How can we block the forward of a=direction:active on SDP from A-leg to B-leg?

5. Configuration
1

When ProSBC handles NAT of A-Leg on a NAP with NAT “Remote Method for RTP”=”Parse Direction Attribute” it shouldn’t forward the a=direction:active to the B-leg. Right or are we missing something?

We have the A-leg NAP configure with NAT “Remote Method for RTP”=”Parse Direction Attribute” and B-leg NAP configure with NAT “Remote Method for RTP”=”none" and the line a=direction:active is being forwarded to B-leg. How can we remove a=direction:active from SDP on B-leg?

[Additional information]
We want to replicate or TMG behavior on ProSBC.
On TMG a=direction:active isn’t being forward from A-leg to Bleg. This is what we expect because we need to isolate our clients behavior from operators behavior.
On ProSBC a=direction:active is being forward from A-leg to Bleg. We must not forward.

2

We try on:

  • “Profiles → SDP-> Profile SDP Description” with and without adding “a=direction:both”
    and
    “Profiles → SDP->Determine if and how media encoding payload type remapping should be done.” = “Disable” and “All events”
    Combination of both configurations and none make difference…
3

Forwarding a=direction:active from A-leg to B-leg causes ring back tone problems.
What I’m I missing on configure ProSBC as a Back-to-back user agent (B2BUA)?
Is this issue related with roadmap “TB-4088: SDP Manipulation”?

4

We’ve upgrade to version 3.3.11.67 but the issue persist.

After ProSBC treat the “Parse Direction Attribute” on A-leg NAP it must not forward “a=direction:active" to B-leg. When will this issue be fixed?

Please note that this is issue doesn’t exist on TMG7800 with version 3.0.171

5

How can I download the version 3.3.12.53 to test?
With some operators this behavior causes ring problems.
Is this a misconfiguration or a bug?