MTR Traffic always using the MS relay?

New Contributor

So after installing a couple of Poly GC8 MTR nodes for testing we immediately get the audio and video quality is terrible and worse than the Polycom Realconnect units they replaced.  After some test we did wire shark traffic capture which revealed the MTR unit are not establishing local peer to peer flows. All the traffic that should of been going between two sites about 30kms apart was been sent for a full round trip to the MS relays (100ms away, they are on a mine site near nothing).

 

I did the same testing on our MTR node and found it does the same thing.  

However, when any PC with the Teams Desktop client call each other, the traffic flows stay local and direct (again confirmed with wireshark).  

The MTRs connected in the local switched network there is no firewall to limiting connectivity so I am currently at a loss to understand why the MTR client acts differently to the Teams Desktop client.

 

Initially we wondered if it was because the MTR was not part of the AD domain, but adding it made no difference.  

All policies for Team desktop and the MTR are the same….  

Anyone know any reason to why this would be happening?

1 Reply

So update on this I have done further reading and it appears that Microsoft documents "gloss over" the fact that meetings ALWAYS use take place in the "cloud". The Teams traffic flow document only mentions 1 to 1 calls not meetings.

So for us these MTRs are worse than our existing Polycom VCs since they don't rely on "Cloud" to do P2P between our conference rooms.  

On the upside I am only the person find/fixing these issues and not the person who started this fiasco of a project promising Teams will fix all VC problems!