A geographically dispersed cluster is no different from a standard server cluster, except that one or more nodes of the server cluster reside in another location. This type of cluster is achieved through specialized hardware, software, or a combination of both. There are still distance restrictions due to how long, say, a dark fibre network can be, but a geographically dispersed cluster removes the data center as a single point of failure. (See Chapter 4, Disk Techniques for High Availability, for more information about a dark fibre network.) One barrier to adoption is the cost of implementation, because the network needs great bandwidth across a long distance, as well as a duplicate set of hardware, including the disk subsystem, in both locations. Technically, the secondary location does not have to completely mirror the primary location (that is, it could have fewer nodes and disks as long as it has enough processing power to handle the highest priority applications if something occurs on the primary), but the environment should mirror the production environment as closely as possible.
All geographically dispersed server cluster configurations must appear as a complete hardware solution under the specific geographic cluster list on the Microsoft HCL for the operating system choice you are making. The configuration is a viewed as complete solution, not individual parts that can be put together like a jigsaw puzzle. Microsoft supports only a completely HCL-compliant cluster solution. Please consult http://www.microsoft.com/hcl/ for more information.