The Joe Reis Show Vinoth Chandar - The Future of Open Data Lakehouses
30 snips
Apr 1, 2025 Vinoth Chandar, CEO of OneHouse and creator of Apache Hudi, shares insights on the future of open data lakehouses. He discusses the evolution from traditional databases to modern lakehouses, emphasizing scalable architectures and real-time data management. Vinoth tackles SQL query engine challenges and advocates for a unified data management approach across isolated systems. He also highlights a project aimed at enhancing data interoperability and stresses the significance of open-source solutions in fostering a competitive yet collaborative data landscape.
AI Snips
Chapters
Transcript
Episode notes
Uber's First Data Lakehouse
- Vinoth Chandar built the first data lakehouse at Uber in 2016.
- It was called a "transactional data lake" because it combined data lake scalability with database-like transactions.
The Rise of the Lakehouse
- Data lakes lost popularity, but the need for database features in data lakes led to their resurgence as lakehouses.
- The challenge was combining the scalability of data lakes with transactional features of warehouses.
Key Problems Driving Lakehouse Evolution
- The evolution of data lakes to lakehouses was driven by three core problems: data mutation, fast updates, and change logs.
- GDPR further highlighted the importance of data deletion and management in data lakes.