In this DevOps Chat our guest is Lucas Carlson, newly appointed VP of Strategy at Automic. Of course, Lucas has already had a distinguished career. Here is his bio from Linkedin:
Lucas Carlson is an author, entrepreneur and executive. He also has a background of 20 years of programming to back it all up.
He’s written scalable software that has been downloaded millions of times and used by millions of users. As a keynote speaker in the technology arena and innovation leader in the cloud space, Lucas helps companies discover opportunities for growth.
Lucas wrote a bestselling book on Ruby for O’Reilly called The Ruby Cookbook and has created over a dozen open source libraries, some of which have been downloaded almost a million times.
Lucas was formerly the CEO and founder of AppFog, a massively popular DevOps startup acquired by CenturyLink in 2013. AppFog signed up more than 100,000 developers, raised nearly $10M in venture funding and has become an industry leading Platform-as-a-Service.
Before following his passion and writing books, Lucas served as the Chief Innovation Officer for CenturyLink, in charge of building next-generation Docker and Linux Container products for a $20B company.
Lucas also writes for major tech and entrepreneur publications including Inc, Business Insider, InfoWorld and InformationWeek.
I spoke with Lucas about why he joined Automic. Our discussion proved to be a provocative one, in which we explored Lucas’s view that DevOps is not a great thing for Ops. Ops needs its “own DevOps” that is more geared to its challenges and world view. Of course, many thought that DevOps is just that. Before you comment, have a listen to our conversation to see what Lucas said. I am interested in your thoughts and comments on this.
Alan Shimel: Hi, everyone. Alan Shimel, editor of DevOps.com here for another DevOps Chat. Very happy to be joined in this DevOps Chat by VP of Strategy at Automic Lucas Carlson. Lucas, welcome.
Lucas Carlson: Thank you so much. Glad to be here.
Shimel: Thank you. Lucas, I know you’ve been at Automic now, what, about two or three months?
Shimel: But I’ve known Lucas Carlson for years before Automic and I don’t know if our audience is familiar with you, but you’ve had quite a storied career already at a relatively young age.
Carlson: Thank you. Yeah.
Shimel: Why don’t you give our listeners a little bit of the Lucas Carlson story.
Carlson: Sure. Well, I started out as a programmer: for 20 years I wrote code. I started code when I was in elementary school and ever since I’ve been writing code. And I wrote two programming books during my career. I’ve mainly focused on the open-source side of things so Ruby, PHP, Python worlds.
And then I started a platform-as-a-service startup about seven years ago and I used Cloud Foundry as the foundation for my startup and it was called AppFog. We signed up 100,000 developers, it was a lot of fun. It was a great time.
We were acquired by a telco and it was really interesting to spend a few years inside of a telco and watching them attempting a digital transformation—trying to move their applications into the cloud was a very interesting story.
And after that I quit that job, I moved to Paris for a while—it was a life dream of ours—moved my family, and after Paris I looked around for what to do next and this opportunity came at Automic, and I thought it was a very interesting opportunity for reasons we might get into.
Shimel: Great. Great. And of course, you know, interesting with your background and accomplishments Lucas that Atomic, which is an automation company, is where you wound up and as VP of Strategy there.
So as someone who was a chief strategy officer at the last company I co-founded, Lucas, I can appreciate the title of VP of Strategy. But for our listeners who are saying, “Well, that sounds like just another corporate title,” give them a little insight. What’s your role as VP of Strategy at Atomic?
Carlson: Yeah. Well, my role is to take a company that’s been around for almost 30 years that’s had a product that is battle tested, been around the block, really works, works at scale, has thousands upon thousands of enterprise scale customers, has a lot of pedigree and DNA, and yet is a company nobody’s heard of.
That’s actually why I kind of joined. I thought it was a really interesting challenge to take a company that nobody’s heard of yet has had a product that is—has stood the test of time, is very, very robust, very strong, and see if there wasn’t a way to kind of help that company get recognized for the work that they do, which is actually really great work.
And so I joined the company with that kind of concept in mind. And it’s been very interesting because fundamentally when I came into the company they called themselves a DevOps company and I come from a background—I’ve been doing DevOps for a long time. I grew up around Portland, I hung out with Luke Kaines who founded Puppet and he was one of my mentors when I was building AppFog and—so I knew DevOps inside and out. I used DevOps to build my own startup.
And so I came in and I looked at what Automic had and Automic has a very interesting, strong product, but it’s not DevOps, because DevOps is predominantly—you know, having grown up around it I know you know DevOps is not some mystical union of developers and operations into this perfect combination that can do both and is just as adept at writing Ruby as they are debugging MySQL.
You know, that’s kind of a story—a fairytale that people like to tell, but it’s not the truth. The truth is, DevOps was created by developers for developers to get around operations. They were sick and tired of waiting for operations to get around to provisioning their SQL server so they wrote some code that did it for them.
They were tired of kind of the delays and lags so they wrote code that got around it. It wasn’t really meant to create this mystical union it was really meant to bypass operations.
So, I guess, how does that apply to Automic? So, Automic came in saying they were DevOps—they weren’t a tool for developers and yet I thought they have this long track record. They’re doing something right, but what is it?
And I spent a lot of time thinking about that before I joined and when I joined I became convinced that there is a huge part of the DevOps story that is getting left out: it’s getting forgotten, it is getting sidelined and ignored that doesn’t deserve to be ignored.
I think that DevOps isn’t enough on its own. Having run a platforms and service company myself, having put platform as a service into major name-brand banks, having seen people trying to adopt these DevOps technologies—what I’m seeing over and over again is that one of the biggest struggles about DevOps is a myth that I’m trying to shatter. And the myth is that DevOps doesn’t need operations: that DevOps outsources operations, that operations isn’t necessary for DevOps.
In fact, I think the absolute opposite is true and I don’t think ops is becoming part of dev. I think that ops deserves its own practices equivalent to what DevOps brought to developers. Equivalent and yet different, because I think that fundamentally operations was a different world of developers: always will, always has.
To developers, failing fast is important and breaking stuff is totally acceptable. And going iteratively is great. So if you write some code, it doesn’t work you can fix it really quickly because you failed fast and it makes it better. And that kind of iterative approach is something that works really, really well for developers but it does not translate for operations. Operations cannot fail fast. If you fail in operations and your database goes down in production, that could be millions upon millions of dollars lost for the business.
So it’s not that operations doesn’t want speed—of course they want speed, it’s just that speed is not the primary. For developers, speed is primary; for operations, reliability, robustness, scalability trumps speed. It’s not that they don’t want speed, it’s that they have a different core value set—a different world view—and instead of trying to mush these things together, instead of making them one idealized person that’s never going to exist, what I say is, let’s raise up the ops side of DevOps, bring it to the same level that developers have been raised. When developers adopted DevOps practices, it raised them up to a new level as a power-up—it was a mushroom in Mario. And what we need is to give the ops side a mushroom.
We need ops to level up and be able to play at the same level as developers, but with tools built for operations, not built for developers.
The tools out there for developers in DevOps are not the right tools for operations in DevOps and that’s what –
Shimel: Hold on, I’m going to stop here because I’ve got to jump in—I’m sorry—I’ve been biting my tongue.
So this is actually a pretty—I don’t want to say radical, but a very different kind of view of DevOps and what goes on between dev and ops than I think we hear from most people.
And it actually reminds me I had a conversation—actually a DevOps Chat—with Dave West, who is the CEO of Scrum.org. And you know, obviously, Scrum and part of Agile is really aimed at developers. And though he saw sort of a potential bridge from dev to ops he had a similar view of yours that, you know, we have tools that are written for developers that developers are using that are letting them get more involved in ops but how much is ops really being involved in this DevOps, if you will.
But yet, you know, let’s look—like configuration management tools.
Shimel: Do you think those are dev tools not ops tools?
Carlson: Yep, I absolutely do.
Shimel: Absolutely. And –
Carlson: I could go into that more but I think it’s self-evident. I think that developers use those tools vastly more than operations. I think that the reasons are obvious. The tools were built for developers by developers.
Operations looks at the world through a different lens; they use a different way of approaching the same problems and they kind of—the way that configuration management works embraces a development mindset.
Shimel: Okay, fair enough. Let me ask—so how does this, Lucas, play into Automic and their future toolset? Will we –
Carlson: Well, I joined Automic specifically because I saw this as the biggest problem in DevOps today and that Automic has solutions on the ops side of DevOps. Automic’s solutions are the level-up for ops teams. They don’t say—like one of my biggest pet peeves right now is this rip-and-replace mentality. You see it … with Mesosphere, with Docker containers, with Cloud Foundry, with OpenShift. All the big vendors out, there they all have the same promise: “Take everything you’ve done, throw it away, import all of your applications into our platform that will do it all for you and it will solve all your problems.”
Well, I’m sorry, but that’s just not going to happen. Those platforms are generally underdeveloped, undertested at scale, they are a brand new way of managing applications that is untested. They don’t embrace the vast majority of enterprise scale applications architectures. They basically force you to rip and replace. And rip and replace is not just a thing for the code it’s a thing for all the IT processes that have been known to work for the last 10, 15, 20 years.
So you’re trading a potential promise for a solution to all your problems, but the risks are exponential and vastly untested. So this whole concept of rip and replace I think is very, very dangerous in our industry today and very prevalent.
And I joined Automic because their approach is to embrace what already exists and make it better. Instead of saying, “We’re going to have a new platform for you to import all your apps to,” they have a way to describe the topology of your existing infrastructure and built on top of that and automate that instead of throwing it out and replacing with something new.
And so, I think to me, it really levels up the ops side of DevOps—it makes operations a hero, the co-hero of the DevOps story. It doesn’t deny operations their vital role.
Shimel: Excellent. Excellent. You know what? Lucas, I’d love to see this at a Speak—you’ve got to submit to Speak about this somewhere.
Carlson: Yeah, for sure.
Shimel: Or, maybe we can do a webinar and get into it because honestly, our DevOps chat format being just 15 minutes doesn’t give us enough time to really delve in here and I’d love to get a counterview of these things too and let’s really have a discussion around it. I think I’m going to take that as my job. I’m going to go find someone and we’ll go back and forth on this.
But unfortunately we are just about out of time—we’re at 15 minutes already. As I said, it goes quickly.
Carlson: No problem.
Shimel: This has been a very interesting discussion. I would call it truly an ignite or a fire starter in that you’ve opened the door here and we’re not finished with this one Lucas. We’re going to come back to it.
Carlson: Sounds good.
Shimel: But for now we’re going to call it a wrap. I hope if you’re listening to this you found the conversation interesting. Of course on DevOps.com you’ll also see a written transcript of the conversation so you can hold Lucas to his words here and it’s going to be in writing.
And if you have any questions, comments, feel free to also put them on DevOps.com we’ll get them out to Lucas, and as I said we’re going to try to continue this discussion at a later date.
But for now Lucas Carlson, VP of Strategy Automic. Thank you so much for really kind of blowing the doors up today on this and we look forward to seeing what the follow-up’s going to be. And success at the new position.
Carlson: Thank you so much. I’m very happy to be here.
Shimel: Thank you. Yeah, no, it’s a great company. We’ve worked with Automic for some time and I’ve met a lot of great people there.
Anyway, that’s a wrap for this edition of DevOps Chat. Lucas Carlson from Automic was our guest and this is Alan Shimel for DevOps.com. We’ll see you at the next DevOps Chat. Thanks everyone, have a great day.