Webinar - Dynamically orchestrating RAN and AI workloads on a common GPU cloud

Mobile Network Operators (MNO) typically design 5G RAN infrastructure for peak traffic scenarios leaving compute resources under utilized during off-peak periods. RAN hardware, on average, is utilized at 20-30%. Since RAN is now disaggregated and requires acceleration both RAN and AI could run on the same GPU infra.

In this webinar we demonstrated dynamic orchestration of RAN and AI workloads by scaling in and scaling out the RAN and AI workloads based on the RAN traffic in a completely automated manner. This approach ensures optimum utilization of the infra and thereby reducing the ROI for existing CAPEX investments. In fact, by selling unused GPU cycles to an aggregator as spot instances, MNOs can monetize their RAN hardware to further shrink the ROI period to just about two years.

In the webinar, we covered:

Strategies for optimizing GPU infrastructure.
Dynamic scaling techniques for RAN and AI workloads.
Ways to maximize ROI and explore new revenue opportunities for MNOs.

To know more about our offering please visit our web-page https://www.aarnanetworks.com/solutio...

For questions related to this topic, feel free to get in touch with us at [email protected]