This session presents Ursa, the groundbreaking lakehouse-native data streaming engine that recently won the prestigious Best Industry Paper Award at VLDB 2025 in London. As the first Kafka-compatible streaming platform built specifically for lakehouse architectures, Ursa represents a rare achievement: academic research excellence that delivers immediate, transformative production value.
Key Technical Innovations:

Leaderless Architecture: Eliminates costly inter-AZ replication and leader elections
Native Lakehouse Integration: Direct streaming to Iceberg and Delta Lake formats
Zero-Copy Data Access: Unified streaming and analytical workloads without ETL

Real-World Impact:

95% infrastructure cost reduction compared to traditional Kafka deployments
5GB/s sustained throughput at just $54/hour vs $1,000+ for conventional solutions
Fortune 500 production deployments with 99.95% SLA

The presentation will demonstrate how Ursa's award-winning architecture solves three critical industry challenges: escalating streaming infrastructure costs, persistent data silos between operational and analytical systems, and growing demands for real-time AI capabilities.
Attendees will learn practical migration strategies from traditional Kafka deployments, understand the technical foundations that earned VLDB recognition, and see live cost comparisons showing immediate ROI potential. The session includes real customer testimonials and integration examples with modern data platforms including Databricks Unity Catalog.
Target Audience: Data engineers, platform architects, and infrastructure leaders seeking to modernize streaming architectures while dramatically reducing costs.
Takeaway: A clear roadmap for adopting research-proven technology that's already transforming how organizations approach real-time data processing in cloud-native environments.