Oracle Database Discussions

Hi all, I need to do a database tuning and need some help from you people. Below is the situation

A DB needs to be tuned, but looks like users doesn't have much database performance issues as such for now, but they said the server's memory utilization is more than what they expected. If only the OS and the database is running, it shows the memory utilization as 70% and if both appserver and webserver are also started in the same server then the memory utilization goes upto 88-90%. So they are afraid that this will affect the db performance in future. Also, they are not ready to have the appserver and webserver in a different server.

I checked the server details and db details and informed them that they have very less memory level (16 GB Ram and CPU - 8 core only) for such a big db, and that's why they see this much memory consumption and this will end up in a resource contention. But they disagreed this and asked me to have a look at the SGA settings and tune it to bring down the memory utilization. I just checked and changed few SGA size parameters, but not much luck. Also researched my level best and tried few db sizing methods, but it didn't help much.

Can you pls help me to resolve this. Below are the details about the server and db.

Server: HP Itanium, 4CPUx2core (8core), 16GB of RAM
OS & DB: HP-UX and Oracle Db 10.2
Application: HCM 9.0
Peopletools version: PT 8.49

1 Physical Architecture - Webserver, Appserver, Db server all in the same box

Also below link has the list of parameters from v$parameters table and the AWR reports that was taken for 2 different days and for 2 different timings. Also one more AWRSQR report for a peoplesoft process that occupied 13% of the CPU at once.

[DB PARAMETERS AND AWR REPORTS|http://fs11n1.sendspace.com/dl/a70aa3c87b7310996f29b469d441247d/4c0be6ff30477352/u9hhin/db_parameters_and_awr_reports.zip]

Pls check the files in above link, and give me your suggestions to bring down the memory utlization or if possible to bring down the SGA size, without affecting the database performance much. This would be a great help. Let me know if you need any more help

Thanks
N