While geographic redundancy can obviously be a huge benefit for disaster recovery, it is far less obvious what benefit is feasible and likely for more typical non-catastrophic hardware, software, and human failures.
Georedundancy and Service Availability
provides both a theoretical and practical treatment of the feasible and likely benefits of geographic redundancy for both service availability and service reliability. The text provides network/system planners, IS/IT operations folks, system architects, system engineers, developers, testers, and other industry practitioners with a general discussion about the capital expense/operating expense tradeoff that frames system redundancy and georedundancy.