Hi,
I have a setup where a BR2 is connected to a Rajant WiFi mesh node on its WAN interface. Of course, the bandwidth of the mesh network is variable, depending on the circumstances.
This way, it is challenging to configure QoS properly. One way is to use the QoS possibilities of the Rajant, which require ToS/DSCP or VLAN-based QoS. The architecture of the various LANs currently doesn’t allow this, this would be a major change internally.
Are there other ways?
Problem is I have to set the upload/download bandwidth of the WAN interfaces to a static value, e.g. 500 Mbps. A high value is necessary for fast data transfer when the signal is very good. However, a high value has the effect that applications configured with a high QoS priority actually don’t get that priority. When the signal gets worse and the available bandwidth on the mesh network drops to e.g. 2 Mbps, the BR2 still thinks there is still more than enough bandwidth available in the 500 Mbps pipe where the packets get dropped by the Rajant instead of the BR2. A dynamic setting for this interface bandwidth setting would be better.
Another way would be like how it’s implemented in Fortinet, where you can give priority to certain applications once the latency on a link reaches a certain value, where non-priority applications get dropped once this cut-off latency value is reached. Somewhat like we do with the cut-off latency setting in Speedfusion tunnels, but then for applications instead of a whole tunnel.
Are there other ways, or would this be a valuable feature request?