I have come across of many forums & mailing list where people asking for the optimal value (size) for online redo logs files.
I would say, prior to 10gR2, you need to estimate its value upon system redo generation and other database setting factors.
Although, log file size is not the candidate to measure or improve LGWR performance, but, it (log file size) can affect to DBWR and ARCH bakground process performance. Therefore, log file size should be large enough in order to avoid the frequent checkpointing in the database.
Prior to Oracle10R2, there was no direct hint by oracle that you need to set this value as an optimal,or at least start with this value.
In Oracle10gR2, the optimal value for log file can be obtained from v$instance_recovery dyanmic view.
select optimal_logfile_size from v$instance_recovery;
453 (size measured by MB)
Note : The above value is considered optimial based on the current setting of
Oracle documentation suggest the following:A rough guide is to switch logs at most once every twenty minutes.Size your online redo log files according to the amount of redo your system generates.