This topic describes the major features of Blink 3.2.1 and its compatibility with the DataStream API and SQL.
Major featuresBlink 3.2.1 is the first official version based on the Blink open source code. Blink 3.2.1 offers the following major features:
- Job automatic scaling
In Blink 3.2.1, you can improve job performance by using the automatic optimization feature that is based on automatic configuration and automatic scaling. The automatic optimization feature dynamically adjusts the parallelism settings and resources of each operator based on the job running status and incoming data amount. This allows you to shorten job delays. In Blink 3.2.1, this feature is in public review.
- Support for the DataStream API
- Blink 3.2.1 supports the DataStream API. Blink 3.2.1 is developed based on the open source Flink 1.5.1 branch. For information about the DataStream API compatibility between Blink 3.2.1 and Flink 1.5.1, see API compatibility report: Blink 3.2 and Flink 1.5.1.
- Connectors newly supported by the DataStream API
Connector Source Sink Kafka Supported Supported ApsaraDB for Hbase Not supported JDBC ApsaraDB RDS for MySQL Elasticsearch ApsaraDB for MongoDB
- Connectors newly supported for SQL jobs
Blink 3.2.1 adds the support for the following connectors to Blink 2.0.
Connector Type Elasticsearch Stores dimension tables. ApsaraDB for MongoDB Stores result tables. ApsaraDB for Redis Stores dimension tables. ApsaraDB for Redis Stores result tables. Alibaba Cloud Message Queue for Apache RocketMQ V4.2.0 Stores source tables. SQL Server Stores result tables.
- For information about DataStream API compatibility, seeAPI compatibility report: Blink 3.2 and Flink 1.5.1.
- For information about SQL compatibility, seeIncompatible SQL items: Blink 3.0 and Blink 2.0.