You have recently configured an aggregated Ethernet interface between two devices in your
network. The member links do not all reside within the same PFE complex. You must ensure that
the bandwidth and burst size defined in a policer applied to this aggregated Ethernet interface are
properly enforced when the traffic falls out of profile.
Which configuration parameter should you use?
2 Comments on “Which configuration parameter should you use?”
C is correct
Shared bandwith policers are a way of rate-limiting aggregated Ethernet interfaces (or in general, multiple interfaces on different PFEs); this is difficult to do, as you would need to syncronize policer credit counters across PFEs in real time. So the ‘shared-bandwidth-policer’ command just ‘cheats’, by splitting the bandwidth among all members of an aggregate.
For example, if you have an AE interface composed of 2 GigabitEthernet links on two FPCs and you apply to it a shared-bandwidth-policer with a bandwidth-limit of 100M, each link will be subject to policing at a rate of 50M. This works really well.. if your traffic balances well; if your traffic is not well-balanced, you might have… suboptimal results.