Redo log size & performance

We got a performance issue with our Oracle 10g. Suddenly the performance was going down and we was getting a significant increase on wait times. After sniffing around and asking a friend we found that:

1. All our performance pick were related with time when our Oracle was writing the archive log.
2. There is a direct relationship between log_buffer, redo log file size and performance.

So, here is a tip. Your redo log file size must be multiple by 2 of your log_buffer. What I mean is that if your log_buffer is 5 MB your redo log file size can be 10 MB (2 x 5 MB), 20 MB (4 x 5 MB), 30 MB (6 x 5 MB) and so on...