January 2019
Beginner to intermediate
154 pages
4h 31m
English
Instead of processing each record or event at a time, Spark receiver's receive data in parallel and keep it in a buffer of Spark worker nodes. Then, Spark's engine runs tasks over these discretized streams, also called microbatches.
A microbatch is created based on a time window, instead of a number of messages.
Read now
Unlock full access