How Are AI and Automation Impacting Modern Network Demands?

AI and automation empower teams to become more efficient by supporting tasks related to network monitoring, change management, performance, security, and so much more.

However, the power consumption required by the technology will continue to overpower existing networks. Performance challenges, longer processing times, and emerging security threats are only a few hurdles you’ll face because of larger workloads.

In this archived keynote session, Neha Kaul, head of automation and AI technology for Guardian Life, explores the ever-evolving challenges and necessary adaptations for network leaders due to the rise of generative AI and larger workloads.

This segment was part of our live virtual event titled “How to Make Network Management Easier in 2024-25.” The event was presented by Network Computing on September 19, 2024.

A transcript of the video follows below. Minor edits have been made for clarity.

Neha Kaul: For GenAI and larger workloads, we are going to see more and more pressure on our existing networks as enterprises. We are going to see performance challenges, longer processing times, and new threats coming up, especially with AI.

With the black box models, we really don't know what they carry. There has been a lot of emphasis on open-source predictive AI models, including open-source GenAI models. We don't know where they get injected with what.

Related:Is Network-as-a-Service Your Golden Ticket?

So, there's going to be a larger scope on the threats. Now, what do I recommend? In a time when network leaders are looking forward to the changing demands of consumers and their own businesses regarding use cases, I recommend they relook at their foundations.

At the base of the foundation is our network architecture. If we have an optimal network architecture, scaling that architecture becomes very easy. Scaling that architecture would mean adding newer components that can easily be integrated into your existing architecture.

It's important to relook at these architectures, especially from the security perspective with zero trust. As we implement that, we also augment all of that with automation. We're going to talk more about AI in the further slides, but this is how I look at the future direction. Setting the foundation for that future direction.

It wouldn't be an overstatement if I said that in 2025, network automation and AI tools, along with an optimal architecture, will not only be a luxury, but a necessity that most network and technology leaders must look at.

So, in this slide, I just wanted to touch base on what all of us have seen happening in the last 10 years. In 2010, we had IPv6 adoption happening, and it continued through the 2020s as well. But that's the pivotal point where we were able to get larger IP addresses, spaces, and improved network capabilities to deal with most of our networking needs.

Related:Hurricane Helene Communications Outages Show Need for Greater Network Resilience

From 2014 to 2016, it was further augmented by LPWAN technologies coming up. These were low-power, wide-area networks that helped us expand to the wider areas and yet had low power consumption. Because at the end of the day, when you run your data centers and all these high-load networks, you must think about the cost savings.

That's why low power consumption helped. This also helped us connect to the Internet of Things (IoT), which is all your internet devices. These are your mobile phones and many other devices that are network enabled. Homes are even network enabled now. So, all that saw massive growth in 2014-2016.

Somewhere around 2015 to 2017, we had SDNs and NFVs, again, shaping the future of cloud. There is more virtualization happening. SDNs are expanding and decoupling a lot of our software from the hardware.

One point I want to make is that all these technology advancements were helping set a base for the next wave of advancements. When I was putting the slides together, it gave me a chance to look at this evolution and made me realize that none of it was in silos.

Related:Everyone Has a Service Recovery Plan Until They Don’t

Everything that happened as a part of the advancement had something prior to it, asking it to happen, and something in the future where these technologies were setting and filling those gaps.

Watch the archived “How to Make Network Management Easier in 2024-2025” live virtual event on-demand today.