Comparing the performance of DB's with and without DB Links

Question

Hi Experts
I m using Oracle 10G. I have two databases in which the first one uses DB Links to access the other schema objects and the second one doesnot use DB links to access the schema objects instead the permission is granted users to access the other schema objects.
I would like to compare the performance of these two databases which uses DB links and Which doesnot use DB Links. Kindly Suggest How can i achieve this?
Thanks a lot in advance,
Ram

Answer 1

user448837 wrote:
Hi Justin,
I m new to this Organization and the OLD DBA's had done that and it was integrated to the older Version of the Product. After me coming in, We had a discussion and changed it which is the second DB. And in the comparison Front, Its Easy to Just RUN a query and Explain plan it with link and without link. My Real Question is that, is there any other way to measure the Comparison?.It's certainly possible that I'm missing something here, but isn't measuring execution time a pretty good metric here? It seems unlikely that an organization that thought it was a good idea to set up a loopback database link would be particularly interested in the relatively subtle changes you'd see in before-and-after AWR or ASH reports. And it seems unlikely to me that the performance problems that the loopback link created (if it did create any-- it's entirely possible that there is no performance difference at least for the queries your application is running) would be the subtle sort that you'd want an AWR/ ASH report for. It seems much more likely that the presence of a database link caused poor query plans or forced a lot of data to get sent "over the network" (though it presumably never hit the wire). Either of those should cause easily measurable execution differences.
My wager is that you'll get a lot further with simpler metrics (i.e. previously, function X in application Y took 15 seconds, now it takes 5 seconds or previously, we could process X function Y calls per minute, now we can process 2x).
Justin