Function
Used by the Derby system
at runtime in determining when to attempt to escalate locking for at least one
of the tables involved in a transaction from row-level locking to table-level
locking.
A large number of row locks use a lot of resources. If nearly all the rows
are locked, it might be worth the slight decrease in concurrency to lock the
entire table to avoid the large number of row locks.
For more information, see "Locking and performance" in
Tuning Derby.
It is useful to increase this value for large systems (such as
enterprise-level servers, where there is more than 64 MB of memory), and to
decrease it for very small systems (such as palmtops).