The document discusses MongoDB's aggregation framework. It defines aggregation as gathering data together to perform computations and return computed results. The aggregation framework in MongoDB uses pipelines similar to UNIX pipes to perform aggregation operations like $group, $match, $project, etc. on data. It also supports map-reduce operations and provides connectors to Hadoop. The document provides examples of translating common SQL queries to the aggregation framework and discusses concepts like optimization, restrictions and references for further reading.