[Prev][Next][Index][Thread]
cc:Mail UUCPLINK 2.0 Undeliverable Message
User mutual!OCARROLL at TMC_HH@mutual.co.uk is not defined
Original text follows
-----------------------------------------
Received: by ccmail
Received: from uupsi by co.uk (UUPC/extended 1.11) with UUCP;
Wed, 18 Sep 1996 00:24:37 BST
Received: from www19.w3.org by sys4.cambridge.uk.psi.net (8.7.5/SMI-5.5-UKPSINet)
id AAA07928; Wed, 18 Sep 1996 00:23:10 +0100 (BST)
Received: by www19.w3.org (8.6.12/8.6.12) id TAA24518; Tue, 17 Sep 1996 19:21:26 -0400
Resent-Date: Tue, 17 Sep 1996 19:21:26 -0400
Resent-Message-Id: <199609172321.TAA24518@www19.w3.org>
Date: Tue, 17 Sep 1996 19:19:31 EST
From: Tim Mueller -- 610-692-9519 <mueller@silutions.com>
X-ccAdmin: mutual@uupsi
To: www-rdb@w3.org
CC: otisg@panther.middlebury.edu
Message-ID: <009A8855.9D947256.8@pobox.silutions.com>
Subject: re: a question (TIPS for nothing, BIPS for free)
X-List-URL: http://www.w3.org/pub/WWW/Archives/Public/www-rdb/
Resent-From: www-rdb@w3.org
X-ccAdmin: mutual@uupsi
X-Mailing-List: <www-rdb@w3.org> archive/latest/281
X-Loop: www-rdb@w3.org
Sender: www-rdb-request@w3.org
Resent-Sender: www-rdb-request@w3.org
Precedence: list
Boy, you don't want much for free, do you? ;)
> now, question - we have a huge database (well, just the data, not the software
> yet) of about 1-10 or possibly even more records, with about 10+
> columns/fields for each row/record.
Hmm, "big" is a relative term. Big to Mom"s Diner is not the same as BIG
to the Census Bureau (Ever seen a terrabyte of spinning data? Or was that
terrorbyte?) Are we talking 1000's, 100,000's or millions of records? In
database-speak, what's the cardinality? And how many bytes are the
records? Are all the text fields always filled, or might we find only a
few character in a loooong field?
>
> We would like to put that database on-line and make it accessible from the
> Web. We have a public domain database right now, but it appears that that
> software isn't fast enough for such a huge database, and doesn't have all the
> features that we need.
> I was wondering if there are any other relatively cheap or even better, free
> software packages that would be able to handle such big databases and lots of
> simultaneous users and run on UNIX.
Keep in mind that you have to play client-server style games when you start
talking about "lots of simultaneous users" and have a fairly compact
metadata model. There's a contention and concurrency issue lurking under
that rock, which is why things like TP monitors were invented.
> We are looking for something that supports:
>
> - authentication
> - client/server
> - _fast_ queries
> - ability to update/delete/add entries on-the-fly, so to speak, by talking to
> the server.
> - popular and supported query language such as SQL
All of the above is a pretty basic description of a commercial RDBMS,
although _fast_ is something that starts more barroom debates than religion
or politics. If it's a really huge database, you need to design it and
tune it and monitor it. That usually translates into $$.
> - preferably made for using via WWW...
This is a little more specialized, and not really in the direct purview of
the RDBMS. It's more of an add-on tool, whether it comes from the RDBMS
vendor, a third party or MidNite Programs R Us.
>
> I don't know if I'm making much sense and using proper terminology, but hope
> you get the point...
>
> Thanks !
>
> Otis
-------------------------------------------------------------------------------
Tim Mueller mueller@silutions.com
Silutions, Inc voice: +1 610 692 9519
322 N Matlack St fax: +1 610 696 8286
West Chester, PA 19380 http://www.silutions.com/
-------------------------------------------------------------------------------
"You don't have to agree with people to defend them from injustice."
-- Lillian Helman
-------------------------------------------------------------------------------