Not like Spark, Flink is mainly designed to deal with stateful data set(batch) and data stream vary in different use cases. Besides I personally think its hierarchical api system is friendly to us.

This short article includes its Computaion Model, Cluster Architecture, Api Category.

Not like Flink which is short and sharp, Spark acts as huge eco-system which covers from rpc, data storage to computation, scheduler…

I am ganna talk about Spark Core, Spark SQL, Spark Streaming instead of all its features at very short length.