RisingWave
A PostgreSQL-compatible streaming SQL database — continuously-maintained materialized views over streaming/CDC sources, served from the same system, with Iceberg-on-S3 output.
Summary
A PostgreSQL-compatible streaming SQL database — continuously-maintained materialized views over streaming/CDC sources, served from the same system, with Iceberg-on-S3 output.
The "streaming database" answer to Flink's "stream processor": replaces the Flink-plus-OLAP-plus-glue stack with one SQL endpoint for workloads that fit a SQL surface. Also the most-cited editorial voice of the 2026 Kafka-less ingestion shift.
- It is not a general dataflow engine — code-level custom operators and non-SQL logic still belong in Flink or Bytewax.
- Postgres compatibility is the wire protocol and SQL surface, not a drop-in Postgres replacement.
alternative_toApache Flink — for SQL-shaped streaming workloadsenablesCDC into Lakehouse — CDC sources to Iceberg without KafkasolvesLegacy Ingestion Bottlenecks — one system instead of pipeline + serving layer
Definition
A PostgreSQL-compatible **streaming SQL database**: you write SQL against continuously-maintained materialized views, and RisingWave handles ingestion (including CDC sources), incremental computation, and serving in one system. Sits in the "streaming database" category (with Materialize and Decodable) as opposed to the "stream processor" category (Flink, Bytewax) — you declare what should stay fresh, not how the dataflow runs. Writes results to Apache Iceberg on S3 for lakehouse consumption.
The traditional real-time stack is a Flink pipeline plus a separate OLAP database plus the glue between them. For teams whose workload fits a SQL surface — streaming aggregations, multi-stream joins, window functions feeding dashboards and alerts — that stack is two systems too many. RisingWave replaces the pipeline-plus-serving-layer pattern with one Postgres-compatible endpoint, no Kafka required and no JVM to operate.
Real-time dashboards and alerting straight from CDC sources, SQL-only streaming transformations into Iceberg on S3, replacing Flink+OLAP pairs where the logic fits SQL, teams without JVM expertise that need distributed stream processing.
Recent developments
- The reference voice of the "Kafka-less" and "SQL is enough" positions. RisingWave's 2026 comparison series (Flink alternatives, Debezium alternatives, streaming-without-JVM) is the most-cited articulation of the year's ingestion-architecture shift — arguing that for moderate-throughput workloads, SQL-declared materialized views beat operating dataflow clusters. This site's Flink CDC, Debezium, and Bytewax nodes all carry decision framings that trace to this series. Per Apache Flink Alternatives in 2026 (RisingWave) and Debezium Alternatives 2026 (RisingWave).
Connections 6
Outbound 6
enables2alternative_to1solves1Resources 2
The canonical articulation of the "SQL is enough" position — when a streaming database replaces a Flink pipeline, and when it doesn't. Vendor-authored; the decision framework is still the field's reference.
Defines the stream-processor vs streaming-database category split this index uses across its CDC/streaming nodes.