An interview with Thomas Di Giacomo about the state of Kubernetes

I recently attended KubeCon EU 2019 in Barcelona. While there, I got the chance to chat with our CTO, Thomas Di Giacomo, about the state of Kubernetes and its marketplace. We also talked about the community centered culture of Kubernetes as well as some his dreams for the tech world. It was a great conversation and I believe it highlights just why there is this incredible amount of hype surrounding Kubernetes:

 

Andrew Gracey:

We are here at KubeCon [EU] and there’s about 10 thousand people here which is significantly more than last year. As a CTO in the open source world, what do you see with this extraordinary growth?

Thomas Di Giacomo:

So, when we started yesterday with the CNCF Board meeting, it was interesting to see the growth of the number of projects in the CNCF. There were more and more projects, they were more and more mature as well. And with that come requirements to manage the projects properly. We had a great discussion with the TOC (Technical Commitee). And how to deal with that, how to structure better the CNCF Support for the projects.

We are also launching a new way to manage the SIGs (Special interest Groups) that might extend to the TOC later, as well. We are reviewing the plans for KubeCon and maybe doing more Kubernetes Days or smaller events.

AG:

I’m sure it’s hard to pull something like this off. With three of them a year.

TDG:

It takes time and is expensive. We plan to keep doing three each year but maybe with more space between.

We also want to reach out to smaller regions and other countries as well.

 

AG:

With this growth, are there any pendulums that you expect to be swinging back or any large changes that you expect in the near future?

TDG:

In the future, yes. Any open source project is like a product that has a lifecycle. We might not be at the top of the hype for CNCF it not really about Kubernetes but [surrounding] projects.

AG:

And there’s always that drop after the hype before it recovers…

TDG:

Then you go in to, what Gartner calls it, the “Trough of Disillusionment” and we are very conscious of the CNCF as well and are trying to anticipate the signs of that to see how we can avoid that.

AG:

If you look at the normal curve that you typically see, do you think we are past the early adopters? Do you think we are into the mainstream adoption with large companies using it in a very concrete way?

TDG:

It depends on where you look, and what type of industries and regions in the world. I think there’s still a lot of adoption coming. So, we are definitely not at the mass adoption level. The adoption is growing for cloud native technologies.

There are some open source projects that help bridge the current ways of doing IT while trying to benefit the from some of the more recent innovations. We need to bridge those two things as well, it’s not only about adopting the new stuff but also about making sure you can run the old stuff.

AG:

I think we are starting to see a lot of projects that are starting to pull in the older way of doing things.

TDG:

We talk about containers but there are a lot of VMs out there, and a lot of bare metal as well. These things need to work together in real life. I think that’s going to help the adoption of cloud native technologies. Companies like us are helping people to get there.

 

AG:

I’ve had it noted to me a few times that there are a lot of similarities between Kubernetes and the older methodologies such as Service Oriented Architecture, Mesos, etc… What differences are there this time around?

TDG:

So, it was actually mentioned in the keynote this morning.

AG:

Dang, I didn’t get to see it.

TDG:

This talk was about Simultaneous Invention. They took a few examples from history. So basically, what comes together with CNCF and Kubernetes is, first of all, it’s a good base technology, second is that the people are really working together. If you look at the scale of the contributors and the communities, that’s very important and that’s also making things happen. And we are also learning from what happened previously (with Linux, OpenStack, and other open source projects). We are learning from those experiences. When you look at CSI, CNI…  At the end of the day it’s all about the excitement.

AG:

The culture seems much healthier this time.

TDG:

It is, and it’s something that we need to work on because it’s not a given that it will stay like that forever. We need to always keep on recruiting people. Helping people to get started. It was great to see younger faces on stage from different background, different genders, and different countries this morning. There were two people from Finland and India explaining how they started with Kubernetes and the community. They explained their journey and it was very interesting.

 

AG:

What role does SUSE play in the larger ecosystem (and to some extent, the macro trends that we are seeing)?

TDG:

In a way we play the same role that we’ve been playing with our key open source projects such as Linux, Ceph for storage, OpenStack, and Cloud Foundry. We’ve been doing that for 26+ years now, so that’s one part of it.

I think we are also helping by contributing to those projects with things that matter for our customers and partners. And sometimes it’s not always the cool, sexy innovations that are important.

Sometimes, it about boring things to make sure you can easily deploy, upgrade, and manage the solution from a software deployment and operations standpoint. From different clouds as well. So that’s what we try to bring to the communities. We are contributing more and more to even low-level things like cri-o for instance. It has to be solid, stable, secure, and reliable so you can build the cool things on top.

And we do cool things as well.

 

AG:

For the last question, on a more personal note: What crazy dream do you have in tech that you’d like to see realized?

TDG:

I’ve been a developer a long time ago. I’ve been a sys admin. And I like the fact that those things are blurring. I like the fact that developers can benefit from functions and cloud services on-prem. I like the fact that hardware is still improving as well. We get a lot of innovation from the hardware that is sometimes hidden from things like KubeCon. We talk more to the developers.

Bringing those things together would be ideal. So, let’s say you want to do a cool app for, I don’t know, virtual reality or augmented reality. You could reuse some of the computer graphics engines and software application pieces from different places (cloud, libraries) but then also optimize automatically for the hardware that you have available or that you can use in the cloud.

So, let’s say linking the functions to the hardware that they can run best without having to decide, as a developer where to run things in the most optimize way.

AG:

It would be nice if it would see that you are pulling in libcuda and the system sees: “ok, I need a GPU”, right?

TDG:

You build your apps with different pieces of code. Some of them are your pieces, most of them are not your pieces of code and it just runs where they are supposed to run best.

AG:

That’s a really cool dream! Thank you for taking your time to chat.

 

 

Again, thank you Thomas for taking the time at a trade show where time is limited. I definitely enjoyed our chat and I hope the people reading this will as well!

 

By: Andrew Gracey first published at suse.com