On Tue, 2006-10-17 at 16:51 -0700, Anon Mous wrote:
> Hi
>
> I've loved reading all of your thoughts and comments.
>
> Yet, I'm left with the question:
>
> Can we can brainstorm a caching solution that is workable...
>
I think you're making this a little complicated.
A lot of these problems can be solved with something like materialized
views (triggers updating another smaller relation), or they are already
solved by the lower layers of caching (like PostgreSQL's shared buffers,
or the OS buffer cache).
If the application is executing many queries that are exactly the same,
it would be a good idea to look at something like pgpool-II's query
cache. The only way to gain a benefit on querying results that are
already in memory is to avoid the query processing.
Regards,Jeff Davis
