Refer to the exhibit. When data center 2 has a spanning-tree convergence event, data center 1 also sees high CPU utilization on its switches. Which mechanism can reduce high CPU issue in this…

BPDU guard shuts down a port upon receiving a BPDU, which would disable the inter-data-center link rather than simply filtering STP traffic to reduce CPU utilization.

Routers B and C are Layer 3 devices that do not participate in or forward STP BPDUs, so configuring BPDU filter on them has no effect on STP propagation between data centers.

BPDU filter configured on switches A and D at the inter-data-center boundary suppresses both sending and receiving of BPDUs on those interfaces, effectively creating an STP boundary between the two data centers. This prevents STP topology change notifications and related processing caused by DC2 convergence events from ever reaching DC1 switches, eliminating the high CPU spikes.

BPDU guard on Layer 3 routers would cause port shutdown upon receiving a BPDU, but routers do not run STP, making this configuration both inapplicable and ineffective for the described CPU issue.