Received: from gnumed.dhs.org (modem59.malleenet.net.au [203.143.67.123] (may be forged)) by postgresql.org (8.11.3/8.11.4) with SMTP id f9QEKoP17801 for ; Fri, 26 Oct 2001 10:20:51 -0400 (EDT) (envelope-from hherb@malleenet.net.au) Received: (qmail 24011 invoked by uid 8); 26 Oct 2001 14:20:41 -0000 Message-ID: <20011026142041.24010.qmail@gnumed.dhs.org> Received: from pc-00071.e-smith.gnumed.dhs.org (192.168.1.71, claiming to be "there") by bifrost.e-smith.gnumed.dhs.org with SMTP id smtpdJO9jsf; Sat, 27 Oct 2001 00:20:39 EST Content-Type: text/plain; charset="iso-8859-1" From: Horst Herb To: pgsql-sql@postgresql.org Subject: Re: Diferent databases on same query... Date: Sat, 27 Oct 2001 00:24:53 +1000 X-Mailer: KMail [version 1.3.1] References: <22C9D178-C4E4-11D5-BBB1-003065C61C62@mac.com> In-Reply-To: <22C9D178-C4E4-11D5-BBB1-003065C61C62@mac.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Archive-Number: 200110/370 X-Sequence-Number: 5323 On Saturday 20 October 2001 08:53, you wrote: > Andre, > > The only reason to have data for one query in different databases is if > the databases are on different machines. If you're running on PC-type > hardware, you might have to do that some times for performance reasons. > But if at that point, you should ask yourself if you shouldn't invest in > some faster hardware ;-). There are many more good reasons to do so. Example from my own domain, medicine: 1.) one database hold demographic information. This information is needed by administration, nursing, and medical staff as well as by external service providers such as pathology. 2.) another database holds medical information. The information here is highly sensitive. The database admin should be a medical doctor,and it must be unaccessible to clerical staff. SQL permissions are only a *very* weak protection, not good enough for this purpose. For many good reasons, foremost security, this information will not reside in the same database (sometimes not even on the same machine) as the administrative data or the demographic data. 3.) For obvious reasons, it would be bad to replicate the demographic data in order to provide this data to administration, medical staff etc for their own purposes. Thus, we have multiple databases on multiple servers all depending on each other. Some servers are read-only, others need the backend to log every single query for audit trailing. The latter alone would blow the storage needs out of proportion if applied to the whole set of databases. At present, it is up to our client software to sort this out "manually". A minor headache. Horst