Speed at Scale on the Lakehouse with Lakehouse//RT
Overview
| Experience | In Person, Virtual |
|---|---|
| Track | Data Warehousing |
| Industry | Energy & Utilities |
| Technologies | Databricks SQL |
| Skill Level | Beginner |
Data leaders frequently face a challenging architectural decision: when deploying customer-facing dashboards or embedding analytics into applications, they often spin up a separate real-time serving layer to handle low latency, high concurrency workloads. While this can address performance, it introduces significant complexity, governance risks, and infrastructure overhead.
This session challenges the necessity of that serving layer. We will explore how organizations can consolidate their infrastructure by serving high-concurrency end-user analytics directly from the lakehouse with Lakehouse//RT, through real-world customer case studies and benchmarks.
Session Speakers
Himanshu Raja
/Sr Director, Product Management
Databricks
Matt Holzapfel
/Sr Product Adoption Architect
Databricks
Full Summary
Sub-second analytics at scale: how Lakehouse RT reshapes what data apps can deliver
Fast dashboards are easy in a demo. Keeping them fast under real user load is the hard part. The session examines why embedded analytics often feel sluggish and introduces Lakehouse RT, a new warehouse compute type on the Databricks Platform that targets tens-of-milliseconds latency at high concurrency without copying data or fragmenting governance.
FAQ
No. Lakehouse RT queries your existing Delta Lake and Iceberg tables under Unity Catalog. Adopting it is described as selecting a new compute type for Databricks SQL, with no data movement and no parallel permissions model.
It uses a new native engine, Raiden, integrated with Photon. By eliminating JVM overhead, leveraging modern cloud hardware and SIMD, and caching data, metadata, and query plans, it targets tens-of-milliseconds latency even under high concurrency.
Lakehouse RT handles fast serving. ZeroBus provides low-latency ingestion into managed tables that predictive optimization and liquid clustering keep query-ready. A real-time mode in Lakeflow Declarative Pipelines is planned to bring transformation into the same millisecond envelope.
Applications can use the SQL execution API now. ODBC and JDBC driver support is planned, allowing existing SQL-based tools to point at the new engine without changing how they connect.
The speakers stated that Lakehouse RT is priced the same as Databricks SQL. The value proposition is sub-second performance and higher concurrency without adding a separate serving system to run.