Your enterprise customer has asked where they should deploy flow monitoring in their network to monitor traffic between branch offices. What is your design recommendation?

Monitoring only at the network edge captures traffic at one boundary point and misses the full picture of flows between all other branch offices.

In a hub-and-spoke design, remote branches do not route traffic directly to each other - all flows must pass through the central hub site, so placing NetFlow or IPFIX export at the central site guarantees full visibility into all inter-branch traffic without requiring sensors at every individual remote location.

A separate core layer may not exist in a hub-and-spoke WAN design, and inter-branch traffic does not necessarily traverse a distinct core segment.

The data center carries server-to-client traffic but not necessarily peer branch-to-branch user flows, so monitoring there would miss inter-branch traffic.