PuppyGraph

StreamNative and PuppyGraph enable real-time graph analytics on streaming data. PuppyGraph became the first graph query engine integrated with Ursa Engine, StreamNative's lakehouse storage layer, enabling zero-ETL graph queries directly on data stored in Apache Iceberg tables.

Zero-ETL Graph Queries

PuppyGraph provides a graph query layer that operates directly on data streamed through Kafka and Pulsar and stored via Ursa Engine — no data copying or ETL pipelines required. Teams can:

• Query streaming data as graphs using Gremlin or openCypher query languages
• Discover relationships and detect patterns on continuously updated data
• Run graph algorithms (shortest path, PageRank, community detection) on fresh data

Ursa Engine Integration

With Ursa Engine's Iceberg-based storage, PuppyGraph reads directly from the lakehouse layer. As new events flow through StreamNative, they're written to Iceberg tables and immediately available for graph queries. This eliminates the traditional graph database ETL pipeline and ensures analysts always work with the latest data.

Use Cases

The StreamNative + PuppyGraph combination is ideal for:

• Fraud detection networks — trace transaction chains and identify suspicious patterns in real time
• Supply chain visibility — model supplier relationships and track material flows
• Knowledge graph construction — build and query continuously updated knowledge bases from streaming sources