You may have just seen my other question here.
One of the other things the client failed to tell me was that they also wanted a DAG between the U.S. and a site in Europe. I had never set up a DAG, but I did it, and opened a case with Microsoft to make sure I did it right. It seems fine, and everything replicates.
The problem is that when I attempted to test it failing over, it would not. I found information here that leads me to believe they may also need a hardware network load balancer, but I do not fully understand how it works, and how it makes Exchange understand that it has failed over. Could anyone explain that to me? I have looked.
Thank you.
Your servers are probably failing over correctly, but you need a mechanism for the clients to fail over to the other CAS server.
Our solution is to use a load balancer and a custom DNS name for CAS connectivity (casarray.company.com).
You said you're working with only two servers, one per each site, with all three main Exchange roles enabled (Mailbox, Hub Transport, CAS).
You also said that in order to test failover you are "shutting the services off on the server in the US".
In this scenario, it's perfectly normal for your clients to not be able to automatically fail over to the other site. In order for Outlook clients to be redirected, they need to be able to talk to a local Client Access Server, which will then tell them "your mailbox is somewhere else, please go there". If there is not any active Exchange server in the site, they will just try to connect to the last server they contacted succesfully, and they will fail because it's not available.
In order to test if the other site is working, you can manually reconfigure Outlook to point to the remote server. But if you want automatic failover capability, you will need more than one server per site.