Networks have gone through quite the evolution, especially in the past few years as the cloud has increasingly integrated into network operations. Due to increased virtualization, network components are no longer in a centralized location, like a data center.

“This is no longer everything running in a single data center, knowing exactly what device something is running on and being able to go out, look, touch and feel that device in the network,” Wollenweber said. “We’re in an extremely distributed world, and we want to be able to provide things like complete visibility across that ecosystem.”

Complex networks are hard to manage for multiple reasons. Wollenweber said some of the primary factors contributing to network complexity include the following:

• Growth in global internet traffic.
• Accelerated operations.
• Data surges from AI workloads.
• Rising costs for similar products.
• Increased demands for sustainability.
• Rise in cyberthreats.

Opportunities arise with the challenges brought on by network complexity, Wollenweber said. Network professionals should strive for the following goals to have a simpler network.

1. Simpler operations

Simplifying network operations is key to reducing network complexity.

“Take out as much complexity as possible,” said Michael Blake, vice president of IT at PTC, a lifecycle management provider based in Boston, during the session. He encouraged network teams to conduct a total network examination, treating on-premises and cloud infrastructure as parts of a whole network. This examination means using “the same process and procedure across [all infrastructure] so long as you’re as efficient as possible,” he said.

End-to-end visibility is crucial for simplicity and a greater understanding of the network. Automation can facilitate increased visibility, tracking various processes end to end, and it improves efficiency, while lowering Opex. Automation can identify and correct problems within the network as soon as they arise. The panelists touted AI’s integration into networks, especially automation, as an avenue toward simplifying operations.

However, AI’s massive workloads have caused a surge in data processing. Data centers now require hardware that can handle those workloads.

“When we look at things like data center versus a typical office, we see a 50-times power requirement for data centers that is driving huge challenges for sustainability objectives,” said Bill Gartner, senior vice president and general manager of the optical and optics group at Cisco, during the session.

One of Cisco’s responses to this issue was its upgraded 8000 series routers, as well as new pluggable optics and transponders with higher capacity and a smaller footprint. Cisco’s new and upgraded technologies conserve space and energy in the data center, aiding in enterprise sustainability initiatives. These smaller components can now handle the same workloads as their larger, legacy counterparts, if not more.

“You can put what used to be large multichassis systems that were multiple racks and take up an entire room into a single chip and into a single-rack unit,” Wollenweber said.

2. Speed and reliability

As global internet traffic continues to increase year over year, the need for faster and more reliable networks has also increased. According to a 2023 Cloudflare report, global internet traffic grew 25% in that year, with the previous year showing similar results. The growth is no surprise considering how more IoT devices connect to the internet each day.

Wollenweber explained that faster speeds and 100% uptime are important for customers. To enable these things, networks must increase their bandwidth. However, legacy networks are struggling to keep up with bandwidth demands. To increase bandwidth and network speeds, engineers must consider network operation capabilities.

“As more applications come on and more bandwidth is needed in the network, we have to think about how to operate the network more efficiently from a resource perspective but also in speeding up how you can bring these new things [to the] network faster,” Wollenweber said.

3. Future-ready infrastructure

The panelists emphasized the need for future-ready infrastructure to build a simplified network. Anticipating what the network might need in the future stops it from becoming obsolete quickly. If organizations can avoid the need for updated infrastructure later, it saves them money in the long run. Right now, creating a future-ready network means adapting it for AI integration.

“AI is transforming how we build networks in the future,” Wollenweber said.

Though it might be tempting to integrate AI into many different technologies, enterprises need the supporting infrastructure to do it. Adding AI into networks is a step-by-step process, said Dr. Hirochika Asai, vice president of infrastructure technology at Preferred Networks. AI’s integration into automation is an ideal starting point, aiding with both simpler and faster operations.

Ultimately, AI is a new IT muscle, said Kevin Deierling, senior vice president of networking at Nvidia. For AI to become a strong part of a network, the network and data center needs to transform to accommodate it.

Nicole Viera is assistant site editor for TechTarget’s Networking site. She joined TechTarget as an editor and writer in 2024.