This is optimized for bulk analysis. RDS is more optimized for low latency queries. This is because Redshift is column oriented: http://en.wikipedia.org/wiki/Column-oriented_DBMS .

Yep. Redshift is based off of ParAccel. We used that DB for some really awesome large scale analytics on a previous project.

http://www.zdnet.com/amazon-redshift-paraccel-in-costly-appl...

This seamlessly integrates with existing Data Warehouse software solutions and Redshift handles everything you'd normally had to configure, import, read, query yourself using RDS; plus it uses S3 to store its backups and let the client restore their datasets from it. So it doesn't compete with it, my understanding is that it might as well utilize both but in a ready made package that integrates with existing data warehousing solutions.

Amazon RDS is for OLTP workloads. Redshift is a distributed, column-oriented store that's designed for OLAP workloads. For more info: http://aws.amazon.com/redshift/faqs/#0110

Both RedShift and Google's BigQuery are very useful for enterprise applications that involve aggregating data in many ways.