Mark Berly, CTO, information middle networking at Aruba, a Hewlett Packard Enterprise firm, famous that issues like zero contact provisioning have expanded the usage of automation lately. He famous that for recognized, present processes, automation is already well-established. The place AI goes past is that it could actually increase present processes. The excellence is that earlier automation was extra about codifying particular use instances and workflows.
In distinction, the panelists see AI-powered community automation as a shift in direction of extra adaptive, autonomous capabilities that may deal with sudden conditions, fairly than simply predefined duties.
The trail towards totally autonomous networks, the place the whole lot is automated, is likely to be coming sooner or later, however it’s going to take time. Berly famous that he has a self-driving automotive and joked that it has tried to kill him not less than as soon as, which is why he now solely makes use of the function for parking the automotive.
“I do suppose we’re getting nearer and nearer to that autonomous state and actually, it scares me. Like I mentioned, my automotive virtually killed me, I’m certain my community will attempt to crash itself,” Berly mentioned.
Influence of adopting genAI on present community capability
Because the adoption of generative AI (GenAI) continues to speed up, the impression on present community capacities and topologies is turning into a urgent concern.
In a panel session, Gerald de Grace, cloud architect and technical product supervisor at Microsoft, highlighted the immense scale of the problem, noting that the corporate is taking a look at clusters with over 300,000 GPUs. The sheer variety of parts concerned implies that failures are inevitable, and de Grace emphasised the necessity for automated methods to rapidly detect, isolate, and mitigate these points.
