I am looking to bridge ethernet traffic across an 802.11n link as efficiently as possible.
I am using a pair of 433ah each with a r52hn radio card.

So far I have tested setting up one side as ap bridge and other side as station wds and including them in bridging interface along with ether1 and ether2.

I have tested UDP across just the wireless link by running bandwidth test from 433AH to 433AH and achieved 165Mbps.

However when I did two UDP test in parallel through the bridge interface my aggregate throughput varied a lot and so did the CCQ. At times it would be as much as 120Mbps for a couple seconds and then much lower again.

I am currently Not using nv2. I am using both chains and using extension channel. using wpa2-psk. radios set to dynamic-wds with my bridging interface as the default bridge. single ip address assigned to my bridge interface per 433ah.

I believe all other wireless settings remain at their default values.

I do not understand why forwarding traffic from the ethernet port, over the air, and finally out the other ethernet port is causing such a performance degradation. Perhaps someone can shed some light?

What is a better architecture and configuration that will better utilize wireless bandwidth when forwarding data across the ethernet ports?

What is the expectation of plain old 802.11n in a bridging interface when used in just a PTP link?

I did try ethernet bonding and my performance was even worse. I'd love to be able to do > 100Mbps with the 433ah + r52hn in a single data stream. If anyone knows a way that'd be great, being able to do parallel data streams aggregating over 100Mbps is really good to if it turns out that the single stream idea just won't work with 433ah.

I've included a basic network diagram of what I'm trying to achieve. Thanks in advance

if you use regular 802.11 protocol then avoid using the wds, use VPLS bridging. Or if you want to use WDs try nstreme or anther option is to switch to Nv2.
About encryption - try to use only AES instead of TKIP.