TABLE I.

HBase architecture key concepts summary

Concept	Comment
Table/HTable	A collection of related data with a column-based format within HBase.
Region	HBase Tables are divided horizontally by row key range into “Regions.” A region contains all rows in the table between the region’s start key and end key.
Store	Data storage unit of HBase region.
HFile/Storefiles	The unit of Store, which is collocated with a Hadoop datanode and stored on HDFS.
memStore	When write data is uploaded to a HTable, it is initially saved in a cache as memStore. Once the cache size exceeds a pre-defined threshold, the memStore is flushed to HDFS and saved as HFile.
HMaster	HBase cluster master to monitor a RegionServer’s behavior for load balancing.Table operator. e.g., create,delete and update a table.
Regionserver	Serves read/write I/O of all regions in a cluster node. When Regionservers collocate with Hadoop datanode, it can achieve data locality. Subsequently, most reads are served by the RegionServer from the local disk and memory cache, and short circuit reads are enabled.
Rowkey	A unique identifier of a row record in table.
Column family	Columns in Apache HBase are grouped into column families.
Column identifier	The member in column family, also called as column qualifier. Multiple column identifiers can be used within one column family.