Building a Community Around Network Automation with NAF

I'm Scott Robe on with Network Automation Forum. It is pronounced like putting a robe on it works. Um, I'm really happy to talk to you today about network automation forum and the community development or community emergence that we've seen around network automation.

Let me explain that to you. So we're all here today as networkers. We all have some interest in networking technologies, right?

It's not monolithic. There's a lot buried behind that cloud. And we all participate in different organizations like Tech Field Day, like packet pushers, like nano, like the local nugs and the local nugs like network automation forum.

We all have different foci. You know, we have our own autonomous systems, if you will, but we can collaborate and communicate via peering relationships. Yes, I'm intentionally ripping off BGP as a model, uh, for community organization 'cause it really works, right?

We all have our different things that we wanna focus on in our different contexts, but finding ways to cooperate and collaborate are really useful. Network automation in particular isn't pigeonholed into any one of these organizations. Um, everybody has some interest in automation, you know, and all sorts of these different organizations.

Uh, we found some real utility on putting laser focus on network automation. And I'll tell you how that came together. A network automation forum is a forum for anybody who has an interest in advancing network automation.

Um, whether you're an end user, you work for an MSP, you work for a vendor, um, you're involved in the community in some way. If you're interested in advancing the cause of network automation, there's a seat at the table for you. And we intentionally use, um, you know, the forum, uh, concept here.

It's a place to come together and talk. We also like to talk about it as a watering hole, right? Um, Chris Gruman and myself, the co-founders of NAF we're kind of the janitors around the watering hole to make sure the water's flowing.

There's clean seats, um, and there's enough stuff to talk about. Um, but it's not, it's not just us. Obviously there's the broader community, um, that's been growing.

I'll give you some stats on that as we go. Um, the organization is really Chris and myself plus some additional part-time help. Um, if Emily Hilbert is out there, she's our Kayla Anthony, and she does so much legwork for us.

Um, helps us organize these auto con events. And, uh, because I spent a lot of time at Juniper, the first event was auto con zero. Um, 'cause we all know indexing starts at zero.

Um, and we're about to hold auto con three in May. I'll give you some details about that as we go. I'll give you some idea of the collaboration activities that happen in the context and talk specifically about the role of sponsors.

How this all started. Um, in late 2022, I was getting ready to go out and do my own independent consulting work. So I basically started my company tech sales craft in January 23.

But I started hearing about the state of network automation survey that Chris Grundman was running and he picked up from other people that had started it. And when I heard about the survey, I said, but I'd really like to read the results of that survey. And then I said, scratch that.

I'd really like to talk to Chris. So Chris and I got connected through colleagues and we just started talking about what's the holdup? Why haven't we seen greater adoption of network automation?

And there's more than one answer to that. I'll give you some answers here. Um, but we decided, let's focus on the question.

Why haven't we seen the full adoption of network automation? It's a very useful driving question. Um, we started talking with a bunch of our friends saying, is this useful?

Is this a useful discussion? Would you be interested? And we decided, you know, we're just two people.

Let's get 200 of our best friends together to start talking about this. And that led to the very first event, auto Con Zero in Denver. Uh, November, 2023.

That event was planned. We, we pressed the go button in May 23. And, you know, in a very short time, uh, pulled off an event that we planned for 200 people and it ended up drawing about 350 people.

So that was pretty exciting to see that many people interested. Um, it validated the idea that there was a need for a place and a clear desire for focus on discussion of, um, automation content and, uh, community development. So that gave us the, uh, motivation to press the go button on future events, right?

Um, so through these events and through engagement in our Slack channel, um, we support the adoption and discussion network automation topics. Um, there's a seat at the table for anybody here. We kind of covered that already.

Um, we've talked about the community and organization in the events. Um, let's go through some more of those details. And the question.

So the attenuation of automation is not unique to just networking, right? We've seen resistance to automation in all sorts of industries and verticals over time. But in a sense, automation happens, right?

We're all in a sense, lazy, right? We don't, if there's a way to do our job that's easier. There's somebody who finds a way to do it and then they tell two friends and so on and so on.

Um, so we've seen some more dynamics within networking, but we've seen some specific resistance, um, especially for old school people like myself who had this real pride in I can troubleshoot a T one, I know how to interpret the results of a Bert test. Ethernet will never take over sonnet infrastructure, et cetera, et cetera. And, uh, people like myself have learned, you know, the world's changing things are a lot more software dominated and facilitated.

I need to adapt with the way the technical environment's adapting. And that's not always easy. You know, like even with our discussion earlier today on quantum computing, my head still hurts.

There's so much I need to learn to understand that, um, I'm not gonna get there, uh, before I retire, but I'm still engaged enough in the network automation community that I feel like I can still learn. You know, my python skills can still improve and so forth. Um, and there's a lot of people who have come to a similar realization and want to learn and grow and also want to help each other, uh, in this regard.

And that's been a place for, uh, uh, uh, network automation forums. Been a place to really connect people who are both looking for automation help and want to help each other. Alright, so to the question, why haven't we seen the full adoption of network automation?

We'll give you four pretty good answers that show different aspects of the, of the problem. Carol, you're gonna recognize this first one. So Terry Slattery, uh, founder of Chesapeake Computer Consultants.

Um, I had the privilege of working for him back in the nineties. That's where I met Carol. Sorry to embarrass you, Carol.

I missed. Um, you know, he heard us talking about this early on and he astutely noted, um, you're asking the question, but you're not providing any answers. I would like to provide some answers.

forum that basically covers the following, you know, four big items. Culture change is hard. You know, you have momentum.

It's hard to, hard to get out outside the box of what you're doing. Um, there's lack of adoption of network automation architectures, which turned into a really interesting kind of sub-project within naf. There's failure to even think about what it means to have a mature approach.

Network automation. Like there's no commonly understood maturity model, and there's lack of commitment across an organization like us as frontline network engineers. We may see the real value in automating, you know, repetitive tasks.

That doesn't mean anybody up the management stack, um, really agrees or helps or wants to put any support behind it. So that's Terry's set of answers. But wait, there's more.

Chris, you know, wrote on the topic as well, and he has really keyed in on the, uh, identity of the network engineer and how he clinging to being CLI jockeys, I'm horribly oversimplifying it. You should go read his piece. By the way.

The QR codes are the links to where you can go read more on this. Um, moving away from the CLI, when you identify with it as you know, so tightly, as many of us do, um, it's hard, right? And you don't want to change your identity in a sense.

Um, we've seen it, you know, in the industrial revolution. Um, it's not exactly the same this time, but there are common, common thoughts there. Um, and when we've gotta collectively work on shifting our identity, like what's the job of the network engineer or the NetOps team, you know, how we do it versus what we do.

You have to kind of think very carefully about the two, uh, the two different pieces to that equation. You might imagine. I've had some thoughts on this too.

Um, and, you know, since we're gonna keep referencing Phil Gervasi, um, and other friends in the community, you know, Phil and Ethan with packet pushers, um, will help me get this idea out there regarding, you know, we've had cowboy attitudes in the development of, you know, the internet infrastructure, especially in the nineties where you had this special combination of skills, you know, telecommunication circuitry and understanding that. And plus these specialized computers called bridges and hubs and routers and switches. Um, a lot of us like being heroes, like I'm the one that knows how to do this and don't do good, um, in terms of sharing information and how to, you know, spread that wealth to others.

Um, plus the fear of automation is out there and the current AI trend is not helping, uh, that fear as well. Um, these are things that we really need to, you know, take a close look at. I think, I think it's a very interesting, uh, point of view because the original internet routers were very automated, correct?

Um, the NSF net routers were all deployed and configured, uh, from a central push location. They pulled policy out of the PRDB. They were highly automated.

Yep. com file on those routers unless they really had to, um, follow on projects from that. Were nsfnet, were nsfnet related, like the, the RADB and the route servers.

Yep. Uh, were also highly automated. Uh, I know 'cause I was part of that project, right?

Um, and during the early, mid, early nineties, um, there was a definite resistance mainly by those who were, uh, operating commercial vendor equipment. Mm-hmm. It seemed like the commercial, the vendors at the time did not really, well, one, they didn't have a good automation story.

Uh, there wasn't an automation, automation eco ecosystem, if you wanna call it that. And a lot of the operators at the time were not incentivized to automate things, and the sphere of automation did creep in. Uh, so there was a lot of Sistance.

You're teeing up the next slide extremely well, by the Way. So you Haven't, you haven't seen It. Well, yeah, I haven't seen it.

But from what I saw when I was at, at, at the time, um, was a lot of resistance from mainly people who were working from vendor equipment as opposed to those who were building, uh, prototyping and building open source, uh, uh, uh, projects. You should submit a proposal for a talk for a future auto con on that. I'm not kidding.

Right. For people who have seen it at the front lines and have observations like that, that's a perfect type of content we'd like to see, um, in our, in our events. I'm sorry, I didn't mean to cut you off if You Oh no, I'm fine.

Um, so I expect to see a proposal from you in the future. So, um, to those points, I a beer first and I will. Okay, there we go.

Um, at our event in November 24, um, Anna Clayborne, who I believe is the VP of Network Automation at Zao, did an incredibly well researched keynote address on this topic. She didn't just come in and wing it and talk about her thoughts. She did real research and I highly recommend, um, if you click on that QR code, that'll take you to the YouTube, uh, link for her presentation.

Um, very well done. Where she goes through, you know, some of the history that you just touched on coming from legacy telecom practices, um, and how the CLI on certain vendor gear actually discouraged, um, programmability and interoperability across platforms. Like we created a whole certification industry or set of certification industries on memorizing commands for Junos, for EOS, for iOS and so forth, not throwing all those cert programs under the bus.

Right? Some of us have really cool leather bomber jackets, right? Associated with that.

And it was useful, um, for a time, but there were some unintended or perhaps intended consequences as a result of that. And I can't, I can't, uh, portray it the way Anna does. Go, go look at the video.

Um, she goes through a whole set of issues, um, along these lines. You should go read what she, or listen to what she has to say on this. There's more, there's more than just those four perspectives on what the hold ups are, but we see a whole slew of common themes where technology really isn't the hardest part.

Culture change. It's much more of a, a barrier than we ever expected to, um, identifying as a CLI jockey versus I'm here to provide connectivity, right? I'm here to make the network work and how I make the network work isn't the same as making the network work.

So with all these ideas coming into Confluence, there's some really interesting initiatives that have come, come up through naf. They're really, I hate to over, you know, I haven't mentioned the word grassroots, but so much of what we do is very much bottom up. Um, we really exist for end users to bring their stories to the table.

Um, we're not against vendor presentations. We really appreciate our vendor sponsors, um, the workshops that many of them provide and so forth. But we really wanna make sure there's room for people wrestling with real issues in their networks to be able to talk about, um, the challenges, the successes, one of the best presentation titles we've had so far as how I failed at my first, um, network automation implementation and why it's still in production today.

Um, you know, you can check those all out in the n YouTube channel, uh, for specific projects, you know, number one, starting with terms is usually really helpful and that's why we started the automaton, right? Let's, let's make sure we understand we're using the same terms the same way. What's the difference between automation and orchestration?

Um, things like that. Anybody can contribute to that. Um, there is a great network automation landscape document initiated by one gentleman, uh, based in Iceland, um, who has done some incredible research and putting all this network automation tooling in one place.

Um, you can go see what he's done. I I, I will get you the link if you're interested in that. Sorry, I don't have it here.

There is an active working group on developing, um, a network automation framework, right? Again, helping, you know, address things like Carrie Slattery brought up. There is a network service provider boff that's formed within the last six months.

Largely cable co folks right now, but more telco folks are, uh, contributing to that too. Um, we have been, we've seen lots of participation from enterprise, uh, networking sectors, but the telco and Hyperscaler folks are, are coming into, and my pitch to people who have been down serious network automation journeys, come talk about it. Come show how you've hacked your way through the jungle of network automation so that others can learn from what you've already wrestled with.

You know, in very large scale networks, there's talk of the need for vendor independent, um, education and certification for network automation. It's not an easy problem. Um, there's lots of facets to it as we talk about it more, um, but we're starting to reduce some entropy around that.

Um, and there's plenty of room for other things to happen here. Again, all it takes is interest. You know, if there are a couple individuals that see a topic in network automation and orchestration and adjacent tech areas like observability, um, like integration with trouble ticketing systems, you know, the door's open to put ideas on the table.

We're, we're very open to that. Near term futures, long-term futures, we're still gonna keep asking the question. We'll ask the question until it's been definitively answered and we don't need to ask the question anymore.

Um, it's a, it's a good motivator, right? Helps, uh, push the ball forward. Scott, could this be an opportunity to take a snapshot of the nature and progress of network automation?

Do you see it progressing or is it going sideways? Uh, So a great question. Short answer is yes.

We see progress very selfishly as we form network automation forum and have seen participation go up very organically. Just to, to give you the short snapshot, we targeted 200 people for our first event. We got 350.

We then sized for 400 in Amsterdam, may, uh, 2024, sold that out. We then sized for 500 for Denver. November 24, we got about 550.

Um, we're now shooting for 700 in Prague at the end of May, 2025. And we are in the middle of a venue search for November 25 that will accommodate up to a thousand attendees. And that's happened with a very small organization, which means there's a hunger.

We've tapped into a market and we found people that wanna come together and collaborate on this. So that's, that's one very concrete, um, way to put it. But there's, there's technology developments too, right?

You see, I wanna be very careful, um, with implying vendor endorsements. There are vendors that are doing much better with API availability capability and documentation. There are vendors that are doing a really interesting job adapting DevOps to network infrastructure.

Mm-hmm. And I would argue that's not a direct translation in some key areas. That's okay.

The whole idea of leveraging software centric tooling and ecosystem to manage network infrastructure, like, we're definitely seeing progress on all those fronts. So I, this carries over from our, our previous recording. I'm by nature an optimistic person.

Um, but I see real progress and I can throw other data points at you too. So, alright. There's still work to do.

So, so as the question been answered, yes and no, right? Um, the answers are multifaceted. I pointed you to really four decent sets of answers, but there's more to say than just those four perspectives.

We believe there's value in continuing to ask the question to move the ball down the field. If I go back to the beginning of this discussion, sorry, I've been talking a lot. This really isn't a discussion, but thank you for interrupting with questions to make it more of a discussion.

We are all networkers. We will continue to maintain this interest in building collaboration between network automation forum and other orgs. I mean, Tom giving U-S-N-U-A and naf, the ability to talk about these things shows the connections in the community, right?

And U-S-N-U-A coming and helping with some of the things last year at, uh, our event. Um, we have, um, support from European, um, no organization for our prog event coming up in May. Like this is just happening very, very naturally.

And that's like incredible to see that, um, let's look for other ways to cooperate and collaborate. Like, I don't know what they are, I'm just one first, right? And bring ideas to the table.

I have to ask why Prague? Um, so Amsterdam as Amsterdam was our first European event. Um, it was very much an easy button, um, with being able to, you know, being set up for international business, very well, flight availability and so forth.

It was also maybe a little more of the expensive button as well. Um, so we wanted to find, you know, other places where we might be able to do better financially. And, and Prague has hit some important check boxes there.

Alright, good choice. We're, I'm, I did a site visit in February and I'm very excited. The venue's great, the city's incredible.

If you haven't been there, if you can't come to naf, uh, auto Con three in may, find another reason to visit Prague. Um, we don't think network automation and the need for discussion on it is gonna end anytime soon. It, it might, you know, let's just say, Hey, we feel like we make enough progress on this and maybe there is an end to what NAF needs to do at some point in the future.

That's completely possible. Um, we don't think it's gonna happen in the next few years. Um, we've called out auto con three last full week of May, uh, 2025 and Prague Auto Con four.

Um, the date will most likely be the last full week of November before Thanksgiving week. Um, and we structure our events, it's five days. The first two days are very detailed hands-on workshops.

Um, that's typically Monday and Tuesday, Wednesday, Thursday, Friday are the typical conference sessions. Um, 26 and beyond. We have ideas.

We, we know that if we keep growing, if we need to find a venue for more than a thousand people, we need to be looking for that venue now. Um, we are trying to get more proactive on, on those things. So, um, final thoughts.

Um, I'm gonna give you a very personal perspective on this. Um, when I left corporate employment to do my own consulting work, NAF was not part of my plans. I, you know, had the interest in network automation, the connection with Chris was very easy and very helpful.

Doing my own consulting work gave me the flexibility to get engaged in this and just try it. Like we had no idea if this would justify our involvement from a financial perspective. And notice neither of us are ready to stop doing other consulting work, right?

Um, this is a labor of love. Um, but, uh, you know, it's, it's a, it's a good thing for us to be involved in. And I've just really been impressed with the power of community.

Again, I know it can be a trite statement, but seeing people come together very naturally, just providing a place for them to come and discuss this, um, you know, it's not, not unlike what's happening with the local nugs for U-S-N-U-A, we're different in that we're only doing this twice a year. We're providing a much bigger space and it's very focused on a set of topics. Whereas the Nugs and Nano and other, other organizations are very broad in the tech scope.

They, they incur. Um, I've been blown away by what I've seen. Um, and some of the presentations of people who've never given a public presentation before, it's been amazing.

And we encourage that if there are people who haven't done anything like this before and want to get up in front of a crowd, we'll help coach, um, with content, um, with presentations. We're, we're totally happy to help, uh, new folks into the field here. It's, uh, it's been an eye-opening experience for me.

So come check us out. Network Automation Forum. Um, you'll see a join button for our Slack.

You do not have to have attended an event if Slack is open to anybody who wants to participate. Um, we have an active LinkedIn group as well. If you're comfortable giving your email, we can help.

Make sure you get info from us. If LinkedIn, uh, totally goes over a cliff next year, um, we have a way to get in touch with you. Um, all the materials are available, no, no email address needed.

Um, on the YouTube channel, there is a network automation forum, YouTube channel. I encourage you to check it out. Um, if you want to be a part of this conversation, uh, come on in, pull up a chair, there's a seat for you in the discussion.

What other questions can I answer for you? Chris's survey, uh, how much is that leveraged in the conversations or, you know, survey work in general? So Chris's work has been super helpful.

He actually presented on it last November and uh, you can check out that YouTube presentation of his, uh, what he's found so far. And it, you know, he finds more things to add to it each year, No doubt. Okay, Very good.

Thank you so much for the time. Really appreciate it.