Getting Started with Your First Topology - Real-time Analytics with Storm and Cassandra - page 25

Bolts
Bolts are the processing units of a topology. They are the components of the topology that
perform one or more of the following tasks:
• Parsing
• Transformation
• Aggregation
• Joins
• Database interaction
The entire process being performed by the topology is generally divided into smaller tasks
and subtasks, each preferably performed by a different bolt to exploit the power of the par-
allel distributed processing of Storm.
Let's look at the following figure that captures a real-time use case where the location co-
ordinates from various airplanes are tracked and processed to ascertain whether they are
moving on the correct trajectory:
Here, the flight location coordinates are sent by sensors in the plane, which are collated at
log servers and fed into a Storm topology. The Storm topology is broken into the following
bolts that can act on the tuples emitted by the spout:
• The parse event bolt : This bolt filters and transforms the event emitted by the
spout. It converts the information into a decipherable format.
• The location bolt : This is the bolt that extracts the location coordinates from the
tuples it receives from the parse bolt and then sends them across to the next bolt.
• The verify bolt : This is the bolt that verifies the location coordinates sent by the
location bolt against the predefined trajectory of the plane, and if it detects devi-
ation, it sends a tuple to the alert bolt.