In this episode of Telemetry Now, we're joined by a friend of the show, returning guest, and the cofounder of the network automation forum,
Scott Robohn, to talk about AutoCon 3.
Now AutoCon 3 took place just recently at the end of May twenty twenty five in Prague, And, we are recording this in the beginning of June, the very beginning of June, so that way we could, keep it timely, keep it relevant, and get it out to you right away. Now in this episode, we talk about AutoCon 3, some of the talks and workshops. We also talk about what is going on with network automation in the industry as a whole, and we talk about some upcoming events like AutoCon 4 coming this fall.
My name is Philip Gervasi, and this is Telemetry Now.
Now. Scott, welcome back to the podcast.
It really is nice to have you on again and see you again and talk to you again. It is an audio only podcast, so only I can see you, but everyone can hear you today, and we're really glad to have you back. You are a friend of the podcast as we were kinda joking about prior to hitting the old record button. So it is nice to have you on regularly and just kinda get these updates about what's going on in Scott's world. But in the world of, you know, automation and, network automation forum, which we're gonna dive into today, so, welcome.
And I gotta ask, how are you dealing with the, the jet lag now that you're back in the United States?
Well, I'll just say it's it's really good that this is audio only because I don't feel like my body has made it back to US eastern time, my home time zone.
I'm better. I'm still not quite there yet, but, could be could be worse. It's definitely a first world problem.
Yeah. Yeah. For sure. And for the audience, if you're not familiar with the network automation forum and with AutoCon, well, you're in for a treat today. The event was last week, AutoCon 3 in particular.
And, so we really wanted to get this particular episode out right away.
So we're recording today on a Wednesday. Gonna be you're probably listening on a Thursday or Friday, the week after after this last event in Prague. So, Scott, why don't we start off, with a quick level set on what the network automation forum is all about and what these AutoCon events are all about before we get into AutoCon 3 in particular.
Sure. The the the short story is the Network Automation Forum exists to advance the cause of network automation. That's the very short version. And if you think of the term forum, it's a place for people to get together and talk, and that's largely what our autocon events are.
Those talks are, you know, come in the form of formal presentations and lots of hallway track conversations.
We started in November twenty three with, you know, my my business partner, Chris Grundeman and I thinking we need to get more people talking about this. It can't just be our two opinions or two sets of opinions on this. And it's grown steadily with each of the subsequent events. We this, AutoCon 3 was our fourth event because every good engineer starts indexing at zero.
And, we're currently planning for AutoCon 4 in November, twenty twenty five here here in the states. Location to be announced, hopefully, very soon.
Great. Okay.
So on the heels of AutoCon 3, and you mentioned things like growth that you've seen, and, you know, I'm I'm assuming that part of that is the growth in numbers, but Yes.
I wanna really get and and that's that is important. I understand, because there is kind of like an energy that happens when you hit a critical mass, and I get that. And more more opportunities for a diverse discussions and different hallway tracks. Right?
Not just the the main the main line there. But, for for Otocom three in particular, I I'd like to hear about I I didn't get to attend, so I was at AutoCon 0 and AutoCon 2, both in Denver. Fantastic events. I'd like to hear about some of the highlights from AutoCon three from your perspective.
Now I have spoken to a couple of folks at Kentik and outside of Kentik that were at Autocom three. Loved it.
They they used adjectives like electric, and, actually, that was the only adjective I heard. But that's a good one. That is a good one.
And and that was describing the, the main talks, and it was describing the energy, just in the overall conference. But what were some of the highlights from you, from a bird's eye view?
And I know you were in the weeds dealing with stuff like, you know, Wi Fi and and things like that.
Well, the infrastructure stuff was a lot better this time. We'll talk about that later.
Okay.
Let's talk about growth first and not just numbers, and then some of my big takeaways. And that's like it's Scott Robohn's perspective. It's not a comprehensive readout. Yeah.
We'll address some of that later. But in terms of growth, you know, the numbers are tangible. Right? We we started with, you know, three hundred and fifty at the first event, all the way up to over six hundred at this most recent one.
And we're sizing for much more than that for AutoCon 4. But, in terms of skills growth and interest growth in the community, you know, we see, you know, more people asking for more technical content.
That's good. Right? That means that people are trying to actually do more, you know, beyond just writing, you know, a Python script. Mhmm. Looking more into, you know, how do I contribute to open source projects and automation and do more than just the basics. And that's great.
I'll even give a shout out to a gentleman named Francois Cain from Cisco.
Actually just had a conversation with him yesterday on Eric Chow's network automation nerds. That'll come out in July, where Francois said he's seen you know, he he puts, the network automation community into three buckets. You know, the intermediate the beginner, intermediate, and advanced.
And he's seen the intermediate bucket start to fill up and not just have the extremes. Okay. And I think that definitely matches what we've seen, with network automation forum. So there's interest in growing and and extending skill sets. And, one of the things we wanna be really careful with with, the whole Oticon event series is not leaving the introductory crowd behind. So we're trying to think through how do we bring more technically deep content, but still have stuff basically for everyone.
And, you know, as the, eighty song goes, it's tricky, to balance that. Sorry. I had to work that in here. No apologies needed ever. But, so, I mean, the short version of all that is we're definitely seeing numeric growth, and we're seeing growth in terms of the skills people are laying on, layering on for network automation. And I have to assume Yeah. Go ahead.
I have to assume that that means that the nature of the talks and the tracks and the workshops I wanna get into workshops shortly. That is evolving as well. I mean, if you're talking about a growing, segment of kind of the the intermediate crowd, practitioners, operators out there doing this stuff and and turning the cranks and the and the wrenches, you know, it's more than just intro to Python one zero one, but, like, how do I use this to solve this particular problem? And that I mean, to me, that's very compelling stuff. Is that kind of what you're seeing with the development of the, the, you know, the the the speakers, not the speakers themselves, but the, you know, the presentations and and the workshops and that sort of thing?
Absolutely.
Oh, great. Okay. We we definitely put focus on real world solutions, real things that people are doing, real use cases. We really aren't interested in being yet another forum for just advertising and just architecture.
Though there is some value in those things, obviously, in certain contexts, we just not the focus we want network automation forum to have.
But we see that in both the presentation content and the workshop content.
And I I've I've seen that grow and get better in both those areas with every event. You know, more and more people start to get it.
And and just a shout out to the workshop providers in particular, you know, Kentik, Justin Ryburn, you know, rode, you know, drove, one of the one of the sixteen workshops that was delivered. Mhmm. And we really coach workshop, folks to focus on real things. You know, again, it's it's a workshop.
It's not another, presentation opportunity. And I feel like the workshop providers are all stepping up to do that. By the way, they are a mix of vendors and individuals that are just interested in talking about certain things, and that's been really interesting to watch. So Yeah.
When you say talking about certain things, the workshops aren't, like you said, and not another presentation. So it's typically they're more hands on and I'm I'm I'm assuming I'm I did one at AutoCon too, so longer in scope as well. Right. And that's kind of the the whole, accommodating that operator crowd, that intermediate category where it's like, let's go deeper.
Correct. Yeah. Okay. Yeah.
And you always have to have stuff in your back pocket. Right?
Because Mhmm.
Even if you tag, you know, a particular workshop or session as, you know, introductory, intermediate, or advanced, you're always gonna have a range of people in the seats. And so, you know, one of the Yeah. Old old, instructor tricks that I have is, you know, have some advanced exercises that you can hand out to the people that keep finishing the labs and sitting there twiddling their thumbs.
You know, it's always it's always good to, like, oh, okay. You've been able to do this. Now can you try this? Right.
You know, and or leverage them to, you know, help other people in the workshop. You've done it. So you you know, and you've been an instructor. Right?
So you you you understand all those classroom dynamics for sure.
Well, I've been an instructor for sure, but I've also done, a workshop at AutoCon too, as I just mentioned a moment a moment ago. And, my takeaway from that is that workshops are varied. Just walking the halls and seeing the other workshops, obviously, the one that we did was on, the fundamentals of of observability, network observability in particular.
But looking around, they they were varied. It wasn't just purely like Python and Ansible and, you know, an Autobot and and Netbot and those kind of things. It was it was much more varied. So one of the things I wanted to ask you was, and and there's no hype. There's no market texture here. It's a straight question.
What are you seeing in terms of, people talking about AI, considering that a lot of what happens in AI engineering, now this is me with my opinion Sure.
Is more software development than actual AI because you're using maybe a foundational model. You're not building your own models and training your own models. You're you're utilizing something that's built, and then building the software workflows around it. So I see them go hand in hand very, very much.
So I'll say the interest in AI is increasing.
Okay. You know, my you know, over the life of the Oticon series, I was hoping for a lot more AI related content in fall twenty twenty three.
And there wasn't much because I don't think people were doing much in production at that point. Yeah. You know, that was when all the hype, you know, the recent release of Chad GPT, you know, kind of lit the world on fire and started the arms race with, you know, Grock and Gemini and everything else that's come to anthropic, etcetera.
Mhmm. Mhmm.
But I've seen I've seen interest in content increasing. And I'll point people to one particular talk, from Javier Antique of of Cisco, who talked about the concept of getting from automation to autonomy.
And, I can't I can't do his talk justice, but once the information oh, it's you know, the the video will be published, you know, give it give it four to six weeks from now.
That's one of the highly recommended talks, of things that he's, you know, talking about the difference between what does it mean to instantiate an automation versus, you know, what I'm calling protocol like behavior, and real autonomy? I think that's much further down the road, but people are starting to ask those questions. And, you know, I I think the easiest example I can give you is, you know, logical circuit establishment, whether it's s s seven in the telephone network, ATM, SVCs for those who remember those days, or MPLS, you know, LSPs. Right? Whether you're using LDP or RSVP to signal them. You know, those things are automatic.
You don't have to go through and configure seventeen routers to set up an LSP.
You kinda defeats the purpose. Right?
Autonomy, I think, gets us closer to that protocol like behavior that I think is an an interesting and useful future. Yeah.
Yeah. And and you saw that starting to emerge, not not the protocol or the the technology per se, but on Javier's talk and and I assume others, did you see that starting to emerge at AutoCon 3? I I do remember at at AutoCon two in Denver, there was a reluctance on the part of many engineers to you know, there there wasn't a lot of buy in just yet. There was kind of an openness to hear, and that's kinda where it was for me.
And and and by the way, I don't I don't blame folks for that. You know? There's marketing departments that come out with stuff and and sales folks that promise things that break the laws of physics. And and as engineers, we're literally building the stuff that that makes the, you know, the packets go.
So we are justifiably skeptical and hopefully in a healthy way where we're listening Yeah. Willing to listen and all that. And that's I think that's where it was at AutoCon two, but sounds like AutoCon 3, it's starting to open up a little more.
I would agree with that assessment. And, you know, we've seen this movie before.
You know, think about cloud adoption ten, twelve years ago, fourteen years ago. Me put my data and my workloads in someone else's physical infrastructure.
And I especially remember that, you know, with, government related work that I used to do and that fear of, you know, putting things into public cloud infrastructure.
But then, like, the first use case that really hit me was rendering imagery from Mars missions. You don't wanna maintain all your own private infrastructure to serve out all that imagery. That's a burst into the cloud use case that made a lot of sense. You know, we're starting to see similar things emerge, you know, in the AI realm with agents. And if you you and I have had a conversation about this, but one of the things that clicked with me is agents have agency.
They can actually do things.
Right? And we're gonna see more, you know, small useful you can think of them as little bots, that go do very specific tasks, but there need to be coordination and orchestration and routing of messages between those agents to actually get bigger tasks done. Yeah. People are starting to think about that in the NetOps context.
Well, that's why I bring it up because the conversation is not about necessarily training models and building a new model of some sort, whether it's a large language model or some other type of model.
You know, we we've actually been talking about some computer vision, applications recently with looking at circuits and and, quality control and that kind of thing. But, it seems like the the relevance in the discussion is around the software development life cycle that's that's around that surrounds this entire thing. And, okay, now that we're incorporating this technology into network operations or or perhaps IT operations more broadly, and, certainly, they're already using it in security operations and to one extent or another.
You know, how do we write the software and the and the workflows that that get this done? And Right. And then, you know, secure it and have observability into it and all of that. And that's why I see it as such such a, you know, the network automation piece, as as the foundational pillar, a cornerstone of what we do with AI and network operations. And, therefore, it was just very logical to me that it was that it was inevitable to me, actually, that it was gonna be a a thing at AutoCon, and now, I'm hearing, much much more so at AutoCon 3.
So Is that that that brings me to a Yeah. There's another interesting follow-up to what you just said, which is one of my big takeaways, from Prague.
Okay.
There is a emerging and maturing conversation on what it means to acquire additional skills, especially software development related, software engineering related skills for network engineers.
And, you know, we've kinda we've all talked about this historic, you know, I don't wanna be a developer. I'm a network engineer. Darn it.
I would do my best. Doctor McCoy here. But I I can't do either version of him very well.
And, you know, there's some identity issues that we need to really work through. Right?
I I will point people again to when the, when the recordings are published. Claudia Del Luna's opening keynote is really, really interesting on this.
I'm not saying we're exactly in the same place from an opinion perspective, but it really got me thinking in terms of, you know, you can acquire software engineering skills as a as a network engineer.
There's nothing stopping you from doing it. It doesn't mean you have to or everybody has to. But, you can go find other things to help you do your job, whether that's just helping you with logical thinking, understanding of the tooling, like using Github or GitLab or other repositories.
To really help you, you know, move forward, you know, with the future of of networking.
And one of the big things that hit me during lots of conversations last week is this really comes back to being what does it mean to be a well rounded engineer?
Not just a well rounded network engineer.
Thinking in multidisciplinary terms.
You know, you've got, great other pitches on, submarine cables and optical infrastructure at the bottom of the ocean. Right? You know, you've you've talked. And that's definitely something that you've added on to from, like, understanding what you know, those conditions are very different from, you know, cabling in nice, neat racks in a data center.
And there are different physics involved with Well, we hope they're nice, neat racks.
But right. I gotcha. Yeah.
So I guess to to really sum that up and not consume the rest of this discussion with this, you know, we live in an environment where you have the opportunity to learn about anything you want to in ways that have been never been easier, in the history of humanity.
And, I would encourage people to think outside their swim lanes to be educated about certain things. I know I know you have to manage the tension between specialization and generalization.
But, you know, walls that you think are there may only exist in your mind. There's nothing stopping you from going and layering on new skills to help you better be a Mhmm. Fill in the blank engineer.
I remember, around twenty fifteen, twenty sixteen when the term network automation became very commonplace, and that was a a topic of conversation, of podcast, and blog post very, very much so. Right?
And and though the some of those practices and ideas had existed for decades prior, you know, I mean and even even to an extent on box scripting and doing some Perl scripts before Python became primary, all of that stuff existed. But then really just taking a forefront, taking this the the, you know, the main stage for networking did happen maybe about ten years ago. And I remember at AutoCon Zero, the question was, why aren't we farther along with network automation? What's the deal?
So to address your point about being a more well rounded engineer, the analogy, I guess, I use, and I've used for a long time now, is this idea that if you are, let's say, an astrophysicist, you have a PhD in astrophysics, and you are specializing and you specialize in exoplanet, like, discovery and analysis.
Right. Eighty percent of your day is writing code to crunch that awesome space data that you've ingested from, like, you know, whatever telescopes and radio telescopes and and and things like that that you're using.
And does that make you a programmer? Well, not really. A little bit, but no.
But it's just another tool for you to to get the job done. And, sure, you have those, like, pure programmers that that specialize in writing code, and and that's where I would say, Scott, that's where your you know, does does everybody have to be a software developer or an, you know, software engineer? Well, if that's actually your specific thing, yes.
Sure.
But by and large, we use this as a technology, as a tool in the tool belt of all sorts of professions, some of which have nothing to do with, you know, our world in tech and IT and networking. Yep. And that's how I've likened it, for the past decade.
Yeah.
I would I would point people to a book called The Cuckoo's Egg by Cliff Stoll, where I believe Stoll was a PhD astrophysics student Oh, really?
Okay. At some major university. Uh-huh. And he started investigating what looked like rounding errors in charges on, mainframe access via dial up modem.
And that sent him on a path of really understanding how how dial up access worked, and found that they were actually being hacked.
So Oh, interesting. You know, student in one domain. I don't know if exoplanets were his specialty or or what. Right. But he actually found found and built additional skills, you know, just just pretty laid out. It's a, sorry, it's on my list to reread, and, it's been on my mind, so I just thought I'd throw that out there as a good example.
I'll say that when I was, because as you know, I spent most of my my career prior to getting into observability and now AIML and stuff like that, as a VAR engineer.
And, I feel like the main skill that I acquired, I never expected to acquire was, like, being an expert label maker using those, like, p touch label makers Sure. And figuring those out. You never know. But, are there any specific highlights that you feel comfortable talking about, of course, since they're the videos haven't been released yet?
Right. You know, maybe a workshop that you sat in on for a few minutes or a or a talk or a side hall conversation you wanna mention. Are there any specific highlights that you would like to tell us about from AutoCon 3 specifically? I'd I'd really like to know how how it was for you.
It's really hard for me to pick favorites.
You know, my attention is divided during an event. Yeah. You know, because I'm paying attention to all the operation stuff. Mhmm. I will say that I felt like this one has been the smoothest so far, and I did I had a personal goal of making sure I spent more time talking to attendees while I was there, and I was able to at least partially accomplish that goal. I feel like that went much better, than Oticon two, than Oticon one, etcetera.
Great to hear.
But other other content, you know, on the workshops, you know, we look to creative proposals from individuals and companies to bring stuff to the table.
And I was tremendously pleased with every proposal and how how folks stepped up.
Their their new entrance, that did this for the first time, you know, Glueware was one of the companies that, that did this for the first time. I had a lot of back and forth with them. They were, you know, really open with, how they were developing their materials. That was great to see.
I'm not disappointed in any of them. And that that that sounds like very weak praise, and I actually mean it much stronger than that. I I feel like we came up with a strong slate of workshops.
Those are not recorded. If you wanna if you wanna get in on that action, come to Oticon for, you know, in November twenty twenty five. We'll have a new slate of, of workshops for you there. In terms of other content, there are two other ones I'd really love to just put a a little bit of emphasis on.
And they were all good. Don't get me wrong. I feel I feel a little guilty picking, out, you know, anyone's in particular like I already have.
But, John Howard of Proton did a lightning talk on, exposing, secrets. And I I mean, you know, private keys and related issues. And with all the security concerns that we have in networking period and now with the, increasing use of agentic AI mechanisms and MCP servers, I think security needs to be talked about, talked about, talked about, at much higher volume and frequency than we typically do. So it's great to see, some of that deep security content, you know, be presented in Prague.
And by the way, he had a thirty minute talk proposal that we made him scrunched down into ten minutes. On the one hand, that's just not fair. On the other hand, he did a great job with it. Highly recommend viewing that.
What what was last again, Scott?
John Howard John Howard. Okay. Of Proton. And I believe the talk was stop exposing your secrets.
And that would be on the lightning talk on on Fridays, from the event. The other was the closing keynote by Jason Edelman of network to code.
You know, there's we obviously run-in circles where we run into a lot of folks who have done a lot of network automation.
But he is one of the individuals that has been at the epicenter of so many advances in network automation.
Yeah. For sure.
I won't use superlatives.
I won't put him at the the very top summit of the mountain, but he's one of a small group of people who, like, no no one else has done more than seen more than touch more than folks like Jason. And he gave a great historical perspective, in his closing keynote And, you know, pointed things to the future. And I'll I'll leave it to, the listener to go watch the presentation when it's published in a few weeks. But those are the highlights for me. Again, one person's perspective. And, it was it it was great. I can't speak highly enough of how well the event went and how many hands came together to make this work and work well.
Great. Great. And, so that was Prague.
And we have AutoCon 4 coming up in November, location TBD or TBA, one or the other.
TBA. TBA. Are there any changes? Oh, so logistically, the way it works is for the audience's sake, the first couple of days, are workshops. So that's two days of workshops, and then the workshops wrap up, and then the main conference, commences, and that's the next two days. So if you're doing both workshops and the main conference, it's four days, or you can do, I guess, one or the other.
Highly recommend you do all of it and and just really, immerse yourself into that world.
And then the the main conference itself is really a series of of of talks from the stage, the main stage. Is that right? Correct. Okay.
And so And it is it is if you do both, it's a full five days. And the way we Okay.
Five.
The way we do the conference content, we give you we give you Wednesday morning as unstructured time.
It's been a great time for people to start gathering and talking.
And then conference runs Wednesday afternoon, all day Thursday, and Friday morning because by lunchtime on Friday, people are saturated, especially if they've already been at the workshops on Monday and Tuesday.
Well, I mean, even during over the course of those five days, there's other stuff going on. You know this, Scott. There's the birds of a feather gatherings that are happening Right. Talking about all sorts of individual topics and and things like that. And so, so there is actually a lot of opportunity to connect with folks on very, you know, specific issues or just to connect with folks that are like minded. I'm I'm I'm I don't know if there's groups that go running in the morning, but I bet you that there's that too. I don't run.
There is a There's a boff dash runners on Slack.
There you go.
Yep. Perfect. So No boff dash lifters, though. Sorry.
I'm so go go ahead.
Say it again.
So no no boff dash lifters. Sorry.
You can start that if Yeah.
Well, we are a solitary crowd. I don't like to lift with others. Headphones in. Leave me alone focused on the weight.
Now with the with AutoCon 4 coming up later in the fall, are there any changes that we can expect, or is this really working and, we're just gonna look forward to an awesome, a little bit increase in growth and some amazing presentations again?
There are a couple of things we're very interested in layering in. And what I'm about to say are not commitments, but the things that are under active consideration.
Perfect.
There has been interest in having some kind of hackathon activity.
We are exploring that. That's gonna need to be a sponsored activity because you don't charge people to attend a hackathon.
So if anybody is listening, and a spot from a sponsor perspective, there will be specific sponsorship opportunities for hackathon. If you're interested, please reach out to me. I'd very much like to talk to you.
There are active conversations on what would happen during the hackathon.
And a common theme there would be contributing to open source projects, via a hackathon in in network automation. That seems like a really useful goal for a hackathon activity.
That's item number one. Item number two is that we do have, enough room to have a separate track on Thursday all day during the evening. Okay. We are thinking through how we want to handle that.
And there are two main ideas that that could be used for. One would be something that focus on leadership, being able to, you know, do cost justification for network automation projects, things along those lines.
And then more deep technical content is another candidate for using, that Thursday track. We could even split it up and do half a day on business and leadership and half a day on deep cuts Yeah. As it were.
Deep cuts. Right. That's a great idea. That's a great idea.
And and, you know, we're we're thinking it through, and I'm I'm I'm taking a little risk here and exposing that thinking to everyone. You know, just don't take it as a commitment.
Again, if you have thoughts on it and wanna provide feedback, you can, you contact me, or Chris Grundeman, and, we'll take it under consideration.
That's fantastic. I like that a lot. So we do need to wrap it up now. But before we go, can you just share how folks can find the network automation form information online and and then, of course, information about AutoCon three and all that?
Go to networkautomation.forum.
Yes, dot forum is a real TLD. Mhmm. Network automation dot forum will get you there.
That has information, about joining Slack. Our Slack is wide open and is fairly engaging.
You don't have to have attended an auto con to participate in Slack. I also recommend you watch our LinkedIn group. There's pretty regular announcements. And Slack and LinkedIn are where, announcements will go for AutoCon 4. That'll be those will be the first places.
Okay. And if I can add, I would also, to our audience, check out the Network Automation Forum's YouTube channel. I have gone there many a time and watched, presentations from various autocons, sometimes multiple times, just because I was learning something, and I'm like, slow down, professor. I gotta write that down.
So a lot of great information there. Scott, thanks so much for joining me today. And to our audience, if you have an idea for a show or if you'd like to be a guest on Telemetry Now, I would like to hear from you, please reach out at telemetrynow@kentik.com. So for now, thanks so much for listening.
Bye bye.