I'm wondering about traffic shaping best practices. In our Corp Office Network, we have traffic shaping rules on our Main Corporate SSID and at the MX.
We have a use-case where an RDP session is failing when using Corporate SSID (with traffic shaping rules), but works properly on the Guest network (without traffic shaping rules). So my question is - do we need traffic shaping rules on both the wireless SSID AND at the MX? Is that redundant? I'm thinking the doubled rule-set is causing my RDP to timeout before it makes a successfull connection.
Generally speaking, it is always best to have traffic shaping rules, to ensure the important traffic to you is shaped or the bandwidth you need is available.
Traffic shaping works by device type, so let's suppose you applied Traffic Shaping to prioritise audio and video at the MR level, then only the MR will prioritise that traffic, once it makes it onto the next hop such as an MS or MX it will be treated as best effort, just like every other traffic type, but if you apply Traffic Shaping in the MS or MX now that traffic will also get treated with priority over everything else.
Keep in mind that once traffic leaves out the WAN, it's always treated as best effort unless you have some QoS service agreement with your ISP.
Also very important for MR: The layer 2 marking of the traffic headed downstream to your client needs a tag to put it in the correct WMM queue.
Wireless has four queues: Background, Best effort, Video, Voice with each their own max slots times for waiting before tranmission as well as transmit opportunity.
If you are VLAN tagging traffic on the SSID you can literally set the PCP value (6 for voice, 5 for video, 1 or 2 for unwanted traffic). However if it is untagged I'm not sure the AP will tag it correctly, I would have to lab this up and capture the wireless traffic.