In Apache HBase data is stored in a table but a table is not the fundamental unit. HBase is designed for big data and a single table would be unwieldy to store big data. A table is sorted by a row key lexicographically. To store data of scale a table is made up of one or more regions; in fact, it’s typically several regions spread across several RegionServers. A region is a subset of a table’s data. When a table is created, it consists of only one region by default. The whole capacity of the cluster could be said to be underutilized when data is first loaded into an empty table because the data is sent to a single RegionServer.