IDS 200 Lecture Notes - Lecture 24: Bigquery, Bigtable, Mapreduce
12 Feb 2020
School
Department
Course
Professor
Document Summary
F1 is google"s database for storing ads. Highly consistent despite being a large-scale distributed system. Like nosql databases, scalable to run on many instances of standard hardware. Reading should not materially interfere will search results speed. Based on spanner and bigtable (more on bigtable later) A system allowing fast queries over very large tables. For other businesses, using it is vastly cheaper than building a similar proprietary system from scratch. One table is large (even very, very large), one is small: Distribute copies of entire small table with slices of large table. Make a super-large table containing all fields of both. But, eventually, the super-large tables do become too big. Bigquery can process a lot of data very quickly. Spanner is good for a lot of frequent but relatively simple database operations, like transactions. Bigquery is good for relatively infrequent but truly massive database operations. Data warehouse: the entire store of data belonging to a business organization.
