| From: | Andrew Chernow <ac(at)esilo(dot)com> |
|---|---|
| To: | David Fetter <david(at)fetter(dot)org> |
| Cc: | PG Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Extending varlena |
| Date: | 2008-08-18 22:09:13 |
| Message-ID: | 48A9F309.4020109@esilo.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
David Fetter wrote:
> Folks,
>
> As the things stored in databases grow, we're going to start needing
> to think about database objects that 4 bytes of size can't describe.
> People are already storing video in lo and bytea fields. To date, the
> sizes of media files have never trended downward.
>
I always find these requests puzzling. Is it really useful to store the
data for a jpeg, video file or a 10GB tar ball in a database column?
Does anyone actually search for byte sequences within those data streams
(maybe if it were text)? I would think that the metadata is what gets
searched: title, track, name, file times, size, etc... Database storage
is normally pricey, stocked with 15K drives, so wasting that expensive
storage with non-searchable binary blobs doesn't make much sense. Why
not offload the data to a file system with 7200 RPM SATA drives and
store a reference to it in the db? Keep the db more compact and simpler
to manage.
Andrew Chernow
eSilo, LLC
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Alvaro Herrera | 2008-08-18 22:11:36 | Re: pgbench duration option |
| Previous Message | Joshua Drake | 2008-08-18 22:07:38 | PostgreSQL Conference: West - Call for Papers |