Please enable JavaScript.
Coggle requires JavaScript to display documents.
Spark - Coggle Diagram
Spark
Spark vs hadoop
Storage
Spark has no storage
Mapreduce
Yes for both
Speed
spark 10 to 100 X
Resource management
Stand alone spark
Hadoop Yarn
RDD
its immutable
Cannot change the values
new RDD have to be made
Resilient
Distributed
Dataset
Other
To run spark hdfs required? Ans No
spark uses hadoop cluster
Doesnt write to hard disk
No back on multiple nodes
Lienage, keeps in memory the operation
Lieanage
Dependency
Storing in variables