Re: scaling a database

----
What I am wondering is will we reach some kind of limit if we wanted to
grow the database to accomodate 1 million or even 10 million users? Right
now we have a dedicated dual processor server with 1 GB of RAM that
PostgreSQL runs exclusively on. What should we expect to cause problems as
the database size grows? What are the standard solutions? Would you
generally throw more hardware at the problem? Would you partition the
database into multiple instances each with a subset of the users and their
data? Any information that someone can point me to would be greatly
appreciated.
------

Well, it all depends on your definitions. If you're expecting 1
million or 10 million *simultaneous* users, you're going to be
hard-pressed to accomplish that with ANY DBMS, and you're going to have to
throw fistfuls of hardware at it. If each simultaneous was responsible
for one query per minute, you're still talking about 16,000 to 160,000
queries per second, which is quite a bit.

If you're referring to a user as someone that has a username/password
in the database, and could potentially execute a query, then you shouldn't
have any problem with 10 million of them. PostgreSQL can handle very
large databases and tables efficiently. In that case, the capacity of the
server would depend on how many of those users would actually be logged in
and issuing queries at any one time, and the nature of those queries.

To give much more detailed advice, we would really need a more
detailed description of your data and server usage.

steve