Duplicating tables in a distributed database
Date: Wed, 10 Jan 2001 19:50:17 +0200
Message-ID: <3a5ca479.0_at_news1.mweb.co.za>
I have a distributed database (100+) physical databases linked over a "low
speed WAN". One of the problems we have is that for a specific type of
transaction we will require the data to be live a specific database (linked
to a specific entity), and in another transaction we would like the data to
be live in another database (related to another entity). The reason for the
locality of the data being an issue is a performance issue running a
transaction accross different databases. One of the solution we have is to
keep two sets of the data, each in its own table (with its own name)
close to the related entity. The inserts to the tables are relativly cheap
(non-complex) and the transactions that use the
data are optimum as they use data localy (from the same database). The only
issue we have with this is that we are
keeping two sets of the same not, something I suppose you shouldn't do in a
relational database.
My question is if this is done (as normal practice in a distributed db), or is there maybe another way to solve this problem.
Any help will be appreciated
Thanks
Leon
Received on Wed Jan 10 2001 - 18:50:17 CET