September 2017
Beginner to intermediate
360 pages
8h 13m
English
These are the variables that the programmer intends to share to all execution units throughout the cluster. Though they sound very simple to work with, there are a few aspects the programmers need to be cognizant of while working with broadcast variables: they need to be able to fit in the memory of each node in the cluster—they are like local read-only dictionary/index for each node, thus they can't be huge in size, and all nodes share same values thus they are read-only by design. Say, for instance, we have a dictionary for spell check, we would want each node to have the same copy.
So to summarize, here are the major caveats/features of the design and usage of broadcast variables:
Read now
Unlock full access