RE: Storing blobs in database vs filesystem

From: Kevin Closson <kevinc_at_polyserve.com>
Date: Thu, 28 Sep 2006 16:31:57 -0700
Message-ID: <5D2570CAFC98974F9B6A759D1C74BAD0E5B215@ex2.ms.polyserve.com>

>>>

>>>This isn't responsive to the OP, but I was hoping you could
>>>help me understand what advantage you people saw in storing
>>>large volumes of blobs in the database at all.

I'll take a stab. Historically, there have not ben options for getting multiple servers access to unstructured data (read CFS). So at least with OPS and RAC, stuffing these, actually quite-unstructured, pieces of data into the database at least got them accessible to multiple nodes. Also, historically, there were no PL/SQL callouts to get unstructured data. These 2 factors are now historical. You can deal with external data nicely with 9i and moreover 10g and CFS and NAS technology is not spooky anymore (well, I think it might be to the open source folks).

The other force was the desire in the mid 90s for Sybase,Informix and Oracle to be *the* repository for all data...there were musings of OODB and RODB and all those things that never really got anywhere. That is, Informix and Oracle wanted to be media servers in addition to databases and that was at a point in history that predated the web, really.

There are a lot of Oracle sites that blend the best of what Oracle does and the best of what Oracle does to support external unstructured data. Gannett Media comes to mind.

--
http://www.freelists.org/webpage/oracle-l

Received on Thu Sep 28 2006 - 18:31:57 CDT