WebMapReduce on Amazon EC2

In addition to standalone installations, WebMapReduce can be deployed on Amazon's Elastic Compute Cloud (EC2) service. This option is especially appealing if you would like to try out WebMapReduce without much effort. It can even be a cost-effective way to demonstrate map-reduce computing in a class without having to purchase and maintain your own cluster.

Amazon charges fees for EC2 based on usage: what type of machines (or "instances" in EC2 terminology) are running, how long they run, and how much storage and bandwidth they consume. Running a single machine with the "Small" instance type for 5 hours should only cost around $2. Running a 5-node cluster of "High-CPU Medium" instances for 3 weeks could cost around $450. For details, check Amazon's pricing page and its cost estimator.

Tip: Instance type and running time typically account for most of an EC2 bill. Storage and network-related fees are usually negligible.

Getting Started

  1. Sign up for and familiarize yourself with EC2 by reading Amazon's EC2 Getting Started Guide.
  2. Follow our WebMapReduce EC2 Guide to set up a cluster.
  3. Start using WebMapReduce by following our User Guide.

Available AMIs

These are the WebMapReduce AMIs currently available:

AMI ID WebMapReduce Hadoop OS Instance Types
ami-83ef26ea 1.0.2 Cloudera CDH2
(based on Hadoop 0.20.1)
Ubuntu 10.04 Server 32-bit:
  • Small
  • High-CPU Medium