Storm topology
Streams can be partitioned among bolts by using stream grouping, which allows the streams to be routed towards a bolt. Storm provides the following built-in stream groupings, and you can implement a custom stream grouping by implementing the interface:
- Shuffle grouping: Each bolt is configured uniformly to get an almost equal number of tuples
- Fields grouping: Grouping on a particular field is possible to consolidate the tuples of the same field value and different value tuples to different bolts
- All grouping: Each tuple can be sent to all the bolts but can increase the overhead
- Global grouping: All the tuples go to a single bolt
- Direct grouping: The producer can decide which tuples to be sent to which bolt
Get Hadoop Essentials now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
