Usage Monitor Gone Mad
Is there a closed ticket on your account saying your usage has been reset yet? I suspect that as today is the first day of your new billing period your usage may not start to appear until that happens.
I believe the Junipers that handle the Centrals are the main place where the traffic management is applied (as the Centrals are the main choke point on the network) although I would assume that all of the routers on the network are QoS aware and will schedule traffic based on the Diffserv bit set in the TCP packet in times of congestion.
The Ellacoyas sit "the other side" of the Junipers from the customer, and are responsible for inspecting a packet (or stream) and setting the Diffserv bit accordingly.
This results in (and this is an educated guess) any outgoing packets being tagged by the Ellacoyas and then transmitted across the PN network to the peering point, whereby other QoS aware routers *might* take some notice of the Diffserv bit. Inbound (return) traffic wouldn't be tagged until it got back to the Ellacoyas, at which point the packet is handed off to the Junipers to make the decision on whether to transmit down the central, or to drop it based upon traffic loading.
The PN Network has plenty of capacity to handle the traffic from all the centrals - the tagging really is used to make the decision whether to send the packet down the central or not.
This would lead me to the logical assumption that the traffic is metered at the Junipers. However, I actually know that's not correct
The metering information is actually provided by BT every (15?) minutes. This information is recorded (I would assume) by the BT ATM network, and is transmitted to PN frequently.
The VMBU usage is in two parts - the actual metered amount of information is provided by BT. The traffic-type breakdown is provided from the Ellacoyas/Junipers. The figures will often not match down to the byte, but are very very close in practice.
B.
The Ellacoyas sit "the other side" of the Junipers from the customer, and are responsible for inspecting a packet (or stream) and setting the Diffserv bit accordingly.
This results in (and this is an educated guess) any outgoing packets being tagged by the Ellacoyas and then transmitted across the PN network to the peering point, whereby other QoS aware routers *might* take some notice of the Diffserv bit. Inbound (return) traffic wouldn't be tagged until it got back to the Ellacoyas, at which point the packet is handed off to the Junipers to make the decision on whether to transmit down the central, or to drop it based upon traffic loading.
The PN Network has plenty of capacity to handle the traffic from all the centrals - the tagging really is used to make the decision whether to send the packet down the central or not.
This would lead me to the logical assumption that the traffic is metered at the Junipers. However, I actually know that's not correct
The metering information is actually provided by BT every (15?) minutes. This information is recorded (I would assume) by the BT ATM network, and is transmitted to PN frequently.
The VMBU usage is in two parts - the actual metered amount of information is provided by BT. The traffic-type breakdown is provided from the Ellacoyas/Junipers. The figures will often not match down to the byte, but are very very close in practice.
B.
