Data Analytics

Kinesis Data Analytics reads and processes streaming data in real-time using SQL. It can ingest data from both Streams and Firehose. The output can then be written to destinations including S3 and Redshift.


  • Sources provide the streaming data.
  • Pumps define pipelines which filter or transform data.
  • Destinations define where the output stream should go.
  • Runtimes provide a means of transforming and delivering messages on the stream.


SQL has been supported since 2016.

Flink has been supported since 2019. It uses different terminology:

  • Sinks are destinations.
  • Operators are pumps.

Both can be written in Java or Scala.