RunJRun: Parallel Processing for Java on EC2 Instances

Customer Apps>RunJRun: Parallel Processing for Java on EC2 Instances
RunJRun is a very simple system for doing parallel processing in Java, using EC2 instances as compute nodes.


Company: Northbranchlogic
Inquiry e-mail address:
Amazon Web Services Used: Amazon EC2
Solution URL:
Audience: Businesses, Developers
Pricing: For a fee
How does this application use Amazon Web Services?: EC2 instances are used as compute nodes of a parallel proccessing network.
Created On: April 10, 2010 3:45 AM GMT
Last Updated: May 6, 2010 7:16 PM GMT
RunJRun is a very simple system for doing parallel processing in Java, using Amazon Elastic Compute Cloud (EC2) instances as compute nodes.

The basic compute unit is a Runnable, Serializable Java object, a 'task' for short. A user submits a list of such tasks to RunJRun. Each task then has its run() method invoked on an EC2 instance. Optionally, the 'cooked' task object can be stored in S3.

RunJRun uses SQS and S3 to make tasks available to compute nodes. These services scale very well, so RunJRun can handle large numbers of tasks and EC2 instances.

A RunJRun compute node executes one run() method at a time, so that a task has exclusive use of an EC2 instance. (To fully utilize EC2 instances, you may want to make your tasks multi-threaded. See Multi-thread tasks in the RunJRun manual).

RunJRun is written in Java and is implemented as an api.

RunJRun itself is free, open-source software. To use it, you'll need an Amazon Machine Image (AMI) that has the RunJRun compute node software installed. Several such AMIs are available. These are DevPay AMIs - you'll be charged $0.03/hour to use them, in addition to Amazon's regular charges. For details, please go to RunJRun AMIs. Or you can construct your own RunJRun AMI.

For succinct instructions on using RunJRun, see Quick Start. For more details, see the manual.

RunJRun links:
©2014, Amazon Web Services, Inc. or its affiliates. All rights reserved.