5 minutes for 800 calls is too much. We have views with thousands of entries that take no more than few seconds.
However we took some precautions. For example there is NO calculated field in the views (like to Deadline). Also we try to keep fields of linked entities to service calls to minimum on this views. The less joins and calculations the quicker the views.
The bottleneck can be in memory\disk subsystem. If there is not enough memory, the hard faults often take place. (memory pages are read from swap to memory and vice versa). Watch your HDD led. Use perfmon.msc to watch "average disk queue lehgth" counter (it must not be more then 2). How much time does it takes to execute "select * from itsm_servicecalls" running this query while you are connected to oracle from servicedesk server, and compare the time while running this query locally on oracle server. You can also start any network packet analyzer and watch sql traffic - remeber the time, when the packet with you query entered network media, and when reply was returned. IF there is some file transfer service on your aix (samba, ftp) try to benchmark network speed while copying files to\from oracle server. Check for duplex\speed settings of your SD server NIC, SWITCH, look for any errors\collisions on a switch. Turn off antivirus on SD server (only for test time) if there is any..