> We are mandated to recover, within 48 hours of declaration of disaster, our
> SAP database environment to within 48 hours of the failure. We are also
> committed to recovering two exchange servers and a fax server at this same
> level. I move 12 DLT tapes off-site daily to cover this. As a result, I
> *should* be able to recover to within 24 hours of time of failure -- unless
> the system dies on a Sunday. (The Saturday and Sunday tapes actually go
> off-site on Monday. . .)

Sending tapes offsite daily will not guarantee the ability to recover to
within 24 hours of the time of failure. Some of the critical backups at my
site start at 23:30. Our courier service usually removes the latest offsite
tapes at noon. If we had a disaster just before noon, we would lose the backup
started the previous evening, and have to settle for recovering using the
backup started the evening before that. We would end up recovering to 36.5
hours before the disaster. Other sites may well manage a shorter interval
between running backups and getting offsite tapes out of the data center, but
as long as the interval is non-zero, the worst-case interval from last usable
backup to time of failure will be greater than the cycle time for offsite tape
movement.