Vendor has improved unified communication link, there will have one FCOE uplink to Fabric(A and B, for redundancy) to each CISCO fabric interconnect switch. Both A and B Fab can be connected in order to maintain failover. Number cables dramatically reduced. Hence, power consumption less.
In order to send out traffic to uplink, Nexus 5k switch can be used and this supports FCoE and other features too. In DC, generally we prefer Nexus 5k. All zoning, vlan and trunking are done in this switch. There is also redundancy here ie NX A and NX B. Your storage systems will be connected to this NexUS switch. Even backup will be connected to it. In order to send-out traffic from NEXUS, there are other switches. Cat 6K can be used. NexUS 7k can be also used. I heard somewhere MDS switch is also used. This is basically core switch which interfaces to ISP routers.
In order to manage UCS, there is UCS manager which is accessible through GUI. One can manage everything through backend. It is required to login into primay fabric interconnect switch and execute commands. From installation to patching, upgrading(FW and software) can be done. UCSM has nice features.
If vmware ESX is setup on blades and if they form ESX cluster with FT enabled, then your IT environment will 100% up.