It makes sense doesn't it?  Google taking some of its server farms and building a 'Big Data' service along the lines of Google App Engine or Amazon Elastic Map Reduce.   Except the main function for this new service, named BigQuery, is able to take huge data sets, have Google run SQL queries over them in seconds (for billions of rows if you need it) and spit out results.  It's based on Google's internal technology called Dremel, which is a scalable, interactive ad-hoc query system for analysis of read-only nested data.

BigQuery was released to the public this week after a limited preview in November. 

Since announcing BigQuery in limited preview last November, many businesses and developers have started using it for real-time Big Data analytics in the cloud. Claritics, a social and mobile analytics company, built a web application for game developers to gain real-time insights into user behavior. Crystalloids, an Amsterdam-based analytics firm, built a cloud-based application to help a resort network analyze customer reservations, optimize marketing and maximize revenue. This just scratches the surface of use cases for BigQuery.

-- Google Devloper Blog

To summarize BigQuery's features:

• Upload huge data sets (billions of rows possible) and Google will run SQL queries over them in seconds
• No need to to specify/build any indexes
• Accessible via a simple UI or REST interface
• Based on technology called Dremel

You can sign up for BigQuery now and try it out for free (under 100GB per month) and test it with a variety of datasets.  The full pricing plan is available now.