- Gold, 750 points or more
We are currently working on a design and Dual Supervisor's have been specified in the Core & Distribution switches (Sup720's). In previous designs we have NOT used dual Supervisors in the Core or Distribution layers due to the instability that happens during a supervisor switchover - i.e. links stay up but routing protocol's fail and adjacencies are restarted resulting in a period of network instability. Without dual supervisors the whole node is taken out of service and traffic is re-routed around the failure, this results in very predictable behaviour.
Cisco now has NSF capability in the Sup720. Obviously the other switches/routers need NSF awareness in the routing protocol to take advantage of this. Does anyone have any real world experience of NSF? I have found the documentation on CCO and understand how it is supposed to work but I haven't found any white papers or case studies so have nothing to compare our test network with.