Bigtable

from Wikipedia, the free encyclopedia
Bigtable

Bigtable logo
Basic data

developer Google Inc.
category Database
cloud.google.com/bigtable

Bigtable is a high-performance database system from the US company Google Inc. Bigtable is one of the predecessors of Spanner .

history

The development of Bigtable began in 2004. Bigtable supports the MapReduce process and is the basis of many Google products, such as Google Maps , Google Books , YouTube and Google Earth . Particular emphasis was placed on scalability and speed. The database is therefore based on a non-relational structure.

technology

A characteristic of data stored in Bigtable is that records are added very frequently, but existing records are very rarely changed. A bigtable basically consists of a large number of lines that are named by a data record identifier . Any number of columns can be defined within a line. In contrast to relational databases, the number of columns for each row can be different. Only the so-called family columns , which contain the same data type (such as a link that refers to a page), must be known when implementing the database, but can contain any number of instances per line.

Every table is multidimensional . Each entry is given a timestamp for versioning . With the help of timestamps, the latest version of the data can always be considered without editing or deleting older data.

example

The provider himself describes in a publication an example of the use of Bigtable by the in-house search engine: For each new website previously scanned by a crawler , a line is created in the corresponding Bigtable. The first column contains the first content (in HTML syntax) of the page. The multidimensionality is created by later versions of the same page and the assigned timestamp. The following columns all come from the same FamilyColumn "anchor" and each contain the URL of a different page that links to the corresponding page. Thus the number of columns is not limited by the provider.

Individual evidence

  1. a b Bigtable: A distributed storage system for structured data (PDF file, 216 KB)