In the FWSM (or PIX or ASA, I guess) the xlate table timeout only kicks in after all connections have ceased or have timed out. With a dynamic NAT pool with a large number of short-ish term systems active this suggests that most of the addresses in the pool will spend most of their time waiting for a timeout (and forcing further users in an overflow PAT pool if one is configured), or will an address be reused if there are no connections currently on that translation? If the former is the case are there any guidelines or BCP on whether and how to reduce the xlate timeout?
Thanks,
Sam