You're looking at the average to write a particular file. They are *probably* looking at the average service time of all requests. SAME layouts rarely exhibit hot spots or overall average bad service statistics. Cary wrote a paper about it, and I like to keep in mind the old joke about the guy with one foot in boiling water and the other foot in a block of ice. His average of 122 degrees F was a comfortable hot tub! Regards, Mark W. Farnham PS: BORING tends to make it easier to see if you have an actual problem with i/o, but is rarely as flat in i/o profile as SAME. Of course flat is not the same as fast, now is it? -----Original Message----- From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Best, David Sent: Tuesday, December 04, 2007 3:21 PM To: oracle-l@xxxxxxxxxxxxx Subject: How accurate is ADDM? Hey all, How accurate is ADDM? Today we had a severe performance issue with one our of databases. The ADDM recommendation for that period put the most weight on the following: ---[ADDM Snip]--- Action Investigate the possibility of improving the performance of I/O to the online redo log files. Rationale The average size of writes to the online redo log files was 932 K and the average time per write was 3020 milliseconds. ---[ADDM Snip]--- The problem is, our storage team doesn't see the same level of degredation. Any suggestions? tnx -- //www.freelists.org/webpage/oracle-l -- //www.freelists.org/webpage/oracle-l