In this episode of Open at Intel, host Katherine Druckman and SUSE Domain Solution Architect Mark Abrams discuss how to get the most out of Kubernetes for edge solutions, the evolution of the open source community, and the importance of contributing to open source projects. They also touch on the complexities and opportunities in cloud native technologies, the impact of AI, and future developments in edge computing and the open source ecosystem. Enjoy this transcript of their conversation recorded at KubeCon in Salt Lake City last fall.
“The interesting thing about Kubernetes is that it has really solved the problem of failover resiliency. It's high availability, reliability, and scalability. Fundamentally that's what it brings. But in using Kubernetes and the Kubernetes API and cloud native methodologies, we have building blocks. We have LEGOs. Now we're starting to see more of a manufacturing process that allows us to integrate other things more quickly and more easily.”
— SUSE Domain Solution Architect Mark Abrams
Katherine Druckman: Hi, Mark. Thank you so much for joining me in the middle of KubeCon, which is always intense. It's a very busy and packed conference, full of people and sessions, and a lot of things to do, so I appreciate you taking some time with me.
Mark Abrams: I'm excited to be here, and it is interesting to have everybody around while recording.
Katherine Druckman: You can't tell, but there were people walking by us, and we're in a fishbowl of sorts, a little box…
Mark Abrams: A podcast bubble.
Katherine Druckman: Yes, a transparent box, and we can see all the people waiting in line for coffee. That's what is really happening right now.
Mark Abrams: That's right.
Mark's Role and Interests at KubeCon
Katherine Druckman: Mark, tell me why you're here. I know we're going to get into several things, including a book, spoiler alert there, but tell me why you're here. What are you doing at KubeCon?
Mark Abrams: I am here to demo some of the edge solutions that we have at SUSE. I am a domain solution architect with SUSE, and I came to SUSE through the Rancher Labs acquisition. I've always just been interested in little devices. That led me to putting Kubernetes on Raspberry Pi, and I just became the edge guy at Rancher Labs. Really, I'm field engineer, so I go out and work with customers who are interested in learning about the products. And now at SUSE, it's pretty much the same—I just explain the technology, and help them decide if it's the right fit for what they do.
Katherine Druckman: Very cool. I also have an affinity for tiny little devices, so I understand. Do you have a drawer full of single-board, tiny pooters?
Mark Abrams: Oh my gosh, I have a room full.
Katherine Druckman: A room full? Okay. Well, I have a drawer full, but many of them are not used. In fact, I bought a bunch of Pi Zeros back in the day and hardly used them, but they were so cute I couldn't resist. That's embarrassing to admit, but here we are. And they were $5 or something.
Mark Abrams: That's right.
Katherine Druckman: How could I not buy a few of those?
Mark Abrams: Exactly.
Katherine Druckman: Thank you, again, for bringing me a copy of your new book. I wonder if you could tell us a little bit about that experience. I'm really intrigued.
The New Book: Cloud Native Edge Essentials
Mark Abrams: The title is Cloud Native Edge Essentials. We have been finding that ‘edge’ is a funny term, because it means different things to different people. It's really vast, and it covers everything from this Telco edge, to retail edge, to industrial factories, and even mobile units. The military is using it, it's on satellites, and all those use cases have some commonalities, but they also have some differences. The team at SUSE, we have an Edge team, and we have some products that have been bringing cloud native to the edge computers. Edge has been really good at creating operational technology, but they have some issues when it comes to updates, management upgrades, issues when a software application fails. They don't have IT in the fields, and they have challenges—how do you solve the problems? It turns out that Kubernetes is a really good solution to that and that the resiliency of cloud native works really well in the edge space for many of the solutions out there.
The Evolution of Kubernetes and Cloud Native
Katherine Druckman: I've been thinking a lot lately. This is the 10-year anniversary of the Kubernetes project, and I've been thinking about the way that we work has changed so much. I wonder if you could talk a little bit about that. How life-changing has your involvement in the cloud native community and working with cloud native projects, how much has that changed the way that you do your work compared to even five to seven years ago, when Kubernetes was kind of a fledgling project?
Mark Abrams: I've been working with Kubernetes since 2015, it was probably the first time I installed Kubernetes, and it was really challenging. And now we have distributions that make it really easy to get Kubernetes up and running. It has changed the way I work. I've been a software engineer, I've been a solution architect, and now in the field engineering space, and really it's enabled me personally to help customers better. There are issues that we've always had that cloud native solves, and so on the one hand, it's an incomplete solution, and that's why we have so many new companies always popping up in the cloud native space, and the ecosystem is so giant, because we haven't solved it all yet. And yet, we're finding that it's streamlining things and making life simpler. Of course, there's complexity that comes along with it, and new challenges.
Katherine Druckman: Sure. Complexity is a big one for me. The more discrete functionality we add, the more we develop in silos, and the more we... Division of labor, right. The more do, we divide our processes into discrete units that work together, the more complex it becomes and the more dependencies you have, the more, I want to say potential points of failure, even. The more security concerns and all of that. And I wonder how you respond to those challenges, especially in edge computing and in tiny pooter devices.
Mark Abrams: The interesting thing about Kubernetes is that it has really solved the problem of failover resiliency. It's high availability, reliability, and scalability. Fundamentally that's what it brings. But in using Kubernetes and the Kubernetes API and cloud native methodologies, we have building blocks. We have LEGOs. Now we're starting to see more of a manufacturing process that allows us to integrate other things more quickly and more easily. As we find this consistency in interfaces, we're also finding that it's easier and faster to build things together. It's funny, years ago, Fred Brooks wrote a book Mythical Man Month, and he talks about when you add more developers to a project, the project actually gets delayed longer because the amount of communication goes up, and for three people, there's three channels, but for four people, suddenly I think it's seven channels. All of a sudden one person adds exponentially more communication…
Katherine Druckman: Chaos.
Mark Abrams: And that's what's happening with cloud native, right? Oh, I need storage, I need networking, I need backups…
Katherine Druckman: 208 CNCF projects!
Mark Abrams: This CNCF landscape, I was talking to some of the maintainers and they're trying to figure out how to make it more approachable.
Open Source Community and Contributions
Katherine Druckman: I was involved in a little bit of that yesterday doing a small part to try and make it a little bit more digestible. I wanted to pivot a little bit. I've been around open source for a long time. I understand software as a whole, but I'm very much deeply rooted in open source. What attracted you to work in the open source community with open source software? Why open? Why is that important to you and to SUSE?
Mark Abrams: I think for SUSE, it really started as a community, and the open source community and building the community. Personally, it started similar to why I got into little devices is that I was a poor college student and there was this free operating system called Linux, and I could do things like remote into the school computers to do my programming work…
Katherine Druckman: Without having to hoof it to the computer lab.
Mark Abrams: Exactly.
Katherine Druckman: Kids today, they don't know what we went through. They don't know. Remember when... Well, actually a lot of people listening may not remember—we used to have to buy software with actual money in boxes off of shelves in stores.
Mark Abrams: That's right.
Katherine Druckman: That is a true story. It's so weird.
Mark Abrams: It's so weird. I was attracted to the freeness of it, and the open source became interesting because it was accessible. And then I started a company back in 2004, I was using open source technologies. I was doing some stuff with VNC—virtual network computing—and creating this remote online space where people could talk to each other, but I couldn't find investors that would invest in open source. And this was back in 2004.
Katherine Druckman: And it was unknown. You still had to prove legitimacy. So strange, right?
Mark Abrams: Exactly. I took this circuitous route to landing at Rancher Labs where they had open source products, and it was a part of my DNA that I had to leave in order to make a living and being able to come back to it and have this environment, to have, I don't know if there's 10,000 people here at KubeCon this week…
Katherine Druckman: We’re pretty close.
Mark Abrams: That we are all going in the same direction. And open source has just come so far. It's really exciting.
Challenges and Solutions in Edge Computing
Katherine Druckman: Something we just mentioned, you used to have to really legitimize the use of open source software. It was not trusted—maybe a while ago, but not that long ago. And how has that conversation changed? Because I feel while we've come a long way, there is still weirdly enough, as ubiquitous as open source is, there's still a little bit of that conversation to be had here and there.
Mark Abrams: There is. And at the same time, you have parts of the government that are mandating open source because they don't have these incidents where some proprietary software had bugs or security holes.
Katherine Druckman: And there's also, I think, relevant to that conversation, right about legitimacy and perspectives on open source, there's also a conversation to be had about contribution to open source. When things are good and budgets are flush, you don't have to work so hard to get paid time to contribute to open source projects. And when things are less good, that's more of a struggle. We have a lot of projects that a lot of us rely on for our day to day. And I know again, SUSE is one that contributes heavily. And why is that so important and how do you continue to justify, but also encourage and nurture that kind of investment in open source communities and projects?
Mark Abrams: I think it's really, open source has been this grassroots growth. When I shut down my company, I still wanted to put it into open source. I put the code into, it was subversion, it was pre-GitHub. I forget what the repository was, but it was an open online repository. I put it in there, but it still has all the IP headers that are proprietary. I went to the local university and tried to rally students at an open source festival to join me and help me convert this product to an open source product. And that's the thing that I feel the community has developed. I had a hard time doing it myself, but now we have this community that's out there, evangelists. There are so many ways to contribute besides code.
Katherine Druckman: And frequently, I think it's the harder part, dealing with the human problems, the communication and miscommunication is very challenging, but yet really important to the health of any kind of project.
Mark Abrams: It is. And the challenge is that the things that grow, the things that get nurtured, aren't necessarily the things that you chose to use in your product.
Katherine Druckman: Yeah, there you go.
Mark Abrams: And that's where I think, how do we ensure the life of these things? I think that enterprises are starting to find ways to contribute to projects. So sometimes it's the acquisition of the company. At SUSE, we've acquired two proprietary closed source solutions and put them into open source. One was a security product, and the other was an observability product. They are now open source products.
Katherine Druckman: Which products are they?
Mark Abrams: One of them was New Vector was originally closed source and the other one was StackState. We just have the ethos, but rather than rely on the community a hundred percent, we have the viability financially to make more of it. Now we're trying to get community members to contribute as well, other people that are using it. But we depend on those projects for the ecosystem that we're producing. That's one way to do it. There are other ways to do it as well sometimes, and obviously we have contributors to the Linux kernel, but documentation…
Katherine Druckman: Sure. All of that stuff.
Mark Abrams: Evangelism, being here, becoming a cloud native ambassador, those are all important roles because we need to continue to grassroots grow and figure out what's important and make some consistency.
Future of Edge and AI Integration
Katherine Druckman: What are you looking forward to most in the coming months and year as you continue your work on the edge?
Mark Abrams: I have a little bit of a mechanic personality as opposed to an engineer personality. My title is engineer, but I really like to take things apart and build things. And I do 3D printing. I have some projects that highlight what SUSE does, but they would be really physical, tangible things. A couple of years ago, I built a rock, paper, scissor game where we had a robotic arm and the computer would play the robotic arm side. We had to use AI to program the computer to detect a human play. And then we had the robotic arm and it was just randomized. I've been talking to some of the enterprises, some of the hardware vendors about using their hardware to bring that to the next level and do some things that would highlight AI or that would highlight manufacturing. I'm really looking forward to bringing this more to the edge, the industrial edge, and building things that are interesting to show people just how it works and get them to say, "What is that and why?"
Katherine Druckman: Yeah, I can't help myself. Since you mentioned AI and everybody's very excited, how many, I don't know if you were in the keynotes, but many of them were focused on AI, building and deploying AI applications. How do you think openness plays a role in that emerging AI landscape? Because to me, it's important, but, and I don't know if you've followed the conversations around the AI definition and the open source AI definition.
Mark Abrams: I think AI…it's extremely important that we start to really open up what's going on in AI. I think there's a sense of ‘we don't know what's behind the black box.’ LLMs, they…
Katherine Druckman: It's a great mystery.
Mark Abrams: It's a great mystery. How does it work? And it is funny because I've read some blogs from data scientists that are complaining about how AI disrupted their role, but we really need those data scientists because they're the ones that actually understand how these neural networks work and how to build them. But it's a whole new level of interaction with data as opposed to the computer language processing that we've been doing to program things before. Openness is absolutely critical because of the chance for bias. I mean, I've also seen and read some books about the bias that's getting built into these systems.
Katherine Druckman: You want to know that the things that you're building make the world a better place and certainly not a worse one, right?
Mark Abrams: And at least understand the bias when you are getting responses from AI, what are the potential biases that it has so you can…
Katherine Druckman: Or poisoning data sets, right? And generative AI generating potentially harmful responses, for example. I think there's a conversation to be had about documentable reproducibility too, transparency, because a lot of the applications of especially generative AI are going to be in fields that are heavily regulated and require a tremendous amount of documentation about reproducibility and about the steps that you took to get from point A to point B ultimately. And I think openness plays a part there. I'd like to think it does.
Conclusion and Final Thoughts
Mark Abrams: I think there's also the other component. One is that the thing that we're communicating with, let's say an LLM, but there are obviously other backends to generative AI. On the one hand right now, if you use something like ChatGPT, they are taking your queries and the things that you input to ChatGPT and making that a part of their data sets. And at SUSE, we have a stack now that privatizes your ability to communicate with an LLM and you can build RAGs against it. But enterprises have a concern about their IP becoming public. They don't really want their employees just going out talking to ChatGPT. That's another aspect. But of course, our source is all open, from whatever we produce, it's open, but then you get to Ollama or whatever LLM you choose, and how open is that? I've never tried to look into Ollama, but that's where the black box comes back in.
Katherine Druckman: There's so much to look forward to, I think in the next year. I think we're going to have some very interesting conversations this time next year.
Mark Abrams: I agree. This is going to be very interesting to see how we go forward.
Katherine Druckman: Well, thank you so much. Thank you again for joining me and taking us through your story.
Mark Abrams: It was a pleasure to be here. Thanks for having me, Katherine.
Katherine Druckman: You've been listening to Open at Intel. Be sure to check out more about Intel’s work in the open source community at Open.Intel, on X, or on LinkedIn. We hope you join us again next time to geek out about open source.
About the Guest
Mark Abrams, Domain Solutions Architect, SUSE
Mark Abrams has been involved in developing and delivering technology solutions for over 25 years. Mark has broad experience ranging from writing code for backend services, embedded systems, and user interfaces to managing and building technical teams and field activities around pre-sales engineering. Mark founded and led a technology enterprise using distributed methodologies before the modern day cloud existed. Mark was a part of the original team that brought k3s—the lightweight Kubernetes—to market. Mark is currently a proud member of the Domain Solutions Architect's team at SUSE.
About the Host
Katherine Druckman, Open Source Security Evangelist, Intel
Katherine Druckman, an Intel open source security evangelist, hosts the podcasts Open at Intel, Reality 2.0, and FLOSS Weekly. A security and privacy advocate, software engineer, and former digital director of Linux Journal, she's a long-time champion of open source and open standards. She is a software engineer and content creator with over a decade of experience in engineering, content strategy, product management, user experience, and technology evangelism. Find her on LinkedIn.