Bigtable was a project that was initiated in 2004 to manage both scalability and performance of the data used for various projects at Google. The seminal paper that describes the characteristics of the system was released in 2006 (https://static.googleusercontent.com/media/research.google.com/en//archive/bigtable-osdi06.pdf) titled Bigtable: A Distributed Storage System for Structured Data. In essence, Bigtable was a column-store (more on this later) where each value could be uniquely identified using a row key, a column key, and a timestamp. It was one of the first mainstream databases that epitomized the benefits of storing data in a columnar format rather than using the more common row-based layout. Although columnar databases such as kdb+ and Sybase IQ existed prior to Bigtable, the use of the method by an industry leader to manage petabyte-scale information brought the concept into the limelight.

The official site of Bigtable summarizes the key-value proposition:

Since the introduction of Bigtable, several other NoSQL databases adopted the convention of columnar data layout; most notably HBase and Accumulo, which are both Apache projects.

The Bigtable solution is today available for use at https://cloud.google.com/bigtable/ where it can be purchased on a subscription basis. The fee for smaller amounts of data is quite nominal and reasonable, whereas larger installations would require more extensive implementations.