Anyone here knows how Google manages it's DB?
I bet they created their own database system but how do they manage the indexDB? It has to be terabytes of data, do they split it in many smaller DBs and run querys on every singles one while searching or is it really a super huge db?
They store everything in one huge text file
They used to do it like this
https://en.wikipedia.org/wiki/Google_File_System
Don't they use Hadoop or something?
>>58241797
They use spanner. Many nosql databases around the globe synchronized with atomic clocks. The paper describing it has been published.
>>58241951
Hadoop is not a database. They used to use big table(HBase) but not anymore.
>>58241797
Look for white papers
https://research.google.com/pubs/papers.html
This might give you some insight
>>58242265
>Hadoop is not a database.
I assumed it was because I always hear about it in relation to massive data amounts. What is Hadoop then? some sort of query tool?
>>58242350
a distributed filesystem + an implementation of mapreduce
>>58241797
Google has several different kinds of databases.
>BigTable (key/value)
>Megastore (BigTable+ transactions/schema)
>Spanner (distributed sql-like)
>Colossus (files)
>Dremel (sql-like querying over colossus files)
>Chubby (config)
>Piper (source control)
>MySQL here and there
Are you asking about the Web index specially?