I'm sure some folks will disagree with me here but I think there is a difference between a business continuity disaster plan and an identity management disaster plan. Most companies, I hope, have a business continuity disaster plan which basically documents what needs to get done if a server room is flooded, the mainframe catches on fire or the really bad scenarios of buildings or locales being "destroyed".
However, how do you handle the situation of your primary authentication directory "blowing up"? New hardware and restore the latest backup? What about the new identities that were created between the last backup and the incident? Just lost? Sorry, you can't pay your credit card bill?
I know of a bank which moves millions of dollars a day based on a Linux system that authenticates to Active Directory. What happens if AD is not available for 5 minutes? What then?
Let's even go further down the food chain. Your AD (or name your favorite directory) administrator decides to leave on a Friday and deletes or changes a bunch of information. What do you do? Do you restore from the last backup? What about all the changes? What if the admin accidentally deletes an OU? What about any changes that occured between the last backup and the incident?
I would posit that you need near continuous backup for your identity and authentication repositories. If you aren't there yet you really need to be thinking about it...
identity management, compliance