Skip to content

Implementation docs

nathanmarz edited this page Mar 29, 2012 · 7 revisions

This section of the wiki is dedicated to explaining how Storm is implemented. You should have a good grasp of how to use Storm before reading these sections.

  • Structure of the codebase
    • Thrift
    • Java interfaces
    • Clojure implementation
  • Cluster state
    • Metadata in Zookeeper
    • Code and configs managed separately
  • Passing messages around
    • one "receiver port" (virtual port)
    • Kryo serialization
    • worker send thread
    • ZeroMQ
  • Acking framework implementation

Clone this wiki locally