>New info:
>
>According to the execution plan it is performing a clustered index scan on the DB tables and returns ALL of them. This is my bottleneck apparently. Should I use a join hint or some other mechanism?

Are you running SP and batch against the same data?