With Cisco UCS, once the framework is set up, it is pretty easy to manage and expand. It's setting it up that is bad. Of course, the initial setup is all command line on their switches, so it's more about the switch interface than the actual servers. After that, all the NICs are virtual NICs, and don't have a proper physical interface. The PCIe connection hooks into a virtual management interface for the NICs, which have to be then defined in the management software to actually give the server a network connection. Then managing the server blades and the virtual NICs requires a special software, written entirely in Java of all the hellish things, to get a console. It all requires a master to set up. Once set up, sure, it works pretty well, but getting it there is hellish.
With Dell's remote management, iDRAC, it uses a web html interface to do most things. The remote console is in Java, but it is a stand alone app. Plus their NICs are direct, physical, real NICs that don't require any special interface.
Dude, setting up UCS from scratch is not that hard. I've setup numerous ones in our Prod colos. Yes, it is more work to setup initially but once you do it is actually easier to provision and deploy new blades than a typical Dell rack server (yes, we have those too).
*Disclaimer, I work for a company now owned by Dell*