The Architecture of HBase

HBase is column-oriented by design, where HBase tables are stored in ColumnFamilies and each ColumnFamily can have multiple columns. A ColumnFamily's data are stored in multiple files in multiple Regions where a Region holds the data for a particular range of row keys. To manage Regions, MasterServer assigns multiple Regions to a RegionServer. The flexibility in the design of HBase is due to the flexible RegionServers and Regions, and is controlled by a single MasterServer. HBase Architecture uses Zookeeper to manage the coordination and resource management aspects which are needed to be highly available in a distributed environment. Data management in HBase is efficiently carried out by the splitting and compaction processes ...

Get Hadoop Essentials now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.