A/V Edge Server with Load Balancer....still need public IP ADDRESS?

The question has been raised a few times, now that some enterprises are moving beyond the testing phase of OCS 2007 and into large scale production, about the need for a publicly routable IP address when using multiple A/V Edge servers behind a load balancer. First, let me paint you a picture of what a scaled edge topology looks like:

So, you see in this picture we can have multiple edge servers feeding a particular site. In this example we have decided to scale the A/V Edge Server on its own hardware because we anticipate using a high volume of A/V traffic with extended users and federated partners. Since A/V traffic is the most resource intensive traffic, we want to make sure we can handle the load. So, we have load balanced two A/V Edge Servers. By the way, I'm borrowing this diagram from the Edge Server Deployment document. I highly recommend you set this website (https://www.microsoft.com/downloads/browse.aspx?displaylang=en&productID=5EFC9E68-052F-4CAB-9F29-02BFA05A8F2F) in your favorites because it will link you to all the documentation for OCS.

With that said, the question is, I know that the A/V Edge Server needs a publicly routable IP address. So, where do I put it? On the server or on the load balancer? The answer is...both! The A/V Edge Server FQDN needs to resolve to the public IP address of the load balancer. But the A/V Edge servers still need a publicly routable address as well because the allocated ports of 50,000-59,999 will be on the actual A/V Edge server's IP address, not the load balancer. All the load balancer does is select which A/V Edge server to use. Once that selection is made, that external client will communicate directly with the A/V Edge Server itself.