Sometimes, users may notice that the Bits Per Second (BPS) values in the show interfaces extensive
output and the show interfaces queue
output are different. This can cause confusion when troubleshooting interface bandwidth or Class of Service (CoS) related issues.
This article provides an explanation for the BPS values to be different in these command outputs, and includes the formulae to use to calculate different traffic rates.
When checking BPS values on egress interfaces by using the show interfaces extensive
and show interfaces queue
commands, significant differences in the values are observed in the two outputs:
user@host> show interfaces xe-5/0/1 extensive
Physical interface: xe-5/0/1, Enabled, Physical link is Up
.....
Traffic statistics:
Input bytes : 1143907959514 5800027488 bps
Output bytes : 1143916433329 5800027720 bps <<< 5.8Gbps
Input packets: 11987204562 12500059 pps
Output packets: 11987261126 12500059 pps
......
user@host> show interfaces queue xe-5/0/1
Physical interface: xe-5/0/1, Enabled, Physical link is Up
.....
Queue: 0, Forwarding classes: BE
Queued:
Packets : 12429718082 12499952 pps
Bytes : 1691620731707 9999961984 bps <<< 10Gbps
Transmitted:
Packets : 12429718082 12499952 pps
Bytes : 1691620731707 9999961984 bps <<< 10Gbps
....
Notice that the Packet Per Second (PPS) values are nearly the same in the above two outputs but the BPS values are very different (5.8 Gbps Vs. 10 Gbps).
This is expected behavior due to the following:
On Ethernet interfaces, different traffic rates can be calculated as below:
-
L3 rate (bps) = pps * 8 bytes * layer3 packet size
-
L2 rate (bps) = pps * 8 bytes * ethernet frame size
-
L1 rate (bps) = pps * 8 bytes * (ethernet frame size + preamble + inter-frame gap)
It is recommended that users be aware of the type of traffic rate displayed in each command and use the above formulas to convert traffic rates in different layers.