That's true,
you should find the bottle neck and then resolve it rather than coming up with predictions.
Most of the time, the performance hit is from slow running SQL statements rather than the way
JDBC is written. It won't be useful to optimize JDBC code if a query takes 10 seconds to run on the DBMS side.
Some JDBC performance tips are: Use connection pools. Under the right situations, use batch updates, reuse compiled PreparedStatement.
[ April 06, 2007: Message edited by: Chu Tan ]