about doing. And there’s the, there’s the validation of that from this, like, from being accepted, i.e., like, you submit a talk and it gets accepted, therefore there’s validation that the thing you want to talk about is interesting. There’s also the validation of, like, and input into, like, a design process from an audience.
Bridget: So I think I’ve also seen, and I think it might have been you actually, I definitely have seen people who have worked at, like, Chef and Puppet and whatnot sit down at a DevOps Days, propose the topic configuration management And then sit down and they don’t talk. They just let everyone else talk and they’re like, free market research, this is fantastic.
Gareth: [00:22:32] Yeah, I’ve never actually done that, but like, I tend to be a bit more pointed.
Bridget: But yeah, but it’s kind of a take the pulse of what the community wants to talk about.
Gareth: Yep.
Bridget: Oh, okay. So by the way, quick sidebar on the topic of DevOps, you do not camel case it for DevOps Weekly, at least. And I, for DevOps Days, I don’t capitalize any of that and I put it all in one string. And I feel like this is kind of one of those touchpoints, you know, flashpoints in our community. Like, what’s your opinion about the camel casing?
Gareth: I think in the same way as— so DevOps as a thing is marketing in the purest, best sense. It’s basically an invented word that brought together a lot of people and still does too. Patrick didn’t camelCase it, therefore no one else should do. However, people did. I’ve lost that discussion based on a number of times, just like on, this is what people are doing on the internet elsewhere. I’m like, but everyone can be wrong and that’s fine. I don’t let it bug me because that would drive me crazy. The other areas you’d have to go down the same rabbit hole for is quite large. Um, but on the other hand, I’m, I’m stubborn and I, I probably won’t change it. Um, there’s, there’s a pull request, uh, on a government publication somewhere where, uh, I’d, I’d written the original draft. It ended up as, uh, uh, lowercase. So I was like, ha, yes, the, the, the UK government’s usage of DevOps is correct. Um, after I left, someone, uh, changed, changed that and I, I had to go back and sort of, but Their argument for doing so was that DevOps is a pop— is a pop culture, and it’s not. It’s a single made-up word. Um, and but yes, I, I’m, I’m, yeah, I’m talking far too much about a topic I, I’m not bothered about.
Bridget: [00:24:37] I just, it’s, it’s just, it entertains me because it’s these little rabbit holes we can go down, you know. It’s just funny. Um, but speaking of like debates we can have, I went to go check for purposes of this podcast, I went to go look up when that recent adventure we had, um, in Amsterdam was. And I was a little horrified to see it was fall 2015, um, because that seemed— it doesn’t seem like it was that long ago. It was, you know, a little bit ago. But when we did Velocity Debate Club, remember when, uh, Courtney Nash had us, uh, I think they had like, I don’t know, a slot to fill or something like last minute at Velocity in Amsterdam. And, uh, we’ll link to this in the show notes because it was super funny. We We were discussing whether containers and microservices were good or bad for your business. And the entertaining part was, we started out with me defending them and you dismissing them, and then we switched sides halfway through.
Gareth: I’ve forgotten the name of it. It’s a sort of style of debating. And I think it reinforces the whole, like, taking different sides, I think, teaches you like a lot about a topic. You’re like, and yeah, people are going to have like strong feelings sometimes. That’s fine. If you’re— if you practice on the— your ability to take both sides of an argument, you’re likely to then have more empathy for someone when you actually genuinely do sit on one side. And when you have strong, like really strong views and would find it harder to argue like the other side, like you can sort of think, well, yeah, but like actually I can start thinking about it. I can force myself to think about it from your side. And if you’re trying to win an argument on the internet, then that’s not super important. If you’re actually trying to like come to a consensus or learn something, then like I think it’s a really useful skill. And yeah, who knew that it would be quite as much fun? I think even the art, even like known left, the people seem to rate it quite highly. Um, I think It wasn’t just good fun for us.
Bridget: [00:26:43] But the reason I brought that up is because I thought it was kind of interesting to ponder. Now it’s 2 years later, I’m wondering, and I didn’t go back and I’m not even sure if there is a video of it. There probably is. I didn’t go and try to watch it, but I’m just curious how your thinking has evolved because, I mean, fill us in. You just joined Docker. Where do you think the containers discussion is now? Like, paint a picture for us.
Gareth: So I think, I guess, answering a few of those questions, containers now are in that sort of messy period where it’s not about containers. And I think people like a word and attaching loads of things to a word, that’s fine. But actually, where Docker the company sits today is actually like, Yes, there’s this technology. Yes, there’s this sort of buzz. These, like, people have found loads of different ways of making this sort of like fine, like doing things with the technology that like Docker sort of shipped. Um, Docker the company is then trying to make a business, like build an actual like business around helping like ultimately large, large organizations like modernize their, their technology, their infrastructure, their applications. And there’s a load of gnarly problems there that are not about containers. They’re not about— I think sometimes it’s useful to talk about the implementation detail stuff. And there’s a load of really smart people working on a whole bunch of whether it’s kernel-level things, whether it’s the next layer up, a load of the the fundamental building blocks of like how we— like what technology is going to look like next. Um, I’m not actually one of those people. I know like my— like just in terms of my background, my interests, I’m like, they’re not the things that I get really excited about. Um, I’ll happily geek out about some of it, but I think how they come together to form like higher level, like more user-facing, more business-facing things Um, it’s sort of interesting. And that’s, and that’s why I joined Docker specifically as a product, as a product manager as well. I actually, I’d said Docker’s got like all these like bits of technology that lots of people are using and actually a like product set in the sort of Docker Enterprise Edition stuff that like large companies are adopting. And how can all of that work better together? How can it go like faster for everyone? Um, And I think similar to sort of— I’ve had a lot of experience in sort of the open source community and then worked for 2 sort of companies with deep open source roots. And how you get— how we scale open source with new technology companies is one of those sort of really interesting problems that I’m personally interested in. I think it’s important for the sustainability of open source in general. Um, and there’s sort of a circle to be squared there.
Bridget: [00:29:52] Yeah, and that’s, that’s really interesting to me too, because obviously you’ve worked now at a couple of vendors that do really foundational, really fundamental, really important open source, and yet there’s also a business model there. And if you’re working on product stuff, um, or if you’re working on the customer, you know, enterprise-facing product stuff in particular, Like, how do you decide which stuff is, you know, free open source? Um, you know, what’s, what’s Moby and what’s Docker? Like, how do you make that? How do you draw that line? How do you decide from a product point of view?
Gareth: I’m, yeah, I’m obviously new, new at Docker, so I can’t really speak about like sort of previous things there, but I think a big part of it is like the open source is strategic, not tactical. So what’s the most benefit to the people who will actually give you some benefit back? And sometimes the most benefit might be actual money, actual revenue, actual direct revenue. And open sourcing the things that are going to give you direct revenue might be a bad idea. Having said that, the biggest benefit to something being open source might actually be a huge bit, like, big user community.
Bridget: [00:31:12] I mean, Docker’s buzz in the ecosystem is, is undeniable. It’s a gravitational force.
Gareth: And I think that, like, those are the arguments for and against are often contextual. Like, if, if you’re a very large company versus a sort of a startup that’s actually trying to, like, make a business, then the arguments are different. Um, I think, uh, Luke, Luke Kinnear, um, sort of the founder, um, like of Puppet and sort of good friend, he’s wrote some really good posts more recently. Um, I’ll try to dig out the links and we can add them into the notes because like, and obviously he’s got a huge amount of experience there, huge amount of hindsight as well. Um, and made a lot of decisions throughout Puppet’s life of deciding, well, Puppet was open source. Puppet was licensed under one model. Puppet changed its license. Some parts were open sourced later. The thought processes, I think, as an industry, we’re still learning what that looks like. We’ve had, obviously, open source companies for a while. But up until very recently, they weren’t public. The previous generation of those companies were all— MySQL AB was huge, but obviously it was acquired. Everything sort of— a lot of it— that might have been the best thing for those companies, but as a community, we lost the data. Obviously, public companies have to provide information about how they’re actually working. And we’re starting to see open source software companies go public. So Hortonworks, Cloudera, Mongo, there’s a couple of others now. There’s a couple of others maybe on the cusp. We’ll suddenly start having data in a few years that we never had before about the open source model for companies.
Bridget: [00:33:14] And that’s really interesting too, because you just mentioned a couple of companies that definitely, if we think about like a Gartner hype cycle, you know, big data, however we want to put it, is something that we were pretty hyped up about a few years ago. And I don’t know if we got to disillusionment or not, but we definitely are like, yeah, okay, data, that’s definitely important. It’s part of the overall, oh, now we’re going to be excited about IoT or whatever. Um, I’m kind of wondering, like, yeah, there you go, compute on the edge, you know, fog, whatever. Do I want my computing in the fog? I don’t even know. But that kind of— it makes me wonder, like, is there a relationship to be teased out there between, um, where on the hype curve something is and when, as an open source business model, it actually makes sense? I don’t have an answer for that. I’m not sure if that’s even a question, but I’m curious.
Gareth: I think there’s lots of people who have pointed out open source is where innovation is happening. And then obviously, open source companies invariably try and capitalize on that. But also at the same, I think that lifts open source, but it also results in products with features that large organizations need. And a lot of that, The sort of the current state of large-scale IT is there’s so many things that can be improved by things that we sort of broadly speaking know the answer to. But actually implementing that at scale in a highly regulated environment, in organizations— I mean, coming back to the sort of DevOps topic where organizationally they’re not set up to take advantage of how people have been using technologies. So, like, so much of the sort of microservice conversation isn’t about technology, it’s about the people and teams and independence. A lot of people will say, actually, from a technology standpoint, it’s sort of terrible. Like, you’re introducing latency all over the place. You’re now managing so many additional components. You’ve added all this operational complexity.
Bridget: [00:35:20] But complexity is fun, Gareth. Come on.
Gareth: But there’s also— but the flip side of that is, like, you’ve reduced the complexity from the perspective of an individual team that can now work far more independently. You’ve moved complexity around.
Bridget: Tim Noakes likes to talk about conservation of complexity.
Gareth: When you can make those trade-offs, there’s actually, yeah, I can pay the higher operational cost for a higher developer sort of team autonomy throughput thing. You can reason about technologies, architectures, platforms. And a lot of those things assume that, like, a lot of those things assume that, like, actually changing your team structures is trivial, right? And if it’s like a 20-person company ultimately, and you’re growing, well, you’re sort of changing your team structure all the time because you’re growing. What worked at 15 people doesn’t work at 20. What works at 20 doesn’t work at 25. If you’re a huge, big, large organization and you’ve been there for 50 years or several hundred years, like the UK government, that change of your organization in order to take advantage of technologies in the same way as they’re used elsewhere in much more fluid places is near on impossible. So the question is—
Bridget: [00:36:39] You can’t just reorg the government.
Gareth: Yeah. So can you get some of those benefits in areas where That change is harder. And I think that often when you look at across different companies that are based around open source software, but actually are fundamentally software businesses selling to that audience, that’s what they’re trying to do. They’re not trying to say everyone needs to pay for this innovation. They’re saying, we can get this into more people’s hands that can’t adopt it. And I think that is part of that hype cycle. It’s part of actually The early adopters don’t look like the majority. And going back to the Velocity talk I did a few years ago, the idea that everyone is a software company is interesting, because it’s easy to say and everyone agrees with it. But then if you say, well, what does a software company look like? And you look at Google and you say, well, give or take, they’ve got like 40% of their headcount, 50% of their headcount is in R&D. It’s basically engineers, designers, product project people. If you extrapolate that to every company is a software company and you go, so every company then has, let’s go for the lowball of 40% of people, that’s obviously insane. It’s crazy. It doesn’t work. Ultimately, there aren’t enough software engineers in the, in the US to staff like the FTSE 200, FTSE 100. Like, oh, we need other like models that scale better for companies that are not going to be first. They’re not going to be the early adopters.
Bridget: [00:38:23] Well, and sometimes people like to say fill in the blank is software, and you can make that argument, but it’s also people, and it’s also deliverables, and it’s also end users. Like If you do, if what you produce is cars or insurance or government, software is a huge part of that. And software is also not your entire raison d’être.
Gareth: Yep. And you’ve also got this spread of like, yes, on one hand, there are conversations going on about like, sort of the future is serverless, or the future is service meshes, or the future is service meshes in serverless environments.
Bridget: Are we over unikernels, by the way? I feel like the unikernel hype came, and I haven’t heard as much about it lately. Can we have serverless unikernels? I don’t know. Do unikernels have to be an evolutionary dead end?
Gareth: I don’t think they have to be, but I think we’ll all always have these out-there conversations. On the other hand, there are larger organizations spending an awful lot more money than people are spending on any of those things on getting onto virtualization.
Bridget: [00:39:31] I know.
Gareth: Not only are they spending a lot of money, they’re going to save loads of money in doing so. They like, it’s going to free up people, it’s going to save money, it’s going to make them a lot more flexible than they were before.
Bridget: Yeah.
Gareth: And like, just those two, those two things could totally be happening at the same time in like the, the industry.
Bridget: I’m thinking of like a giant airline, uh, having an outage because of a data center in the southern US. And it’s like, hmm, okay, if that airline can just lift and shift some stuff into the public cloud, they spread out that risk. They’re much less likely to have that outage. They’re not going to incur so many losses because of something like that. And that’s kind of orthogonal to whether or not they are running all of that with functions as a service, or if they maybe lift and shift whatever their web sphere and sadness is into some containers and stick it in the cloud, and then keep iterating in their new product lines.
Gareth: Yeah, there’s like the— I think the messy problems of how you change like organizations and software at like in environments like that are, I mean, to me at least, a lot more interesting than, than just the pure technology conversations. I like, I, I like hanging around with loads of different people and like some of the people I really like respect are just like, they’re They’re just into the technology for nearly for its own sake, not like, not quite, but nearly. And there can be a lot of fun. And the flip side of that is like, I also have friends who are like, I just do the people stuff. I just like, I’m actually mainly working on like sort of classic service management, like, and then like, these are the improvements we’re making. And I like, I like all of those things, like maybe not all at the same time. And I think that’s the, that’s how I’ve ended up sort of being interested in the whole DevOps community, the banner, the newsletter, the rest of it is because they were the people who came together. They were all often generalists. They were people who often are doing different things at different times, or they’ve moved from one place to another and taken more responsibility, or they’ve gone from hip places to large, messy enterprise organizations, or vice versa. I think that That’s interesting to me. It’s like those people are trying to go like, oh yeah, one moment you’re working at Etsy and the next you’re working at a sort of large, big bank. And yeah, I think from those people come interesting insights about like, that thing we were doing here is totally relevant here, and basically you can just do it in both. Or actually, no, we tried the same thing here and it didn’t work. And those stories are really interesting.
Bridget: [00:42:25] Yeah, and you keep bringing us those stories. And so that— I think that more than anything else, the fact that the newsletter does a really good job, or you do a really good job in the newsletter, of finding the interesting and cutting-edge technological advances and the practitioner stories and the analysis of how these changes in the industry are affecting people and people are affecting them. Like that, I think that’s why I keep reading the newsletter. It’s like, it has a really good, you have a really good, um, broad with depth, but broad perspective there.
Gareth: And the reality is people are writing really good there. And like, as a, as a broad community, like people are writing really good quality posts. I mean, like there’s this, I don’t think I was saying to someone like, I, I basically collect links during the week. People send me things. I see them on Twitter. I see them in other communities I’m part of and I Instapaper things. And then once a week, go through and I always get there and go like, oh, I didn’t really collect many things this week. Then I go, ah, wait a minute, there’s loads. Some weeks are definitely busier than others. Some weeks are more high quality than others. But I’m always super impressed just how much good stuff people are writing. One of the— John Willis says DevOps is sharing, is a fundamental precept of it. People have been like, are just constantly sharing, even whether or not they consider themselves part of that DevOps community or not now as well.
Bridget: [00:43:59] Yeah, I like it. Oh, and I have to ask, um, for our listeners, because I wrote an ACMQ article and you linked to it in the newsletter, and I think it was like last week or something. Did you link to it because you were coming on my podcast?
Gareth: No, luckily. Again, I, I, I found it in Like the topic and the article were really interesting. I think things like that in particular. Last week, actually, I was really strong on people and culture side. And the sort of— there was the post about experience actually trying to transform an IT team. There was one about swarming. I like the technology stuff, but it’s really easy to— it’s easier to find the technology stuff than the good people stuff.
Bridget: Yeah. Yeah. No, that’s, that’s cool. Okay. So we’re almost out of time, but I’m, I’m thinking like final question, 10,000-foot view. Um, can you give us a picture of the, what you see happening since you have, you know, your finger on the pulse, et cetera, et cetera, broader industry trends? Where are things, where, where’s the puck going? You know, what should people be paying attention to and learning from? Like, What’s your broad picture for early 2018? What should people be looking at?
Gareth: [00:45:15] I’ll be boring and say context. I think that you can chase technologies, but if you do so with a view to implementing them in your environment, you’ll always be— you’ll never stop. It won’t be— if you’re chasing the technology, whether that’s Like whatever that is, that’s Kubernetes, that’s Docker, that’s serverless, that’s what— like you’ll always be chasing the next technology and you’ll miss the fact that you’re like, well, what’s the business problem you’re solving? What’s the context you’re in? Like, is that the— is that relevant to where I am right now? Like, and I think stepping back and like, like I would like to say that the technology industry will reach a certain point and like What we’ll do is maybe step back a little bit and have more grown-up conversations about context and solving problems. I’m not sure that’s actually what’ll happen.
Bridget: People like their shiny toys.
Gareth: [00:46:16] Yeah. I’m not against shiny toys. I’ve had enough dalliances there. Pushing the technology envelope forward is really important. But I do think some of the conversations about where they’re useful are better driven by a passing interest at least, and some information about the context. What organization are you in? What problem are you solving?
Bridget: I love that. What problem are you trying to solve?
Gareth: I think it’s not— and where? Are you having the same problem in 2 different places? That might— they might want the same answer, they might want different answers. And without that context, you’ll— again, you’re just chasing the technology, and you’ll always be unhappy then, because I think you’ll— then you’ll be chasing the next technology afterwards.
Bridget: Yeah, that is— that’s a really good insight.
Gareth: Getting— as an industry, it was getting better at sort of technology that stops, um, in terms of like, it’s, it’s good enough, it’s stable. Because we know that happens. There’s quite a lot of technology that has sat for a long time that basically does its job. Large organizations in particular will have a lot more bits of technology running than they do developers working on them. I think we ignore that discipline a bit. Like in terms of sort of the modern tooling, modern technology, um, like how, like what would, like what sort of, maybe what tools and technologies or processes or what, like would we need to build, to build technology that lasts like 10 years, frankly? Um, like it’s sort of like some environments are better there than others, something like some— yeah, I’m kind of—
Bridget: [00:48:17] I’m thinking of like the space program, you know, they build stuff and they send it off to space and years later it’s still reporting back and you’re like, wow, that still works.
Gareth: And the fact that we’re surprised and like amazed by that is sort of depressing. Um, like, it’s, it’s, it’s possible. Um, and I think a lot of the like sort of arguments for not doing that elsewhere is, is really like how that’s done is really costly.
Bridget: Yeah.
Gareth: Actually, it’s really costly because it’s niche. Um, and are there tools and technologies that can be applied there that could mean that we could all do that? And you see, like, and like, when you look at the sort of stability of like the JVM sort of style applications versus, um, like Ruby on Rails, like, which I was like, I’m a huge, big fan of like Ruby on Rails. Like, I just love how to Big impact technologically. I did a whole bunch of work there. But coming back to a Rails application after 2 years that’s not been worked on by anyone else is actually— there’s a burden there. There’s a trade-off that’s been made to make that able to go. And it’s like studying those differences, like what led to that? It wasn’t by design. It wasn’t just a trade-off. It was like, what properties of software would lead to things that we could just build and would keep running?
Bridget: [00:49:36] So, I think what I’m hearing is a conference talk you’re going to be giving in 2018, probably at, like, Monktoberfest or something, and it’s going to be entitled Built to Last, and it’ll be about building software for the long haul.
Gareth: I’d love to be expert enough to give that. I think, like, it’s—
Bridget: But this doesn’t need— I mean, I feel like this isn’t necessarily expert in that you did it, it’s more a survey. Of the practices that can be observed, you know, kind of a field study, if you will.
Gareth: Yep. No, I certainly think it’s an interesting area.
Bridget: Like, nice.
Gareth: Whether I have time to actually do more than ramble on a podcast, we’ll see.
Bridget: Well, it— and this is what always happens is we run out of time and we say that it’s going to have to be another podcast because I think that would be a very worthwhile other podcast. So to wrap up, let’s say community and event stuff. I’m in the next few weeks, I have 2 weeks at home now, then I’m going to be in Bergen, Norway for BoosterConf, and then I’m actually going to be in London just for a couple of days for a party community event thing at the new Microsoft Reactor space there. That’s March 15th, and then I’ll be at SRECon in SF running a Kubernetes workshop on March 20th. 26th, and that just takes us to the end of March. Uh, hashtag vendor life, like travel season is starting up again. Oh, but what is yours? What does your, uh, next couple months look like, Gareth?
Gareth: [00:51:10] No, I’m not too much travel, partly down to the new job, apart from sort of internal things. I’m, I’m, I work, well, often with the Cambridge team, but we’ve also got an office in Paris. I’m gonna get a trip to Paris, see some folks there. Um, uh, and then back to London Uh, yeah, start of March, sort of 7th for QCon. So I’m, uh, I’m helping run, uh, with Anne Curry, an ethics track at QCon, which has been— it was one of those things where we were like, this will be good, and everyone was like, are you sure? And now everyone’s super excited about it. So we’re looking forward to that.
Bridget: We’ll put a link in the show notes to, uh, QCon London and your track with Anne Curry because that will be something people will want to check out. Right, awesome.
Gareth: We didn’t even talk about that topic, but yeah, certainly ethics and technology is again separate to the sort of— well, I said I’m sort of maybe getting more into a lot of the separate to pure technology things. I think ethics for technologists is going to become a more important pressing issue if as an industry we’re going to own our destiny. Yeah, yeah, we’re, we’re surprisingly powerful, and the companies that we’re like sort of part of are increasingly powerful. And I think like comes— that comes with a whole bunch of responsibility and eventually regulation if we don’t actually get our act together. Um, so I was having a— how people are— the robot’s going to be like everything from that. Yeah, and I think there’s But us having a sort of like a lack of an ethical education for technologists, a lack of an ethical sort of conversation anywhere, I think means that we’re totally unprepared for those conversations.
Bridget: [00:52:57] Yeah, no kidding. I mean, I, I took the appetizer sampler of liberal arts classes just because I wanted to, but my— at least when I did my computer science program, it did not have a lot of requirements around, you know, learning to be a well-read or you know, reflective human being. Like, that just— there was barely time for that in the curriculum. Yeah.
Gareth: And there were like lots of things. I think it’s like that. I think it’s le— it’s probably less about the sort of the pure education side. And actually, like, because that’s like having a load of new engineers and junior engineers like with an understanding of that would be good. On the other hand, you still have this— you still have our generation of people who will run havoc. So that we like— we might build the crazy like evil robots And then they would wipe out all these ethical technologists.
Bridget: The ethical juniors would be like, uh, we learned that you should consider—
Gareth: yes, I think there’s— I think that’s one of the areas where, like, conferences have a, like, an interesting avenue to get to, like, a lot of people.
Bridget: Um, and ultimately, then grown-ups already working in the industry might listen to you, might listen to Anne, and might, uh, pay attention to this ethics track.
Gareth: [00:54:05] The nice thing with the ethics track is there’s— we’ve got some real experts that they can come listen to, not— and me and Anna just helping curate.
Bridget: Well, I will look forward to— I did see that your program is posted, so I will link to that in the show notes and, and tweet about it and tell people they need to come check that out. Um, let’s see, we also have open CFPs for lots of DevOps Days, um, including DevOps Days Minneapolis, which I run, uh, the GopherCon CFP, um, link in the show notes closes March 15th. The conference is August 27th to 30th, and, uh, there’s discount codes for a lot. ADO2018 will get you a discount on a lot of DevOps Days, ChefConf, GopherCon. So if you have an upcoming conference you’d like to see promoted on ADO, you can fill out the handy form at arresteddevops.com/conf. So let’s see, um, links to anything we want people to check out. Gareth, do you have anything you want people to go check out?
Gareth: [00:55:06] Apart from the ethics track, oh, I’ve been hacking on something recently called Kubeval. Basically a local validator for Kubernetes configurations. If people are playing with Kubernetes, feel free to have a look at that. It’s one of my pet projects at the moment.
Bridget: Nice. We will put a link in the show notes to that. Speaking of hacking on fun projects, I’ll put a link in the show notes to this IoT Dev Kit. That they gave us at our Microsoft Cloud Developer Advocacy Summit. And it’s actually— I finally started playing with it and it’s really cool. So that’s something to check out. And I also want to signal boost, there’s a Male Ally Summit in New York on March 29th that Natasha Green and some other folks are running. So put a link for that too, because I think it’s a lot of times those of us in tech may have noticed that tech is full of dudes, and most of them are well-intentioned and wonderful humans who would love to make tech better for those who aren’t dudes. So, um, having the insights from fellow male allies will, I think, be really helpful there. So yeah, um, Gareth, thank you so much for joining the podcast today. This was super fun.
Gareth: [00:56:23] Yeah, no, it’s been good. Oh, thanks for having me.
Bridget: Absolutely. So, uh, head over to arresteddevops.com/devops-weekly for this episode’s show notes. And the site also has our newsletter, Patreon, all the Arrested DevOps stuff you could possibly want. Visit arresteddevops.com/itunes, leave us a review in the iTunes store. I’m told that helps people find the podcast. I don’t know how these things work, but, uh, I’m Bridget at Bridget Kromhout. We’re Arrested DevOps. And remember, there’s always DevOps in the banana stand.