The scaled single-site Edge Server topology enables you to quickly add new servers to the edge network as your environment grows and more users begin to consume services. This topology also provides for increased availability. If, for example, a Web Conferencing Server fails, other Web Conferencing Servers can take over the workload.

In this topology, hardware load balancers are required on the external and the internal sides of the edge network. Hardware load balancers are required because the potential for latency over software solutions is lower. Software-based load balancers are not supported. For more information, see the Microsoft Knowledge Base article titled "The Windows Server 2003-Based Network Load Balancing Service Is Not Supported by Communications Server 2007 in Production or Lab Deployments," found at http://support.microsoft.com/kb/939800.

It is important to remember that whether the load balancer is serving one server or ten, the servers are all referenced by the virtual IP (VIP) of the load balancer. In the case of the DNS, the only IP address that needs to be registered is the VIP of the load balancer. In fact, it is better not to register the servers separately. Also, when you configure certificates in a scaled single-site Edge Server topology, the subject name should be the DNS name of the VIP address of the load balancer, not the host name of any server in the array.

Figure 7-2 shows four consolidated Edge Servers. Each of the servers provides the Access Edge role, the Web Conference Edge role, and the A/V Edge role.

Figure 7-2. Scaled single-site Edge Server topology

Hardware load balancers provide a distinct advantage in scaling your edge architecture—you can add servers to provide for greater availability and greater load handling. In other topologies, the only way to scale your consolidated Edge Servers is to scale up, meaning a server that has more processing power and memory. The scaled single-site topology also enables you to scale out. Instead of using fewer, more powerful servers, you can use more but less powerful servers that are typically more cost effective. The other benefit, as mentioned earlier, is the ability to realize availability as an enhancement. In the topologies that we have discussed to this point, if an Access Edge Server fails, your enterprise will not be able to communicate until it is recovered. With the scaled single-site Edge Server topology, you have more than one consolidated Edge Server in place so that your enterprise will continue to operate if one of those servers fails. Being down during recovery is no longer a production threat because the other servers will continue to provide service.