Explore the powerful capabilities of Storm, an open-source distributed realtime computation system, in this comprehensive 57-minute conference talk. Learn how Storm simplifies writing and scaling complex realtime computations on computer clusters, processing millions of messages per second with guaranteed delivery. Discover the fundamental concepts of Storm, including streams, spouts, bolts, and topologies. Gain practical knowledge on developing and testing topologies in local mode, deploying them on Storm clusters, and understanding fault-tolerance mechanisms. Delve into advanced topics such as Distributed RPC for parallel processing and transactional topologies for idempotent computations. Through examples of production usage, grasp Storm's potential in solving various realtime computation problems, from stream processing to continuous computation and distributed RPC.
Storm: Realtime Distributed Computing with Nathan Marz - ETE 2012
Overview
Syllabus
Intro
Basic info
Before Storm
Scaling
Problems
What we want
Use cases
Storm Cluster
Starting a topology
Killing a topology
Concepts
Streams
Spouts
Spout examples
Bolts
Tasks
Task execution
Stream grouping
Streaming word count
Distributed RPC
DRPC Example
Computing reach
Reach topology
Guaranteeing message processing
Transactional topologies
Example #2
Storm UI
Storm on EC2
Starter code
Documentation
Ecosystem
Questions?
Taught by
ChariotSolutions
