If I remember correctly SQL Anywhere provides a method to calibrate the real local IO speed, so that the optimizer is not using anymore the 40x speed assumption for sequential reads compared to random reads of the database files.
Does anyone know more details?
e.g. is the calibration time dependent on the database size?
I guess you are refering to the ALTER DATABASE CALIBRATE statements:
No, I don't know more details...
answered 13 Dec '10, 16:12