It depends on how you use the db link -- and what you're comparing it to.
A db link is, in it's simplest use, just a database to database connection.
There is some overhead to manage the distributed transaction, but if you're
pulling data over the db link then there really is no distributed
transaction -- that is, no need to do commits or rollbacks on the remote
node. Anyway, your main concern will be to keep the data movement over the
link efficient. If you do stuff like pull a row or two at a time within a
loop (a common performance problem) instead of selecting the set in a single
SQL statement, or do a lot of distributed joins, you will see performance
problems compared to a non-distributed system.
--
----------------------------------------
Mark C. Stock
www.enquery.com
(888) 512-2048
"Arne Bippes" <sceddie_at_gmx.de> wrote in message
news:ddad38c.0310290511.793e4f11_at_posting.google.com...
> Hi,
>
> we are thinking about using Oracles DB_link feature to connect two
> Databases.
> Currently the data is transported via MQSeries.
>
> Does anybody have any experience about the Performance. Do I have to
> expect increased or decreased Performance?
>
> Greetings
>
> Arne Bippes
Received on Wed Oct 29 2003 - 07:47:34 CST
