An increasing prevalent new family of stores are column family stores, which partition rows so that a table can stride across multiple machines. On each machine, the row data is structured as a multidimensional sorted map. The distribution helps in scaling the store to a large amount of data, while the sorted attribute helps in doing things such as range scans. This design was first promulgated by the Google BigTable team (https://ai.google/research/pubs/pub27898).
In the following sections, we will take a detailed look at an example: Cassandra.
