Greetings, all.
In order to mitigate a single point-of-failure in the hardware architecture in blade server chassis, our customer wants to be able to utilize both the dual 10GE and the dual 1GE LAN-on-Motherboard (LOM) connections each blade has to the I/O ports of the enclosure in an active/standby configuration. The server engineering team believes this is only attainable using the VMware vNetwork Distributed Switch. The network side of the shop requires the advanced security, management and QoS feature-set provided by the Nexus 1000v distributed switch.
I was able to track down a document which outlines what we'd like to do, and it seems the 1000v is indeed capable of such an implementation:
http://www.cisco.com/en/US/prod/collateral/switches/ps9441/ps9902/white_paper_c07-607716.html#wp9000506 . The network diagram illustrates our design - 7k, 5k double-sided vPC with dual port-channels, one 2x10GE, one 2x1GE down to the 1000v.
Is there any additional documentation regarding this design? I'm looking to provide details of configuration, best practices, limitations and the like before we move forward. For example, how are the port-groups in vCenter, and subsequently the port-profiles in the 1000v coordinated?
Any help on this is greatly appreciated!
- Greg.
