I have noticed the since a few years that TX and RX statistics in bridged interfaces (eg: a WiFi interface attached to the LAN) are inverted, I have searched for an answer for why this happens but I have not found any canonical source.
Is this an OpenWRT practice or a Linux one? Or is it a standard across all OSes?
Does anybody know why or has a link where this is explained?
The main issue for me is that traffic monitoring charts of OpenWISP Monitoring will display the upload/download in inverted fashion for this kind of interfaces that are member of a bridge, which confuses some users, so I would like to understand if I should invert it on the application level and if so when is right to do so, but before doing so I would like to clear these doubts which are puzzling me, there must be a reason for this practice but I can't find much info on it, maybe I'm using the wrong search terms..
It shouldn't be too confusing, the traffic displayed is not [actually] inverted.
"Inverted" compared to what?
the WLAN
the WAN
It would be inverted to the comparison LAN in those cases.
Without getting too technical, imagine you're standing on the opposite side of a door..or you stand at the middle of a [physical] bridge observing traffic in one direction, then stand on the opposite side or turn 180 degrees, respectively. You see the same packages (packets) going in and out; but you see them as a mirror/reverse of your previous opposite orientation.
Do you know an URL where I can find more information about this standard practice?
Initially it is a bit confusing, not only to me but pretty much with all the people I interacted with, but after reading why it happens and explaining it to others, then it is not confusing anymore and becomes straightforward, I totally get the analogy. The same happens also in the RADIUS accounting protocol (input/output octets are inverted).
However, when presenting information to users, which may not be technical people, most network management software inverts these values again behind the scenes to present it to users in a way that is familiar to them.
For example, if they want to know how much download and upload traffic is consumed by users, either via the RADIUS accounting protocol (when available) or from the interface statistics, these applications take the necessary steps to make sure that values shown in "Download" and "Upload" reflect what the average user expects: Download and Upload from the end user point of view, not from the point of view of the bridge or the NAS. That's the point I'm trying to explain.
I just want to make sure to invert the right values and I'm looking from some technical source which I can use as a reference. If anyone knows, that would be great in clarifying this doubt.
