We have three T-1s that form a single Multilink; that one multilink services five tunnels. Every tunnel and the Multilink is set to BW 4.632Mbps. Both NetQos and Concord eHealth show a single tunnel using twice the BW than the Multilink. We do have QoS set on the Multilink, and each tunnel is encapsulated via GRE. I've got to explain to the customer how a tunnel can consume more BW than the Multilink...any ideas?