How to predict a part of work which will go to storage system after migration to Exadata ?

Hello all,

I'm analyzing the customer environment look for the best candidate for Exadata. I try to predict a part of work which will go to storage system after migration to Exadata (the Smart Scan share )

1st part.

Customer has some databases. I've got the AWR for each.

In AWR I see:

1st DB: 85% CPU time, db file sequential read 8%,  db file scattered read 0.43%.

2nd DB: 70% CPU time, db file sequential read 8%,  db file scattered read 0 on TOP5.

3rd DB: 80% CPU time, db file sequential read 17%, db file scattered read 4.5%

In other words, IO share is now very big, and the multiblock reads is insignificant.