risingwave

RisingWave is an event streaming platform that replaces traditional complex stacks (such as Debezium, Kafka, and Flink) by unifying data ingestion, incremental processing, and low-latency serving. It is designed to provide fresh, queryable data for real-time applications and AI agents. Key capabilities include: - **Multi-source Ingestion**: Supports webhooks, native CDC from databases (PostgreSQL, MySQL), event streams (Kafka, Pulsar, Kinesis), and historical batch data from S3. - **Incremental Processing**: Uses materialized views to perform continuous computation, ensuring end-to-end data freshness under 100ms. - **Low-Latency Serving**: Maintains results in an internal row store to serve SQL queries with 10-20 ms p99 latency. - **Open Storage Integration**: Writes to Apache Iceberg tables for long-term retention and analytical access, utilizing Apache DataFusion for vectorized query execution. - **Cost-Efficient Architecture**: Stores internal state and materialized views in object storage (S3) to reduce costs and enable elastic scaling, with an optional disk cache for performance. - **Compatibility**: Connects via the PostgreSQL wire protocol and provides an MCP server, CLI, and Skills for AI agent integration. Common use cases include real-time monitoring and alerting, feature stores for ML, live dashboards, real-time data enrichment, and streaming lakehouses.