Kubernetes Cross-Zone/Region Simplified: Harnessing Bacalhau for Efficient Distributed Compute

This conference talk explores how organizations can overcome the challenges of processing exponentially growing distributed data by implementing compute-over-data strategies with Kubernetes. Learn how to bring machine learning compute directly to data sources rather than using unsustainable centralized processing approaches. David Aronchick from Expanso demonstrates practical solutions to address massive data transfer costs, regulatory compliance issues, and network reliability problems that currently slow ML infrastructure adoption. Discover real-world implementation examples, including an energy company managing 15,000 microgrids, where processing data in place ensures compliance, reduces costs, and improves reliability while maintaining centralized control. The presentation showcases open-source tools like Bacalhau that provide practical patterns for modernizing ML infrastructure in distributed environments, particularly valuable for processing data from distributed deployments or analyzing real-time sensor edge data.