Subscribe to stream
The hit series on the hyperscale data center ecosystem

Panel: How is the modern data center responding to the growth of complex computations?

With the increased demand for artificial intelligence, machine learning and many other complex workloads of the future, it is becoming increasingly challenging to process terabytes of bandwidth in one location.

With network technology and the infrastructure required to support it (e.g. NICs, switches and cables) evolving rapidly to address CPU bottlenecks, we ask our panel to walk us through the evolution of data center processing.

Do we know enough about smart NICs and their power to free up server processing for primary tasks? And, can they have a future outside the hyperscale data center? And, what about the server - are DPUs a game-changer for workload offloading further up the stack? Is the industry ready to embrace innovation at hardware and software level in the era of accelerated computing?