Re: Are we losing momentum? - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Are we losing momentum?
Date
Msg-id 10857.1050448704@sss.pgh.pa.us
Whole thread Raw
In response to Re: Are we losing momentum?  ("Robert E. Bruccoleri" <bruc@stone.congenomics.com>)
Responses Re: Are we losing momentum?  (ow <oneway_111@yahoo.com>)
Re: Are we losing momentum?  (Sailesh Krishnamurthy <sailesh@cs.berkeley.edu>)
Re: Are we losing momentum?  ("Robert E. Bruccoleri" <bruc@stone.congenomics.com>)
List pgsql-hackers
"Robert E. Bruccoleri" <bruc@stone.congenomics.com> writes:
> I beg to differ with Tom Lane's opinion, but schemas do not solve the
> problem with multi-database queries because of the following reasons:

> 1) When dealing with large databases, the use of multiple databases
> reduces the risk of wiping out all the data, and reduces the recovery
> time in case of accidents.

> 2) Multiple databases allow for different management policies on each
> database, whereas schemas require some consistency across them all.
> In a heterogeneous working environment, this is a signficant issue.

> 3) PostgreSQL should strive for heterogeneous multi-database queries,
> so that applications currently using other systems could be slowly
> migrated to PostgreSQL by moving portions of a database from other
> vendors to PostgreSQL. In my work, the lack of PostgreSQL - Oracle
> connectivity is a disabling impediment to wider PostgreSQL usage.

Please keep in mind that I was replying to a poster who said "cross-db
queries on the same server (meaning same postmaster, for our purposes)
are trivial; why hasn't Postgres got them when everybody else does?"

Your above arguments are all good ones, but they presume a scenario that
is much different and *MUCH* harder to implement than local "cross
database" queries.  My point is that schemas solve the same-server
problems that the original poster was interested in.  I did not say,
nor mean, that there is no need for cross-server queries.  But that is
a different problem.  Today we can only offer dblink; maybe someday
SQL-MED.
        regards, tom lane



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: First draft of new FE/BE protocol spec posted for comments
Next
From: Tom Lane
Date:
Subject: Re: Foreign Database Connectivity