57. DockerCon 2019 – A Preview Show

Chris: In episode 57 of Mobycast, we look ahead towards DockerCon 2019, discuss what to expect and throw down a few predictions. Welcome to Mobycast, a weekly conversation about cloud native development, AWS, and building distributed systems. Let’s jump right in.

Jon: Welcome, Chris. It’s another episode of Mobycast.

Chris: Hey, John. Good to be back.

Jon: Yeah, good to have you. It’s just you and I today. Rich is spending some family time and we are out of practice because you were on vacation.

Chris: I was, in a whole different time zone.

Jon: A whole different time zone. Did you go to Montana?

Chris: That would be a different time zone. I kind of wish it was one only time zone away, at least right now. My body is massively confused—my internal clock. I was in Italy, nine hours difference between Seattle time, and then ended up having an unexpected multi-day layover in Chicago. My body know doesn’t when to go to sleep, when to wake-up, when to eat. Everything’s coming up whack.

Jon: It probably doesn’t need to it. You did plenty of that in Italy.

Chris: Yes, I did. I definitely carbo loaded for sure. But I worked it off; we walked. According to my Apple watch, I think it was about 80 miles total, and spread over seven days.

Jon: That’s fantastic. That’s really great. Yup, no trips to Italy here. We just did Kelsus stuff and tried to hold down the fort while you were away, not an easy task but here we are back. Today we’re going to talk about DockerCon. It’s coming up in two weeks, well, by the time you are listening dear listener, it’ll be one week. It’s coming up in a week and you’ll be there. This is your third one, Chris?

Chris: This will be my fourth, actually.

Jon: Fourth DockerCon.

Chris: Yeah.

Jon: Nice.

Chris: Yeah. It’s kind of interesting that I hadn’t gone to conferences for quite some time and then really DockerCon was the one I got back into it with and that was back in 2016 in the Seattle version. It was one of those things that was super easy for me to go to, there is no travel. That’s coincidentally when I first started using Docker. It was just awesome timing. I’d started using Docker in January 2016 and then the conference I think was in June. It was just perfect timing. That edition was just […] and it’s a great conference. It was really good.

Jon: Very cool. Yeah, it’s changed over the years. Let’s talk about what’s coming up and then maybe later in the conversation we can talk about how its changed over the years. What are you expecting this year? What’s going to happen?

Chris: This year is going to be interesting in that I don’t expect anything too terribly dramatic. Last year was a big shift in culture and content. We’ve talked about that on a previous episode of Mobycast when we recapped the 2018 version. But big changes happened in 2018 with Docker with the departure of both Ben, the previous CEO, as well as Solomon, the CTO. Kind of like a one-two punch there, changing of the guard, and brought in new CEO, some other new management. Previously, Docker was all about innovation and just lots of different directions and technology, and they absolutely made some missteps there in trying to be overly ambitious but at least there was a lot of excitement, there was a lot going on, a lot of innovation.

Last year, it was made very clear that, “Okay, new management brain coming in.” Steve Singh, the CEO, it’s all about Docker as a business and really funny in what that business model is. A really dramatic reduction in innovation. The number of product announcements that they had, much, much less and just really loud and clear that focus on, “What’s the business model? Making money,” and really having this thing grow up and understanding that it is a business, it’s not a technology innovation lab, if you will. It’s like, “It has to make money.”

I really expect this year to be more of the same. I don’t expect a lot of new product announcements. I expect them to continue to focus on the enterprise edition of Docker. Focus on hybrid environments, focus on legacy apps and monoliths, and how do you easily give that support to the enterprise companies so they can containerize their existing applications without rewrites, and continue the big bear hug around Kubernetes.

Jon: Right. When you expect, it’s kind of like, lots of education, lots of best practices, and lots of, “Hey, here’s a unique way to use this existing thing. You could use something that you might not have thought of before,” and not so much announcement. None of those types of sessions or maybe a very few of them.

Chris: Absolutely. I definitely expect a lot of lather, rinse, repeat and just more practical like, “How do you use this stuff.” To some extent, what else are you going to do with Docker? It’s very robust, stable, it’s mature, it has everything you need for containerization. In that respect, there’s just not a tremendous amount of new tooling or product around it. Unless, it’s driven more from a business model standpoint.

Jon: Sure. I have in front of me here that DockerCon has grown over the years; 500 in 2014, 4000 people, 5500, last year 6000. More people are using Docker. It’s still growing. Do you expect DockerCon to actually be bigger or do you think it might shrink a little bit this year?

Chris: For sure. When they release numbers, it’s not going to shrink. Whatever the real numbers are. Last year, it was, I think about 6000. The previous year was 5500. It’s not a huge jump and so this year I would expect they’ll probably say it’s something like, 7000 attendees. I kind of doubt that will be the case. It’s at the same venue; it’s in San Francisco at Moscone.

Jon: Funny ways of counting attendees might be like, “Oh, well, if you’re a gold sponsor then everybody in this organization of your company is an attendee.”

Chris: Yeah, absolutely. Amazon too, so 200,000.

Jon: Right.

Chris: They’re definitely not nearly as big as some of the other conferences around. But again, it’s a different flavor. You have someone like AWS and its conference with re:Invent. It’s a commercial conference versus DockerCon, especially over the last four years has transitioned more to like, “This has more in the field of open-source and community and what not.” I think that limits the attendees if you will. It feels a lot like some of the O’Reilly conferences like OSCON.

Jon: Cool. What’s the layout of the conference? What do you expect there? How long is it? That kind of thing?

Chris: It’s pretty compressed especially compared to something like re:Invent. It’s really two days long. Officially, I think it’s four days, but the first day is a pre-day where there’s paid workshops that you can attend. The fourth day is a recap of popular breakout sessions. It’s really the two middle days, Tuesday and Wednesday, are the core days of the conference. Each day starts off with a keynote session and then it’s followed by the breakout sessions. Basically, two days of full content and then some additional stuff on the tail ends; at the beginning and the end. About 100 breakout sessions which is, I mean in a way…

Jon: […], right?

Chris: It’s kind of refreshing. You can actually look through the catalogue, register, and it takes you 10 minutes versus re:Invent, I think I spent probably 5-10 hours going through that, trying to line up my agenda for it because there’s 2000 plus.

Jon: With like three weeks of FOMO.

Chris: Yeah. Definitely, much, much smaller, more manageable and more compressed.

Jon: Right on. I’m going to actually skip a little part of the outline since we were just talking about sessions, maybe you can talk about the tracks that it has.

Chris: This has remained pretty consistent too across the various DockerCons. It’s a mix of high-level stuff, some low-level stuff, and then everything in between. Some of the main tracks are, they have a Docker Tech talks. These are sessions that go deep on the products and technology that make up the Docker platform. It’s pretty hands-on, pretty pragmatic information.

There’s a Docker for Developers track which is really focused on the developers that are using Docker on their machine, on their laptops. How do they really get the most, so they’re using the tools like Docker Compose, Docker for Mac or Docker for Windows and what not.

Then they have a track Docker for IT Infrastructure and Ops. That’s your DevOps type stuff. How do you actually use Docker in production? How do you deploy things? How do you run them? How do you maintain them? How do you monitor them? Those kinds of things.

They have a Customer Case Studies track. Obviously, this is customers that are using Docker and they get up there and talk about how they’re using it, the benefits they get, inspire other folks to do the same.

They have a Black Belt session track. This is usually really deep into the guts, the internals, of the Docker platform and related things. I’ve spent a lot of the previous DockerCons in Black Belt sessions. I can definitely attest, they usually go really, really deep. The other thing to point out too, at DockerCon, their sessions are 40-minutes long versus re:Invent is 60-minutes long. That 20 minutes makes a big difference especially with these Black Belt sessions, 40 minutes, going really deep, like pulling apart how network works inside Docker or how does storage work in device […] drivers, and what not to get all that within 40 minutes. They usually go through; their slide decks are 100-slides long. You just feel you’re like fire hosed.

Jon: They’re not actually trying to give you a little less content so that it’s more manageable; they’re just trying to shove it all into 40 minutes.

Chris: It’s absolutely what it feels like. It feels like it’s 60 minutes or even more of content in a 40-minute session.

Jon: Wow.

Chris: That’s kind of being in your toes, pay attention, otherwise, you’re going to miss it. A few other tracks they have, they have an Ecosystem track which is just the whole community that makes up Docker; all the vendors and other open source projects and partners and what not which is interesting because you don’t really see that. I mean, you see a little bit of that at something like re:Invent but not nearly as much devoted to that as you’ll see at something like DockerCon.

And then they have a Transform track which is a bit more of a wild card, open-ended one where it’s just like, “How has this enabled various transformations or changed culture and processes,” and things like that.

Jon: One of the tracks there caught my ear because it reminded of a tweet, I saw from Clare Liguori yesterday. She was asking people that develop with Docker, how they interact with Docker on their machines while they’re developing. It was a poll. It was like, “Do you use Docker CLI? Do you use Docker Compose? Do you use some sort of IDE type of connectivity, like a button to run your container? Or other?” And then it was like 75% was Docker CLI, 20% or even more was Docker Compose, and then very few people were on other or IDE capability.

She shot back, “Huh, kind of funny. I really enjoy using…” I think she might be using VS Code or something or full-on Visual Studio, “…I enjoy using that because it just automatically will attach my debugger to the container—to the process running inside of the container—that way, I don’t have to do that manually and I can have a debugger going.” It just kind of took me to this trip down memory lane of times in my career where I’ve had a debugger available then got into a technology where a debugger is not the way you code and then going back with mobile development and Objective-C and it’s like, “Oh, the debugger is back. What a luxury.” And then often a JSF we’re not doing debuggers right now anymore.

The whole of thing of a debugger is that they’re awesome, they’re incredible, they’re so cool when you have access to them but they’re also painful to set up in certain situations and Docker is one of the situations. I’d be curious for you to go, snoop around that track and just see how mature that stuff is and whether it might make sense for our team to start doing that.

Chris: I think we’ve touched on this on a previous episode about the Docker learning curve and it kind of changes; you just have to change the way you work a bit. Debugging inside Docker, I mean, the tools have been there now for years. The S Code, in 2017, they showed it off just being able to debug running containers and […] and all that kind of stuff. You can make all that work. It does take a little bit more, if you have to look into it and set it up, just make it happen. It’s just whether or not it’s something that you really need. Like you said, to some extent, gotten away from using your IDEs within a created debugging and step in through code.

I know me, personally, the first 20 years of my career, I use Visual Studio and using a debugger was integral to what I was doing. I would almost always, I would be stepping through code as a way of just doing unit testing or some integration testing, just going through and making sure it was doing exactly what I expected it to be doing. When I switched over to Node in 2012, at that point, there really was no good debugger at all for Node. It was like, I went from this really nice integrated IDE with integrated debugger and profiler and everything else, and now I have a text setter and that’s about it.

Jon: Well, I don’t want to go too far down this but debuggers are more technologically advanced. The ability to really see what’s happening in code is an advancement in technology that happened because people wanted to see that. Going away from debuggers is a trade-off. It’s like, “I want these capabilities of this new language and I want to be able to […] in this new way of doing thing and I’m willing to give up my debugger for it.” But if you can get the debugger back and sort of have the best of both worlds—I know we’ve talked about this separation of having your container sort of not ever be allowed to have only-works-on-my-machine type setup—but if your IDE is really good at helping you manage that, if your IDE helps you set-up those wallets, and let’s you debug fully and then let’s you be confident that when you’re really seeing it, you’re taking away all those little things that you don’t want to release into production such as for example, a shared file system, then that can be a great thing.

I guess, my argument here really comes down to things like, there have been times in my career where I’ve used the debugger and I’ve noticed, “Oh my god, I didn’t realize that was happening.” I didn’t realize it because the code was working, everything seemed happy, no problems, no unit test is saying there’s something wrong. But the debugger was like, “Oh my god, why is that file handle still around?” I would’ve never known that without this debugger or without a production outage. You know what I mean? There’s real value in debuggers.

Chris: Absolutely. Maybe the tools are out there whether you’re doing Docker development or not, there’s quite a bit of a tooling community. You can do the debugging inside the container, you just set it up right and learn how to do it. It doesn’t take that much time. If you like debuggers and you really want that capability, you can have it with Docker. As a side, I think what happens with most people is that when they’re to debug, they just debug it running outside of Docker which is a big no-no.

Jon: Yeah, exactly. That’s kind of what I wanted you to poke around there. Another thing we talk a lot about in Mobycast is the current feeding of your own tech career. I think that a lot of the times developers spend a lot of time learning the new technologies, the new languages, the new ops stuff, and then before you know it, you haven’t spent any time updating your development process. Learning new hotkeys in an IDE or trying out a new, more powerful development practice, that stuff, it’s like, “I got my way I like it. I’m just going to keep doing it that way.” But updating that part of your skillset can be very valuable.

Chris: Indeed.

Jon: Cool. I think we’ve spent quite a bit of time here. Oh my goodness, we just took a nice, long detour. Where should we go next, Chris? Let’s get this back on track.

Chris: Sure. Definitely, it’s kind of a mission accomplished in that. This is all about a previous show of DockerCon 2019. We’ve talked about some of the history, what we can expect with this new one, what the content is like, the layout. This is, again, different than bigger conferences like re:Invent. In re:Invent, it definitely makes sense to talk about a lot of more practical tips, you need to pace yourself, you need to wear comfortable shoes, and hydrate and sleep, and all that kind of stuff. The good news here with DockerCon, it’s just more relaxed; it’s not nearly as intense. It is a great conference with a lot of good technical content and pragmatic content, lots of passionate people there that are really into it, and you can totally geek out. From that standpoint, it’s just a lot more manageable and in some ways, maybe even a bit more enjoyable.

Jon: I want to ask you a really important question because this could play into some future episode of Mobycast because we like to go through some talks. You’ve already started to look at your agenda, right? Are there any particular speakers that you’re excited to hear or watch?

Chris: Off the top of my head, not really. There is some talks from folks at Netflix. There’s one on just […] container security, theoretic and practical implementations of that at Netflix, so that’s interesting.

Jon: Do you happen to have in front of you who’s giving that talk?

Chris: That is Michael Wardrop.

Jon: Cool.

Chris: There’s one specifically around node, Node.js. That’s done by Bret Fisher who’s kind of one of the Docker Captains in is longtime member of the Docker Community and he’s spoken before. There’s the Dockerfile Best Practices, […] of this year is done by a couple of different folks. There are a lot of the same speakers as before. I wouldn’t be surprised if Abby Fuller is here again. Mostly, I’m interested to see, “Okay, what’s new? What are some of the pragmatic little tidbits that are going to make life a little bit easier?” And also, just kind of get the overall feel and flavor of what’s happening in the community and where it’s going.

Again, we’ve talked about in the past about what’s the future look like for Docker, is it going to stay independent or is it going to get acquired, are they going to figure the business model, and then also this whole Kubernetes. Last year was such a huge, dramatic switch for them. Previously, they did support Kubernetes, but they were really focused on Swarm, and last year Kubernetes was everywhere. This year, there weren’t as many sessions that are around Kubernetes however if you go to the DockerCon website, the primary subtitle for it is, “DockerCon is the #1 container industry conference for all things Kubernetes, microservices, and DevOps.”

Jon: Wow.

Chris: It doesn’t even mention Docker. They have eight or nine paid workshops, at least two of those are on Kubernetes. They’ve definitely doubled down on Kubernetes so it’d be really interesting to see just where they go from there.

Jon: It will be interesting because Kubernetes has been getting a little bit of heat lately. I’ve seen sentiment turn against it pretty hard on Twitter. Just yesterday I think or maybe it was two days ago, Tinder released a tech blog basically going through the millions of dollars and weeks and months that they spent moving over to Kubernetes and how difficult that was for them. Kelsey Hightower, one of the Google Cloud people, said a while back that, “A great use case for Kubernetes is if you have developers releasing applications that are different and disparate into one area, if you’re making your own Platform as a Service, look at Kubernetes; if you’re not, don’t.”

Making a Platform as a Service may be some things that some enterprises do, maybe a telecom company does that, maybe cable companies—big, big companies, big enterprises do that for sure and they maybe do it on-prem, Kubernetes is right for them. But for everybody else, it’s not. It’ll be interesting to see if Docker which we expect is really still trying to embrace the enterprise and try to figure out how to monetize the enterprise. It would make sense for them to still be all about Kubernetes where it makes sense.

Chris: In a way, Kubernetes kind of have the same problems now that AWS has. We’ve talked about what AWS, the rapid pace of innovation, 1000 releases a year, new services all the time, and it’s just really hard to keep up. In a way, Kubernetes kind of has the same issue. It’s a huge beast, it’s being added to constantly, and it’s actually really hard to comprehend all of it and to be well-versed in the entire platform or even in core pieces of it. That’s a big limiting factor in adoption by these folks that don’t have 100 plus person IT teams.

Jon: Exactly.

Chris: It’ll be interesting to see if Docker does something there to make it much more palatable and easy to just work with without having to go spend months and months figuring out how to get Kubernetes to work and to run it reliably in production.

Jon: That would be interesting. That could be our prediction, maybe we could leave everybody with that; that if Docker is going to resurrect itself, DockerCon is going to resurrect itself as a place to be, and that is in some sort of onboarding ramp to Kubernetes that’s realistic for smaller organization would be potentially the way to do it.

Great talking to you. I hope you have a good time at the conference. When we get back, we’re going to have so much to talk about.

Chris: You bet. Alright. Sounds good.

Jon: Talk to you next week.

Chris: See yah. Bye.