I am looking at deploying an ACE to load balance some Terminal Servers and some Citrix Servers for a large serverfarm consisting of approximately 6 - 10 realservers. We are currently using DNS round-robin to load balance these services but it is not working properly because of the fact that we cannot detect realserver failure - hence the use of the ACE health monitoring.

I face the following potential problems:

1. There is a possibility that we cannot change the IP's on the existing servers that are to be loadbalanced. This imposes a problem for a routed implementation.

2. Assigning a new VIP for the serverfarm might not be possible as it requires reconfiguration of all client machines, which is a large population.

Any comments/ideas? I have thought of using a bridged mode implementation (trickier than L3 Routed implementation). Can I still achieve load balancing without using a VIP when using bridged mode?

Any thoughts on achieving this in bridged mode?

Thanks in advance for any ideas/pointers.