Mongo Scaling Demo
==================

This is a sample algorithm to demonstrate one way to create an algorithm that:
  * scales horizontally
  * is ephemeral
  * uses MongoDB as a persistent data store

This demo was created to suppliment the talk "Designing Algorithms That Scale
Horizontally With MongoDB", presented at MongoRaleigh and MongoDallas.


GETTING STARTED:
================

1.  Launch MongoDB locally.
2.  Run "bundle install" (after installing ruby, rubygems, and the bundler gem)
3.  Run "ruby consumer.rb" in a terminal
4.  Run "ruby replayer.rb 10" in another terminal (10 can be changed.  It's the
       number of weather samples to be written per second.  Increase this and
       the number of consumers you run to see how the consumer algorithm scales
       with increased load)
5.  Run as many consumers in seprate terminals as you want.

You can create and destroy consumer processes at will.  This demonstrates the
ephemeral nature of this algorithm, and you can also see that as you scale it
up, each consumer does less work.