Maximizing Global Potential - Cost-Optimized, High-Availability Workloads Across Regions

Learn how to design and implement cost-optimized, high-availability workloads across multiple regions and cloud providers in this 31-minute conference talk from CNCF. Discover strategies for running compute-intensive stateless workloads like AI inference and video encoding by distributing them across regions to optimize both cost and availability while avoiding resource shortages in single regions. Explore multi-cluster scheduling techniques and Karpenter's dynamic provisioning capabilities across preemptible and non-preemptible node VMs, and understand how to integrate with cloud providers' pricing and availability APIs for unified cost-efficient scheduling and autoscaling. Master traffic routing strategies to maintain service responsiveness and efficiency regardless of pod location, and gain insights into building resilient multi-cloud architectures that adapt to fluctuating costs while ensuring seamless workload execution.