Single node parallelism with PL/Proxy?

I have seen some suggest that single node parallelism is possible, but all of the material I have found so far deals with parallelizing across nodes. I am interested in getting large static queries to exercise multiple cores when a query maximizes on a single thread.

In a DWH environment, I would like to use PL/Proxy to do the following:

• develop functions that query a local database on a single node

• individually querying partitions

• aggregating the intermediate data sets into a single result

Is this possible?

Sure it's possible: check this blogpost http://www.depesz.com/index.php/2011/12/02/the-secret-ingredient-in-the-webscale-sauce/

StackExchange Database Administrators Q#9097, answer score: 3