Agentic AI, Automation, and the Future of Securing Networks at Networking Field Day 39 – Tech Field Day Takeaways

I'm Tom Hollingsworth, event Lead for networking, and here are my takeaways from Networking Field, A 39 Networking field. A 39 was a great event to round out 2025 because the one hot topic that's on everybody's lips was first and foremost at this event. Of course, I'm talking about ai.

You probably think to yourself, well, what does a GPT algorithm have to do with networking? And my answer is a lot more than you think because there are so many different facets to what AI offers, especially in the realm of networking that you might be remiss in thinking, well, this is just an easy problem to solve or I don't necessarily know that there's a whole lot people can do about it. In these takeaways, I hope to show you that there's nuance in this conversation and the companies are attacking it in a very different way to bring you the network of the future, something that is resilient and reliable and capable of expanding to whatever you want to do with it.

My first takeaway from Networking Field A 39 is that root cause analysis is one of the most concrete use cases for ag agentic AI solutions. In networking, not only does AI enhance the speed with which engineers can find the real causes behind issues, but also evidence-driven guidance offered by the agents means that you always have the backup data for your conclusions. AG agentic AI can also fully automate remediation workflows to ensure that problems are resolved as quickly as possible.

Root cause analysis is something that has always been a struggle for people in the networking space because we never quite understand all of the underlying issues and we tend to solve the symptoms rather than the problem. We hope that if we solve enough of those symptoms quickly that we can accidentally solve the actual problem, and as is often said, there's nothing more permanent in it than a temporary solution, which is often what causes these larger systemic issues is a temporary patch for a smaller problem than it escalates out of control. You need to have some kind of a visibility into your entire network to pull together all of those pieces and give you the body of knowledge necessary to ensure that you're solving those problems.

This was illustrated perfectly by our friends at Nokia because they have a platform called event-Driven Automation or IA that showed this off very well in the demonstrations at Networking Field Day, you can use natural language to query the database to understand what's going on. You can tell it to look for things that might not seem right and you can configure notifications and triggers to let you know if something falls outside of those parameters so you you can start working on it immediately. It's the kind of knowledge base that really helps turn junior network administrators into more senior people and senior people into the kind of sages that we come to rely on in the industry to make sure that we're resolving these issues as quickly as possible.

My second takeaway is that data governance is becoming more important as workloads, transit clouds and enterprises all over the world. One of the best ways to keep data secure and transit is to properly build out a secure network transport. One of the biggest things that we deal with on a daily basis is not only making sure that that data is secured, but also being able to provide reports for auditors to ensure that we can prove that the data was protected in flight.

A company that is doing an excellent job in this space is grapht. They have a network as a service offering that allows for enhanced data governance and reporting. For example, centralized policy enforcement means that rules are applied equally everywhere, so there are no gaps.

You no longer have to worry about executives working under one set of policies and knowledge workers working under a different set of policies, and the gap in between the two allows for people to get in and take things that don't belong to them. Decoupled governance means that data doesn't need to be decrypted to apply these policies. One of the biggest advancements that we've seen in networking over the years is the ability to treat traffic that's encrypted as if it's not for enhancement of policy enforcement and things like that.

What all of this is partnered with robust transport security like microsegmentation and end-to-end encryption, and it's a properly built overlay network, and that is something that people really need to have in their toolkit. You need to be able to ensure that the data is safe as soon as it leaves the location all the way to its destination. When you think about the amount of data that is being uploaded to the cloud for AI purposes, you know how quickly things can go awry if that data stream were to be intercepted or manipulated in some way, and that's something that we're starting to see a lot with attackers who are looking maybe not to steal that information, but to force incorrect conclusions by midstream manipulation with a properly built encrypted end-to-end overlay like the one from graph, that's not something that you have to worry about.

My third takeaway is that the future of AI networking is not just GPUs as outlined by Cisco during this event. There's a lot of design work that goes into the front side fabric that drives data being delivered to those GPU clusters. I know a lot of people in the industry tend to focus on the high speed interconnects between GPUs or between the GPUs and the storage network.

The value of the front side network that is the one that is sending data into the AI cluster cannot be understated. I know that we talk a lot about using ethernet for this, but you have to understand that the ethernet has very specialized protocols for things like high speed packet delivery with minimal congestion, special kinds of congestion notification that allow packets to be reordered or take different paths to reduce latency. Things like per packet, spraying mean that multiple paths can be used to increase bandwidth, but more importantly, enhancements like adaptive routing ensure that the packets arrive in the proper order to be processed.

And in this particular instance as outlined by Cisco, it can matter with nanoseconds because the amount of time that your GPU clusters spend idling or waiting for a response before they can return the data that they've been investigating, that has a very real cost to the organization. And if you don't understand that, you probably shouldn't be designing one of these networks in the first place. The Cisco presentation did an amazing job of highlighting just how important each of those steps are and where those situations need to be addressed, upgraded, and possibly rethought.

My final thoughts from this event are networking is in a very good place right now. We are facing challenges from new technologies like AI that are not only forcing us to re-engineer the way that we build these networks, but also think about the way that we use them. I've heard from a lot of people that AI will really help our network operators focus on the important parts of their job instead of reinventing the wheel every time.

We need to come up with a new way to monitor or maintain something. And not only that, but AI on the backend is gonna drive a lot of product refreshes and a lot of new thinking in the way that we need to architect our networks. I think that there's a long future ahead of us for everything that we're working on, and networking Field Day is actually one of the places where you're gonna be hearing the most about it.

Big shout out to Dustin Demers for being a first time delegate at this event who joined a number of alumni from over the years to learn about what's going on here. Don't forget that you can check out all of the coverage from this event at net, the Networking Field Day event page, and you can read more on all of our delegates personal pages and on their LinkedIn. You can find links there.

Thank you very much for watching this episode of Tech Field Day takeaways on the Tech Field Day plus YouTube channel. If you enjoyed this, please be sure that you subscribe and share your thoughts with the Networking Field Day community. In the comments.

Be sure to follow Tech Field Day on X, Twitter, blue Sky and Mastodon for more updates and coverage, and be sure to check out the presentation videos and the Tech Field Day website and Tech Field Day YouTube channels. Our next event is gonna be Tech Field Day Extra at KubeCon, so make sure that you are tuned in and following along there. Thanks for tuning into this episode of Tech Field Day takeaways, and we will see you soon.