From: | "Thomas G(dot) Lockhart" <lockhart(at)alumni(dot)caltech(dot)edu> |
---|---|
To: | M(dot)Boekhold(at)et(dot)tudelft(dot)nl |
Cc: | Armin Schloesser <armin(at)ap-kas(dot)ie(dot)philips(dot)com>, Bruce Momjian <maillist(at)candle(dot)pha(dot)pa(dot)us>, pgsql-hackers(at)postgreSQL(dot)org |
Subject: | Re: [HACKERS] PostgreSQL Backend as SW Gateway to Oracle |
Date: | 1998-07-10 05:37:32 |
Message-ID: | 35A5A89C.23BC02AC@alumni.caltech.edu |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
> > in getting simultaneous multiple db access within Postgres...
> > thinking of trying to implement this as a Postgres "master database"
> > with hooks deeper in the backend to call out to a remote database as
> > a separate session. Sort of like Ingres implemented their
> > distributed databases. Haven't done anything with it though...
> Cewl... wouldn't this enable us to run PostgreSQL on beowolf-like
> clusters? :)
Well, no. I haven't coded on a beowolf system (and my Linux Journal with
a writeup on it has gone wandering :( but a beowolf must be a MIMD
system with (perhaps) a shared file system. So, we would need a
medium-grained or coarse-grained decomposition of the backend to
distribute a single session across a cluster.
However, what I proposed would allow a single database, or parts of a
single logical database, to reside on one host, with access from a
client hitting multiple hosts to find all the tables, so one could
distribute the load if several pieces or many databases were involved.
Doesn't need to be a beowolf, just a networked set of servers.
- Tom
From | Date | Subject | |
---|---|---|---|
Next Message | David Maclean | 1998-07-10 06:18:59 | Constraints in PostgreSQL |
Previous Message | Dr. Michael Meskes | 1998-07-09 22:10:50 | Leaving for vacation |