Litmus Edge + Zerobus Ingest: Streaming Industrial Data to the Lakehous
Industrial-scale telemetry streaming from Litmus Edge into Databricks with Zerobus Ingest.
Industrial AI depends on OT data. But getting that data from the factory floor into the lakehouse usually means batch exports, custom ETL or Kafka clusters to manage. This white paper from Litmus and Databricks lays out a simpler path. Litmus Edge connects and normalizes plant data at the edge. Zerobus Ingest then writes it directly into Delta tables governed by Unity Catalog. In a simulated benchmark, the setup streamed about one million telemetry messages per second.
Inside, you’ll learn:
How Litmus Edge and Zerobus Ingest stream OT data directly into Delta Lake, with no Kafka brokers, connectors or custom ingestion pipelines
Benchmark results and the full test setup for streaming ~1M messages per second across three parallel Zerobus streams
How edge store-and-forward and Zerobus backpressure control keep data flowing through network outages and workload spikes
How Unity Catalog makes industrial telemetry governed and ready for analytics, AI and natural-language queries as soon as it lands
Best practices for spreading load across parallel streams, normalizing data at the edge and sizing infrastructure at scale
Authors: Vishvesh Shah, Director of Product Management, Litmus · Victoria Bukta, Member of Product Staff, Databricks
