I had a chance to sit down recently with Nigel Kersten, chief technical strategist at Puppet. Besides the fancy title, Nigel is one of the smartest, nicest people you will meet. Always a pleasure catching up with him. We spent most of our time speaking about the Puppet 2017 DevOps survey, which is now open right here.
We also spoke about some of the findings of the past five annual DevOps surveys and the DevOps space in general. Great, informative conversation. As usual, the streaming audio of the our chat is below, followed by a written transcript. But please remember to take the 2017 DevOps Survey. Enjoy!
Alan Shimel: Hi, everyone, Alan Shimel, editor in chief, DevOps.com here, for another DevOps Chat, and today’s DevOps Chat guest is none other than Nigel Kersten, chief technical strategist at Puppet. Nigel, welcome back.
Kersten: Hey, Alan—thank you. It’s great to be back, and always good to talk to you.
Shimel: Absolutely. Nigel, you know—well, first of all, it’s always great to have you on the show, and we’d love to have you on as often as we possibly can, but specifically today, I wanted to talk a little bit about—it’s that time of year again where Puppet will be launching their, the annual DevOps survey, and I wanted to spend a little time talking about it and cluing our audience in, into what’s in store for this year.
Kersten: Absolutely. Always happy to talk about the survey. So I think, you know, this has been—we’ve been running this for quite a while now, this is now the sixth year. And I think one of the things that’s really fascinating about the survey is, we survey generally technical practitioners, but also managers as well. But, in the course of surveying all of these practices and technical behaviors and how teams communicate with each other, we end up coming to a bunch of really great conclusions around the actual business outcomes.
And this is why I think it’s so important for people to be going out and filling out this survey, because I see customer visit after customer visit, every time I’m out there on the road, so many people come up to me and say, “We’ve been trying to change the way we work inside our company, and your surveys turned out to be a really useful tool for handing to our senior management, and saying, ‘This is the sort of outcomes we can expect if you let us work the way we actually want to work.’”
Shimel: Yes. And you know what, Nigel, just a little bit of a frame of reference, I got the number while I was at Gene’s last DevOps Enterprise Summit in San Francisco, but I was speaking to Nicole, Dr. Nicole Forsgren. What was the total—you may not know—but the total amount of people who have taken this survey at this point over the past five years?
Kersten: There are actually over 25,000 people, which I think is fantastic, and particularly even in the first few years, we’re definitely really focused on the sort of North American market. We’ve been seeing this growing more and more in Europe, in Asia, and we’re starting to see that, really, DevOps has become a global movement and is being recognized as such within the enterprise, not just sort of the, you know, webscale hacker groups.
Shimel: Yeah. Well, that’s certainly for sure. And Nigel, this year, is Dr. Nicole—well, they’ve actually started a company now around DevOps statistics and research.
Kersten: Yeah, so we’ve been working with Dr. Nicole Forsgren and Gene Kim, who many people know from the author of, “The Phoenix Project,” and ex-CTO of Tripwire, and the inimitable Jez Humble, who many people know from his Lean Enterprise and Continuous Delivery books. We’ve been working with them for about the last four or five years and, yeah, this year they’ve formed a new company called DORA—DevOps Research and Assessment—where they go in and help companies do benchmarking around their actual sort of DevOps practices.
Kersten: And so I think that combination of industry experience, great sets of contacts, and really, the statistical rigor that people like Nicole actually bring to this—
Kersten: —make this to be much, much more than just a vendor survey. Because it’d be easy for us at Puppet to just survey the Puppet user base and come to a bunch of conclusions. But I think the really interesting fact is when we get to talk to people—more people than are just using Puppet, people across the whole industry, some people who are just coming to DevOps for the first time.
Kersten: So I’d really highly encourage anyone who’s listening, you know, go out and fill out the survey when we launch next week, and don’t just feel like, “I don’t really know very much about DevOps, I shouldn’t fill out the survey.” We’re really looking for input from people all across the spectrum.
Shimel: Absolutely. And of course, Nigel, by the time people hear this, it will be next week, and we will include the URL in the show notes so that people can click right out of this article to the survey.
Shimel: Nigel, I don’t know if you would know this, even, but is there sort of a shortened URL that people can look for, to go to?
Kersten: Yeah, we’ll give you a shorter URL by the time we actually launch this podcast out.
Shimel: Okay. Not a problem. We’ll put that in there. So Nigel, six years in, though, if you’re not moving forward, you’re dying, right? [Laughter] As they say. What’s different about the survey this year?
Kersten: Yeah, so I think we’ve really shown, over the last few years, that there’s a couple of practices that have really been formalized when it comes to DevOps enterprise adoption. We’ve seen that things like version control being heavily used in operations teams and that infrastructure being shared with your development groups, we’ve seen the role of continuous delivery and lean product management practices. So we really feel that, over the last four or five years, we’ve really shown a bunch of this stuff, that it really isn’t changing very much.
So we’ve switched focus a little bit this year. So one of the interesting things we’re really looking at are the role of infrastructure architecture patterns in DevOps environments. And I don’t just mean, you know, “Are you adopting containers, are you doing macroservices, service-oriented architecture?” But we’re really looking for information on, architecture-wise, in terms of the way your application is set up, is your team self-sufficient and autonomous? Can you make changes quickly on your own, without lots of dependencies across the organization, and can you experiment really quickly?
Because I think we’ve seen from Agile and from all of the DevOps enterprise movements over the last few years that the ability to do small tests and iterate quickly on the results of those tests to set your larger direction is really critical to both advancing your service, but doing so quickly and more reliably. Because I think that’s been one of the really huge sort of new things for senior management who haven’t been so familiar with these practices to realize is, adopting DevOps practices doesn’t actually mean sacrificing quality for speed or vice versa. It actually means you’re producing better, high-quality software more quickly.
Shimel: Absolutely. So, Nigel, as I think you mentioned, this week, or the time people are listening to this, the survey will be open. Just in terms of time frames, how long will it be open and when do you think people might start getting a look at some of the data in the report?
Kersten: Yeah, so we should be launching later on in the year as far as the results go. We tend to leave it open for at least a few weeks. Some of that sort of depends upon, you know, making sure we get a sufficiently large statistical base to get the conclusions that we’re actually looking for. So, the faster people want to see the results, the faster we want to see them fill out the survey. So, basically, head off and go and do it now.
Shimel: Right. [Laughter]
Kersten: As soon as we get that critical mass of data, we’ll be launching the results. So you’ll see us talking about this at Velocity, at the DevOps Enterprise Summits and a bunch of the different events happening.
Shimel: Oh, I’m sure, I’m sure—looking forward to it. And being written about on DevOps.com, I might add. [Laughter]
Shimel: For sure. So, Nigel, if you were king for the day, you know—
Shimel: —what would you, you know, and you looked into your crystal ball, what would the headlines be from this year’s survey?
Kersten: If I looked into the crystal ball, what I would really like to see is that we’ve drawn some really strong conclusions around the benefits of DevOps practices in the public sector and in nonprofit organizations, because that’s been a constant source of feedback for the last few years, that we’ve managed to draw really great conclusions for profit-oriented companies.
But I think we’ve also seen, with the rise of the U.S. Digital Service in the U.S. and .gov.uk that’s been running for quite a long time in the UK and the Digital Transformation Office in Australia—we’re really starting to see some of these practices start to be applied to the public sector. And, honestly, that’s in all of our interests, for government infrastructure to work more efficiently and deliver better services that progress more rapidly to all of us.
So that’s something I would really, really love to see, that we can draw those conclusions, which I’m feeling pretty optimistic about. But I think the big headline I would like to see, continuing that thread of, “How do we make sure that the survey and the results in the report that we’ve come out with are useful for practitioners selling upwards in their organization, as well as senior leaders trying to implement change?” is—give your teams more autonomy. Make sure that you’re designing your infrastructure in such a way that they can test on their own, that they can rapidly experiment on their own, and that you can actually progress things much, much faster than we’re you’re needing lots and lots of meetings just to create major changes.
So the more we can actually absorb the lessons of service-oriented architecture, and not just create a mass of complicated services, but create clusters of services that teams can work on with a high degree of autonomy and flexibility. That’s really the headline I would love to see—teams that are able to experiment and are autonomous will produce better results for the business.
Shimel: Absolutely. Nigel, let me ask—go off on a little bit of another tangent, here. You know, one of the things that I kinda, I find fascinating and I tend to come back to and I get as much of as I can is the Project Blueshift stuff that you guys do at Puppet. Gareth Rushgrove is deeply involved in it. And while, today, you know, there’s certainly a strong container focus on it, it really is about sort of future-proofing Puppet, right, and making sure that the folks at Puppet are aware of, you know, what’s the next big thing coming down.
What can we look for in the survey that may give us a clue in terms of what is the next big thing and hence, you know, something for Blueshift to be aware of?
Kersten: Yeah, so I think we’ve generally tried to stay a little bit away from stuff that’s too vendor-specific research for us. But I agree—so you’ve totally captured what Blueshift is about for Puppet. It’s very much about us making sure that, you know, Puppet’s a highly successful project, you know, and used by tens and tens of thousands of companies across the world.
But we’re in a very fast-changing technology landscape right now and I think it’s really critical for us as a company to make sure that we’re not just resting on our laurels there. We’re seeing lots and lots of things change. It’s entirely possible that, in 10 years’ time, everyone will be on completely serverless infrastructure, so to speak. And, you know, you might have been able to hear my air quotes somewhat around serverless, because there’s always someone’s servers running that code somewhere.
Shimel: Right. [Laughter]
Kersten: That old saying that the cloud is just someone else’s server.
Shimel: Uh huh.
Kersten: And I think, really critical for us, we’ve seen that Puppet, the language, has been an incredibly portable model. We started off in the UNIX and Linux space, and it’s proven to be hugely successful in the Windows space for a lot of our enterprise customers, and people really like having that common language describing all of your infrastructure. And we see no reason why, just as you’ve seen the Puppet container build stuff that we’ve done, that we can use the Puppet language to describe and compose containers and maintain that audibility.
We can also describe all sorts of infrastructure, no matter where things come from. So, for us, Blueshift is very much about keeping that eye on the sort of next level horizon over the next six to 24 months, and just seeing what’s emerging and making sure that Puppet stays agile and that we start baking in the features to Puppet the language and the platform that will enable us to all take advantage of those things.
Shimel: Sure. And I—
Kersten: I think it’s really exciting. I don’t think we’ve seen infrastructure technology change as fast as this, ever.
Shimel: It’s crazy. And, you know, I don’t know if it’s the gadget boy in me or, you know, I’ve lived in tech for so long, but it’s that—it’s not even a Puppet thing to me, Nigel, it’s a thing. What is the next thing? I think it’s a little bit of how I got into DevOps, it’s a little bit of how I got into computers and technology to begin with. So I’m always interested and curious.
Shimel: And sometimes, you know, the nice thing about this survey is, it does cast such a wide net of responses that you—
Kersten: Yeah. And I wanna make sure that people don’t feel too left out as well. Because, you know, I often talk to people who are in big enterprise environments, managing important legacy applications, and some of them are still struggling with getting relatively basic principles adopted—of just using version control, using infrastructure as code.
Kersten: And sometimes you talk to them and they get a little demoralized, going, “You know, everyone’s talking about all this cool technology out there, and we’re not actually getting to use it.” But I think the reality is, most enterprise environments, the software they’re running is delivering actual business value, and it’s important to keep making sure that’s as efficient as possible.
We saw from the “DevOps Handbook” that Gene and Jez and all of those people are involved in that you can apply these principles to mainframes, you know? You can see people taking incredibly legacy technology, applying these modern ways of working to them, and getting really, really huge benefits.
Shimel: Got it, got it. Nigel, we’re almost out of time, unfortunately. You know, we can talk all day, but people only listen, it’s the curse of the modern, you know, the 10-minute or the 10-second sound bite and the 12-minute podcast—but first of all, good luck with this here survey. You know, I encourage anyone listening to go, please click on the link and fill out the survey. Can’t wait to see what some of the results are this year.
Continued good luck with Puppet, too, as you guys are really, you know, coming out of Puppet Conference and meeting with Sanjay and some of the exec team, I was very excited about the possibilities, and I’m sure you are more than me—
Shimel: —about, you know, what’s on the horizon here for Puppet, so good luck with that as well. And thanks for being our guest on DevOps Chat.
Kersten: No problem, Alan. It’s always great to catch up with you.
Shimel: Okay, my friend. Nigel Kersten, chief technical strategist at Puppet, thanks for appearing on this DevOps Chat, and thanks for telling us more about the upcoming devops survey, or sixth annual devops survey from Puppet.
This is Alan Shimel for DevOps.com. Have a great day.