Amazon Kinesis Analytics lets you use normal SQL to create queries that run on data that is constantly updating and arriving in massive volumes. The service is linked to Amazon's other Kinesis services: Kinesis Streams and Kinesis Firehose, which provide ways to send data into Amazon's cloud. The output from a Kinesis Analytics stream can be routed to other Amazon services such as Amazon S3, Redshift, and Elasticsearch Service, or sent to your own custom endpoint.

The idea is that you can use it on data sources that change rapidly in real time such as stock market data or web site analysis. The advantage it offers is that a lot more developers are competent in SQL than will ever learn more specialist languages.

The underlying technology behind Kinesis Analytics comes from SQLstream Blaze. This is a stream processing suite that offers real-time information from the integration, analysis and visualization of high volume, high velocity data. Amazon has licensed a subset of the core technologies of SQLstream Blaze.

In Kinesis Analytics, queries run in a processing window that takes a subset of the data coming in. You get a choice of three types of window - for a specific time period; sliding windows that give you the most recent set of data of a specific size; and custom windows when you can't really group on time.