Skip to content

SamzaSQL: Streaming SQL implementation on top of Apache Samza and Apache Kafka

Notifications You must be signed in to change notification settings

milinda/samza-sql

Repository files navigation

SamzaSQL: Fast Data Management with Streaming SQL and Apache Samza

SamzaSQL is a scalable and fault-tolerant SQL based streaming query engine implemented on top of Apache Samza with support for interaction with non-streaming data sources. SamzaSQL define a minimal set of extensions to standard SQL to enable fast data querying and data manipulation. Queries expressed in streaming SQL are compiled into physical plans that are scalable, fault-tolerant and executed on Apache Samza. SamzaSQL front-end, which does query parsing, validation and planning is built on top of Apache Calcite. SamzaSQL is architectured to support variuos data formats such as Avro or JSON and variuos traditional relational and NoSQL stores using plugable extensions. SamzaSQL provides a window operator implementation that provides timely and deterministic window output under unpredictable message delivery latencies, and deterministic window output with node failures and message re-delivery.

Please note that this repository contains a copy of streaming SQL related components from Samza's samza-sql branch in addition to the features I added over them. I am planning to contribute new features and improvements back to Samza once this project become stable.

License

Apache License 2.0

About

SamzaSQL: Streaming SQL implementation on top of Apache Samza and Apache Kafka

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published