Re: Modelling large trees and hierarchies

"Paul Arnold" <paula_at_pivetal.com> wrote in message news:b6ee5aa3.0406162326.204e76cb_at_posting.google.com... > "Dawn M. Wolthuis" <dwolt_at_tincat-group.com> wrote in message news:<caqm68$jel$1_at_news.netins.net>...
> > "Paul Arnold" <paula_at_pivetal.com> wrote in message
> > news:b6ee5aa3.0406161236.6b2e1205_at_posting.google.com...
> > <snip>
> > > Any suggestions/foresight/tips that may help us in the database
> > > modelling would be most appreciated?
> >
> > Are you tied to a SQL-DBMS or open to other options?
> > --dawn

>
> Hi Dawn,
>
> We are pretty much tied to MS SQL Server/Oracle DBs at the moment....
>
> Our previous attempt involved generating highly optimized and
> compressed XML on the server which was then passed across to the
> clients.  This was fine for trees upto 1 million records, but it just
> won't scale past that due to the Microsoft XMLDocument's very
> inefficient memory footprint.
>
> We can also not guarantee the spec of the client machines, so want to
> avoid anything that is highly memory dependant.
>
> Why do you have any neat options that we could consider?

Nothing "way cool" just solid, fast, easy, agile, but old-fashioned and backed by big blue -- IBM UniVerse or UniData. You could keep your XML data model for the most part, and just have a real engine (rather than XML docs) behind it. --dawn Received on Thu Jun 17 2004 - 22:36:04 CEST