What does HBase mean in big data knowledge points?

HBase is a distributed, column-oriented and open source database. This technology comes from the Google paper Bigtable: Distributed Storage System of Structured Data written by Fay Chang. For example, Bigtable used Google? Like the distributed data storage provided by the file system, HBase provides functions similar to Bigtable on Hadoop.

HBase is a sub-project of Hadoop project of Apache, which is different from the general relational database, but a database suitable for unstructured data storage. HBase？ Distributed database has the following remarkable characteristics.

1. Large capacity: Tables in the HBase distributed database can store thousands of rows and columns.

Second, column-oriented: the data of HBase is stored according to a column in the table, and columns can be dynamically added according to the data, and various operations can be performed on the columns independently.

Third, multi-version: the data storage of each column of the table in HBase has multiple versions, and each column corresponds to a piece of data, but some data will correspond to multiple versions.

Sparsity: Because the columns in the HBase data table are allowed to be empty, and empty columns do not occupy storage space, the data table can be designed to be very sparse.

5. Extensibility: Because the underlying layer of HBase depends on HDFS, when the disk space is insufficient, machines (namely DataNode service) can be dynamically added to increase the disk space, thus avoiding data migration like relational database.

Sixth, high reliability: HBase is above HDFS, and HDFS has its own backup mechanism, so when HBase cluster has serious problems, the copy mechanism can ensure that data will not be lost or damaged.