Google BigQuery is a cloud-based big data analytics web service for processing very large read-only data sets.
BigQuery was designed for analyzing data on the order of billions of rows, using a SQL-like syntax. It runs on the Google Cloud Storage infrastructure and can be accessed with a REST-oriented application program interface (API).
BigQuery, which was released as V2 in 2011, is what Google calls an "externalized version" of its home-brewed Dremel query service software. Dremel and BigQuery employ columnar storage for fast data scanning and a tree architecture for dispatching queries and aggregating results across huge computer clusters.
BigQuery in its Dremel form has been used inside Google to track device installation data, create crash reports and analyze spam. Since its inception, BigQuery features have continually been improved. In early 2013, data joins and time stamps were added to the service. Later in the year, stream data insert capabilities were added.
Mike Mason explains how Google BigQuery works in ten minutes.
See also: columnar database