Technology

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.

6 connections 2 resources 1 post

Summary

What it is

A PostgreSQL-compatible streaming SQL database — continuously-maintained materialized views over streaming/CDC sources, served from the same system, with Iceberg-on-S3 output.

Where it fits

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.

Misconceptions / Traps
  • 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.
Key Connections
  • alternative_to Apache Flink — for SQL-shaped streaming workloads
  • enables CDC into Lakehouse — CDC sources to Iceberg without Kafka
  • solves Legacy Ingestion Bottlenecks — one system instead of pipeline + serving layer

Definition

What it is

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.

Why it exists

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.

Primary use cases

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

Latest signals
  • 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

Resources 2

Featured in