Why would it be any different than setting it up on a regular rackserver environment. The only thing which might be a problem is when the clusters use multicasting between each other. There are some topic here which describe some problems with multicast without an IGMP querier in that particular vlan.
The multicast issue is really only applicable to NLB, not a MCSC.
Multicast functionality has been discontinued in Windows Server 2008 failover clustering, and cluster communications now use User Datagram Protocol (UDP) unicast. (Port 3343 is still the common port that is used by Microsoft failover clusters.) There are now new messaging processes that are internal to the Cluster service itself. When nodes transmit and receive heartbeats to confirm that each node is still available, the nodes use Transmission Control Protocol (TCP) rather than the less reliable UDP.
In my experience classic shared storage failover clusters have been just like rackmount servers. I setup a separate vNIC and VLAN for the heartbeat cluster private network, but that's about all I did that was UCS specific.
There is some addtional info in this thread: https://supportforums.cisco.com/thread/2013429