Real-Time Big Data Analytics

Storm internals

The moment people start talking about Storm, a few key aspects of this framework stand apart:

• Storm parallelism
• Storm internal message processing

Now, let's pick each of these attributes and figure out how Storm is able to deliver these capabilities.

Storm parallelism

If we want to enlist the processes that thrive within a Storm cluster, the following are key components to be tracked:

• Worker process: These are the processes executing on the supervisor node and process a subset of the topology. Each worker process executes in its own JVM. The number of workers allocated to a topology can be specified in the topology builder template and is applicable at the time of topology submission.
• Executors: These are the threads that are spawned ...