but I do see a problem with what you wish to do and how you've implemented the network infrastructure to do what you wish to do. Therefore Exchange is impacted [as well as other network resources].
As I said, if you add an exchange server into a network with another exchange server, the old guy will see the new guy [be aware of it]. But that does not mean the the new guy is automatically configured to be a fail over.
Even more importantly, you do not mention if the second DC which is home to the new guy Exchange is itself a GC server and a DNS server. As a minimum, if the first DC holds all roles and there is no other DC which has roles assigned to it including a backup DNS server, if the old guy goes down, your whole network becomes inaccessible [no DNS in which clients can find resources, more importantly, no DNS means no AD functionality. further no GC role in which user can authenticate, so they would have to use cached credentials and really only log on to the local machine.
this is an article on Exchange 2000 dealing with high availability. In this article, you'll find some of the more common methods for scaling Exchange and maintaining a high availability mail service. How you want your network will determine which types of methods used for high availability. Note: the most commom is to cluster Exchange servers.
http://www.microsoft.com/technet/prodtechnol/sql/2000/deploy/harag05.mspx#E5HAC