RisingWave is a distributed SQL streaming platform that continuously ingests data from event streams, databases, and data lakes, transforms it in real time, and serves low‑latency query results via a PostgreSQL‑compatible interface. It combines built‑in storage with a decoupled compute‑storage architecture, eliminating the need for separate databases, caches, or message‑queue layers, and supports sub‑100 ms freshness for use cases such as real‑time dashboards, fraud detection, and IoT telemetry processing.
Funding
Funding not disclosed
Founders
Product
Problem
Building real-time data pipelines typically requires stitching together multiple systems—message brokers, stream processors, state stores, caches, and serving databases—to achieve low-latency analytics and operational insights. Managing these components increases operational complexity, latency, and development effort, while exact‑once guarantees and scaling can be difficult to achieve.
Solution
RisingWave provides a single distributed SQL streaming platform that continuously ingests data from event streams, databases, and data lakes, transforms it using PostgreSQL‑compatible SQL, and serves query results with sub‑100 ms freshness. The platform includes built‑in row storage for low‑latency serving, eliminating the need for separate caches or serving databases. Its decoupled compute‑storage architecture allows independent scaling of processing and storage resources, while exactly‑once semantics ensure data correctness. By using standard SQL instead of custom Java or Scala code, developers can build and maintain streaming applications more quickly. RisingWave also integrates with Iceberg and DataFusion for lake analytics, enabling unified real‑time and batch workloads on a single engine.
Target Audience
RisingWave is aimed at engineering teams building real‑time analytics dashboards, fraud detection, monitoring, IoT telemetry, AI agent infrastructure, trading systems, and streaming ETL pipelines that require continuous SQL processing over event streams.
Features
- PostgreSQL‑compatible SQL engine for stream processing, supporting joins, aggregations, and windowing
- Built‑in row storage that serves low‑latency queries directly, removing the need for external caches or databases
- Decoupled compute‑storage architecture for independent horizontal scaling of processing and storage nodes
- Exactly‑once processing guarantees with instant failure recovery
- Native connectors for Kafka, databases, and data lakes, plus managed Iceberg and DataFusion support for lake analytics
- Ultra mode with in‑memory processing for sub‑100 ms latency workloads