Implement robust continuous applications that never stop and get immediate insights from your data.
Write latency-critical applications with millisecond responses.
Flink can handle millions of events per second in moderate-sized and scale to 1000s of nodes.
Flink is highly available and fault tolerant; the results of your computation will be correct after failures.
Correct stream handling
Flink embraces the notion of event time, guaranteeing that out of order events are handled correctly.
Flink has full batch processing capabilities by treating batch as a special case of streaming.
APIs and libraries
Choose Flink’s own DataStream, DataSet, SQL, CEP, Gelly, and FlinkML APIs, or use compatibility layers for MapReduce, Storm, Cascading, and Beam.
Batteries included; Flink comes with support for Kafka, HDFS, HBase, Kinesis, S3, RabbitMQ, Elastic, Cassandra and YARN.
With over 190 contributors and an impressive list of users, Flink is one of the most active Big Data projects in the ASF.