Interview Phil Estes - IBM Distinguished Engineer, CNCF Ambassador, and ContainerD Maintainer

I apologize in advance for the loud background noise but I couldn't pass up an opportunity to interview Phil Estes our first interview on The Byte where we talk about his first computer, ContainerD, and since Phil is frequenting a lot of conferences and sometimes reviews Call for Papers(CFP) I ask him some CFP tips and general conference tips.

Phil Estes -
ContainerD -
ContainerD Maintainer Michael Crosby -
Tim Berners-Lee's First Web Browser 1993 - of the Web at CERN -

Episode Transcript

Brian Christner: Welcome back to The Byte. In this episode, we're going to be interviewing Phil Estes. Phil Estes, correct?

Phil Estes: Yes that's correct.

Brian Christner: He is an IBM Distinguished Engineer for IBM Cloud, ContainerD maintainer, correct?
Phil Estes: Yeah.

Brian Christner: Member of the Technical Oversight Board for open containers, recently Cloud Native Ambassador, and home-based in Virginia. That is quite an impressive resume to be honest.

Phil Estes: Yeah, yeah. Especially that I live in Virginia and do all that.

Brian Christner: That tops off the cake, right? So we're here in Switzerland today, Phil is actually flying through for a conference that he's attending later this week, or tomorrow actually. And he came by and he's visiting us at Spaces here in Zurich, so that's really cool of you to join us. Now, where are you going after his?

Phil Estes: So this week, got a really neat opportunity to go to CERN. Initially just was going to talk about ContainerD with interested parties there and actually, we've got few other maintainers coming in because KubeCon is next week. There was some ability to kind of add this to people's schedule. So we're going to talk some about ContainerD on Friday but then Jess Frazelle, who I've been working with way back in the Docker open source community days, she had always wanted to visit CERN. So anyway, one thing led to another and now we're both speaking tomorrow in kind of the main auditorium just what they call an IT seminar, give a talk on some topics so I'm going to talk about open source and containers. So yeah it'll be fun.

Brian Christner: Now for those who don't know CERN, they trying to make a black hole with a giant particle collider. And they're actually trying to find what particles make up human beings and all the matter around us, which is quite interesting. And they probably have one of the largest IT infrastructures in the world.
Phil Estes: Yeah, yeah. And actually some pretty interesting historic infrastructure. You know, they've got the NeXT first web server that Tim Berners-Lee ran, that sat at CERN and ran the first few websites. And I think when you visit their data center, you can kind of look through the glass and there's a sign hanging over a certain spot, like the first internet router was here, so it's a pretty interesting place. And like you said, an amazing amount of compute and storage because of all the experiments going on there and they're obviously very interested in cloud and modern technologies to help them kind of operate this infrastructure for the scientists and the researchers.

Brian Christner: I mean that's an impressive facility and it's going to be an amazing event that you get to attend.

Phil Estes: Yeah.
Brian Christner: Now next up, I want to ask you about your first computer. Can you tell us way back when your first computer and what it was and what you did with this first computer?

Phil Estes: Sure. Yeah, so at the time I was in junior high I lived in kind of a rural, not a tiny town but a small town in Illinois. We had the one like a local mall with like a Radio Shack and I would actually walk there after school and play around on theses TRS80 computers which were kind of the early modern PC that you could actually off-the-shelf buy. I know there was some earlier computer equipment you could buy but the TRS80 was kind of commercialized and Radio Shack was pushing it pretty hard. And so anyway, one Christmas my grandfather went in with my parents knowing that I had a strong interest in ... a TRS80 showed up with a two like five and a quarter floppy drives and I think we even had the acoustic coupler for like tape. Anyway, all the hilarious gizmos of that era of computing. And it came with Frogger, a couple of other games.

Phil Estes: But I pretty quickly learned the commands for MS-DOS and got a book on Basic and thought, you know, I just want to see what I can program. So I was just writing silly programs trying to paint things on the screen, make noises. And so yeah that was my first exposure to computers and programming and I guess you could I say I never kind of lost that bug. Just the interest of trying to see what you can make it do. So yeah.

Brian Christner: Nice. I mean, it's incredible when we think back to our first computers and where we're sitting today, it's always a nice journey. So now you're working IBM for quite a long time, actually I remember we discussed this and you what, about two years ago, became a distinguished engineer, or year and a half ago or so.

Phil Estes: Yeah just a year ago.

Brian Christner: So tell us about that journey, how you became a distinguished engineer.

Phil Estes: Yeah, so the cool thing about IBM is that they have, especially as you advance in your career, there's a very clear and discreet path for technical advancement. So for example, there are websites that detail kind of the skills expected, lots of materials to help you kind of understand how to grow and find gaps where you need to work with your manager to find out okay, to be this next level on the technical ladder, I need to do these things or have this kind of scope of my visibility to the rest of IBM. And so distinguished engineer is kind of the culmination of a lot of that because it's the first sort of executive rung on the technical ladder at IBM. And there's really only one above that and it's IBM Fellow, which is a pretty significant accomplishment. There are only a hundred or so active IBM Fellows in the entire company, which if you're in a small company a hundred is a big number if you're in IBM a hundred is a very small number.

Phil Estes: But yeah, so I would say that distinguished engineer is not only about being smart or being technically astute, there needs to be a breadth of something that you're seen as a leader on. And thankfully I'd say I'm lucky that containers and open source and all these things kind of came together at a time when I just happened to get involved and become known as the guy who knew about Docker and containers and then IBM decided to build kind of our cloud platform around that. And so yeah, I mean the timing was perfect for me to kind of expand my scope in IBM to be seen as a leader to where my management and those who supported me could honestly take it forward to, kind of first the Cloud Unit Review Board, and then it's actually a corporate recognition so DE is an appointment at the corporate level.

Phil Estes: You know, it's not something you can sneak into, you got to have a ton of support across IBM, you got to have the right people kind of pulling for you, and so I'm just thankful I had some amazing people around me, management that brought that together. Because it's something I never necessarily thought I would reach at IBM and it's a cool thing.

Brian Christner: I mean it's really an amazing achievement. I mean considering the size of IBM, I don't know how many employees, must be a hundred thousand-

Phil Estes: 400,000.

Brian Christner: 400,000.

Phil Estes: Worldwide.

Brian Christner: So I mean, it's a small, smallest percentage of actually become distinguished engineers. So I mean, that's really an accomplishment. Now you mentioned the open source and how you started with Docker, let's talk a little bit about ContainerD and how you got involved in ContainerD.

Phil Estes: Yeah, so I was working on Docker, the open-source project, obviously 2014, 2015 into 2016 it was hugely popular as an open-source project. It was also under a lot of stress from just the amount of people wanting to kind of give their input, make their mark on Docker, you know, whether it's vendors or independent people. It was being pulled in a lot of directions and of course Docker the company also had specific ideas and strong opinions on Docker the open-source project. And it was causing some tension, Kubernetes Swarm, that was kind of a big excitement in 2016. And ContainerD really came out of a set of discussions with many players at the time, you know, there'd been some public calls for we just need a stable core run time, that's not opinionated, that we can all build on. Docker can continue to build their platform, people that love Kubernetes can build on it.

Phil Estes: So ContainerD, you know again, came out of Docker. Michael Crosby had a huge hand in kind of putting that together. It originally showed up in 2016 as kind of a management layer over runc, which is the OCI layer that also appeared that year. But it was really late that year that, through my involvement with Docker and talking to Solomon (Docker Founder) and all the people at Docker, that we really agreed this is something that should be outside of Docker, should be in CNCF or wherever you guys think is best. And so, you know, early 2017 it was donated to the CNCF. And so again, I felt like it just made sense for IBM to continue involvement to get even more involved in ContainerD than we had in Docker because we built, again, our cloud platform around Kubernetes and having ContainerD as this core runtime across all our platform, which we use Cloud Foundry, which has a container runtime. We have functions and service platform which now uses ContainerD.

Phil Estes: So it's become kind of this underpinning underneath all these layers of [inaudible 00:10:17], you know containers as a service. So yeah, that's kind of the history of how it came to be and why IBM decided that it made sense for us to be involved and why I continued as a maintainer there to be that connection point between IBM product and the open-source side of ContainerD development.

Brian Christner: I mean, ContainerD, I mean within IBM is becoming the standard, but also outside of IBM, I mean Google, Amazon, everybody is relying on ContainerD as the runtime. Is that correct? I mean, how do you see that going forward?

Phil Estes: Yeah, I mean I feel like our adoption has been phenomenal and the cool thing is because, and again I can't take credit for this, Michael Crosby, Stephen Day, Derek McGowan, they had lived through the entire Docker development lifetime and ContainerD was almost a chance to rethink a few pieces to make sure that the abstractions were really, really clean. And so now what's cool to see is it's not just about Docker using ContainerD or Kubernetes using ContainerD, but like AWS Firecracker or gVisor or you know all these kind of new ideas about container isolation. ContainerD just happens to be a perfect vehicle for bringing new kind of ideas around containers and isolation that don't necessarily have to live on top of Kubernetes or the Docker engine.

Phil Estes: So yeah, I think because of that you're seeing, you know, wide adoption. Alibaba Cloud, I like to point at them because they're using it like everywhere in their cloud. They've built their own Pouch Container open source project that sits on ContainerD, that's like a Swiss Army Knife of runtime and registry interactions. And so I think all these things clearly show that it's simple to use, very extensible, and people love how simple it is to start with ContainerD.

Brian Christner: I have to agree with that. I mean it's extremely easy to use, it's understandable, it's documented well. Now for next up for ContainerD, where do you see the direction heading for ContainerD?

Phil Estes: Yeah, you know I think we never want the scope to have this creep of becoming another huge monolithic engine that has every contraption. So we've tried to build in pluggability, we've resisted PRs that want to add a lot of new function directly in it because like the Firecracker team built all their functionality as plug-ins. Most of their code doesn't need to be in ContainerD, so that's the way we see ContainerD growing in functionality is not by us adding function but the pluggability and extensibility allow that outside of the core project.

Phil Estes: So really the core project, stability, performance, better Windows support, which the Microsoft team is working on for our next major release. Again making sure this runtime layer, it allows, not just runc, but all these other variants of Kata Containers and Nabla and gVisor, to have the best possible kind of support for how they use the platform. You know things like multiple containers per VM like we had to shift around the API to make sure that was well supported. But yeah, outside of that I don't see us having major functional additions other than making sure ContainerD stays ... The reason people like it, it maintains that simplicity and usability.

Brian Christner: I mean that's something that's very important in today's age is that not always do we need to keep adding features, sometimes just stable products is what we need. And so that's a brilliant way to go about it. And the last question I have about ContainerD is how can people contribute? So even if you're not a developer, I mean where do you recommend starting with ContainerD if someone wants to get involved?

Phil Estes: Yeah, so I think, you know a lot of the last few months have been busy with some releases and also because we maintain compatibility and release cycles. You know, something we never were able to pull off in Docker is having multiple lines of support, you know, bug fixes or being backported. So that keeps the maintainers fairly busy. So I think anyone who wants to come in and kind of start to look through ways the documentation may be lagging, code, that's always a huge area. We have a website that just has basic information that could be extended with a lot more examples, especially these plug points, like smart people from AWS just came in a figured it out, but I'd be great to have clear documentation. Like how would I add a plug-in to do this? So those are non-development areas that could always use extra hands.

Brian Christner: Absolutely. Now I want to kind of transition this into the next phase, is like conferences. Now you're kind of like a professional on the conference circuit and what're your tips for conferences? I mean, you attend a lot of conferences, you see a lot of them, so I mean, what can you ... And you're going to KubeCon next week, you're going to CERN. Do you have any tips for people and like submitting CFPs to conferences?

Phil Estes: Yeah, I mean CFPs, I've always found, you know just to be honest, I found to be tough because it's ... Especially if we're talking about at KubeCon level or even DockerCon, where there's significant contention over a number of slots that are small and a number of submissions that can be extremely large. And I've even been on review teams for DockerCon, for KubeCons, for other smaller conferences, and you know it can be overwhelming to try and think how do I pick the best talks because the numbers are so large, there are lots of great ideas. So I think some of the best insight that is not necessarily just from me, I've heard it from others as well, is because people are busy, reviewers especially, you have a sentence or two to grab their attention. So, not that you want to over promise, but you need to pack those first few sentences with like what's the real value you're going to give to people coming to this talk. Because a lot of people spend a lot of time kind of with backstory and it's like people just don't have the time to get to where you're going. So yeah.

Brian Christner: It's really like fire sale or as a resume, you know, you really want to catch somebody with a cover letter and just really pull them in.

Phil Estes: Yeah.Brian Christner: Now, since you're also going, what are some conference concepts that you really enjoy, that you've seen? Like open spaces or like at DockerCon we saw like a, what was it, they open space that you can submit talks to on the side. I mean, that's kind of a new concept that's taking over conferences. Do you see anything else grabbing attention?

Phil Estes: Yeah, so like you're saying, I think Hallway Track is an old term of just standing around in the hall that's been formalized, we've seen at conferences like a DockerCon. Which is really valuable. I mean, I think there's a slight bit of abuse of ... Like you see a lot of sales pitches being put into Hallway Track like, you know, come let me tell you about our product. Which is fine, I mean people can self select out of that obviously. I think the other thing, again, that is at DockerCon but I've seen other similar ideas around it is just connecting people because it's very interesting, like you said, when I attend a conference that's my community, so to speak, it's a ton of fun because I'm going to see a lot of people I know.

Phil Estes: When I go to a conference that's not necessarily ... You know I was just at CraftConf in Budapest last week, which is a huge cross-industry conference, and so it's not a bunch of container people. And it is a different feeling to walk in and like oh man I don't really know anybody, how do I connect? And thankfully if you're a speaker, sometimes there's a speaker event where you start to mingle. But anything that a conference can provide to connect people in, like, at DockerCon, it's very specific, the Pals program. I think that anything like that to try and help people that their company may have paid good money for them to be there and yeah if they feel disconnected they may just go back to their hotel and really miss out on connecting with people or listening to the talks. So I think that's area conferences, especially really large conferences, it's overwhelming for like a total newbie.

Brian Christner: I mean I feel that also, I mean the talks are amazing but the talks are always online so if you miss something you always go online, but I find the networking and talking to people actually building things, I mean you get just tons of value out of this.

Phil Estes: Yeah. And, you know, I guess being an introvert of sorts myself, I was never in to kind of walking the booths of like an expo hall, but I've learned, and I've sort of forced myself to learn, that's a great way to actually find out what's happening in the industry. Because people at those booths would love to tell you, you know maybe it's a product pitch in some sense, but it can be really valuable to kind of get a pulse for what people are ... If it's a container conference, what are different people doing with containers, what's the view on security, and what people think about the value of containers for industry such and such, you know finance or. So that's another way, it takes a kind of stepping may be outside your comfort zone at times but just strolling around an expo haul and connecting with people there.

Brian Christner: That's a great tip, thank you very much. Well, that's all the time we have for this episode. We really appreciate Phil coming on our first interview of The Byte and we wish him success at CERN and KubeCon. Any last words you want to tell us?

Phil Estes: No, it's sort of becoming a habit to stop in Zurich, sadly Brian always sees me right after an international flight when I'm still a little out of sorts, but it's always cool to be welcomed here and go off to some other beautiful place in Switzerland. But yeah, thanks for having me.

Brian Christner: Absolutely, thank you, Phil.
© 2020 TheByte