We are in the process of building a disaster recovery center several miles away from our main campus, which has been made accessible via a high bandwidth Ethernet connection over radio. Rather than creating a new subnet for this location, it has just become an extension of the main campus. We decided to try this for redundancy purposes.
The disaster center has a bundle of T1s of a lesser capacity than that of the main campus. Of course, when the radio fails there is only one way to access the disaster center. Obviously, because it is the exact same subnet, the main campus wins out the best path to the subnet in the OSPF area. This naturally creates a situation where users on campus can't make it out of the local subnet destined for addresses at the disaster center, and users external to the campus subnet don't have the proper route for those same addresses.
I just have a few questions about this situation. Is this even the way we should be doing it? Are there any other design recommendations for purposes of the network infrastructure that we have available to us? Is there a solution to the obvious routing problem we have here?
Thanks for any input,