Networks shift focus from speed to operational agility amid AI-driven traffic complexities

As cloud and AI workloads surge, organisations face new challenges in maintaining network performance, with operational flexibility, automation, and skills shortages now at the forefront of enterprise network evolution.

Enterprise networks are moving into a phase where sheer bandwidth is no longer the only measure that matters. As cloud services, artificial intelligence and more distributed applications drive ever-larger data flows, organisations are finding that the harder problem is not simply carrying traffic, but doing so with enough consistency, visibility and operational discipline to keep pace with demand.

Brett van Rooyen, NEC XON Adtran Product Manager, told ITWeb that the rise of cloud and AI workloads is changing what businesses expect from their networks. He said enterprises now need greater capacity, lower latency and more predictable performance as data volumes grow. The pressure is being felt across both mature and emerging markets, where teams are often expected to support increasingly complex environments with too few specialists.

That skills shortage is emerging as a central constraint. Van Rooyen said many customers do not have the in-house expertise to manage modern network estates, which is helping to push more companies towards integration-led strategies and managed services. The point echoes a broader industry concern: as networks span more clouds, more tools and more layers of abstraction, the operational burden can outstrip the benefit of adding raw infrastructure alone.

Byren Meintjies, senior manager of global business development at Adtran, argued that the shift is being driven as much by application behaviour as by network architecture. Speaking to ITWeb, he said workloads are becoming more distributed and more real time, with heavier data movement placing fresh demands on transport networks. That, in turn, is increasing interest in higher-capacity optical systems and transport designs that are easier to run at scale.

One approach gaining attention is IP over DWDM, or IPoDWDM, which combines IP and optical layers in a way that can reduce the number of components to deploy and manage. Meintjies said this can simplify operations, improve efficiency and make expansion less cumbersome. The theme is consistent with wider industry analysis from Accenture, NTT Data and TechRadar, all of which argue that legacy network structures are struggling to cope with AI-era workloads, particularly where latency, reliability and monitoring are concerned.

The broader conclusion from the Adtran and NEC XON executives is that future networks will be judged less by headline throughput than by how easily they can be operated. Flexibility, automation and tighter integration are becoming just as important as capacity, especially in organisations that lack deep internal network expertise. In practical terms, the enterprise winner may be the one that can manage complexity most effectively, not the one that simply buys the most bandwidth.

Source Reference Map

Inspired by headline at: ^[1]

Sources by paragraph:

• Paragraph 1: ^[2], ^[3], ^[7]
• Paragraph 2: ^[1], ^[2]
• Paragraph 3: ^[1], ^[6]
• Paragraph 4: ^[1], ^[2], ^[4], ^[5]
• Paragraph 5: ^[1], ^[3], ^[4], ^[7]

Source: Fuse Wire Services