No Voice Channels between 2nd Netlink Switch and Remote IP Phones

[ErnieE]

Hi,

Here is our client's setup:

Two offices in different cities with two SV8100s netlinked thru a VPN Tunnel. A number of remote offices that are connecting using IP phones(All DT700). Remotes are configured for NAT.

Problem: The office with the secondary switch can call a remote office but there will be no audio on the call. The same holds true for a remote office calling an ext on the secondary switch. Calls to and from the primary switch extensions work fine calling remote and secondary switch extensions.

They have a large number of offices so VPN tunnels between all the offices would be hard to manage.
we have done Wire shark traces and have determined that the traffic from the remote phone to the second switch is being directed to the Primary Switch WAN. The traffic from the Secondary switch is being directed at the WAN of Remote Phone. so voice channels are traveling in a triangle.

We can not setup a router port forward that works to get the traffic from the Primary Switch that is meant for the secondary switch to the secondary switch.

Has anyone else been successful at setting up this kind of configuration. Or any other suggestions.

 

[belevedere]

Typically a no voice issue with netlink is related to the PAD card. Either a port is not opened up or there is an discrepancy with the IP addresses related to the PAD card.

 

[gmannatl]

Where did you do the packet capture?

If you are in the secondary site and you go to the internet are you going through the tunnel and out of the Primary's internet connection? For example if you have a Network Security Gateway that needs to be used at the Primary site? Go to whatismyip to confirm.

I would be interested to know how you tested to confirm that traffic coming from the remote is hitting the primary site? Signaling traffic WOULD go the the primary site, so just seeing some traffic coming to the primary does not mean it is passing the VoIP there. What kind of traffic was it Sip or RTP?

My guess would be (assuming your internet from Sec is not going out Primary) open all of the RTP ports in the secondary sites firewall to the PAD in the secondary site. It could be you are seeing the SIP signaling at the primary, but the RTP can't get to the second sites PAD.

 

[OzzieGeorge]

Try checking if all RTP ports are permitted inbound on the router!

 

[ErnieE]

To All:

I have all of the prescribed ports for the DSP resources on the secondary switch forwarded on the Secondary router. (The port range for the 2 DSP on the secondary are in the 20000 range, the DSPs on the primary are in the 10000 range).

I placed a wireshark between the phone and lan and at the secondary switch between the switch and it's LAN port.
There is an RTP stream coming from the Secondary Switch(thru it's router WAN Ip) to the phone. I can decode the stream and it has audio but it is being rejected by the phone (I assume) because it is not from the address that it expects it from. When I look at the SIP call initiation sequence, I see the extension on the 2nd switch that I want to talk to being marked as <EXT>@<PRIMARYSWITCHWANIP> and the RTP outbound packet's destination IP is the Primary Switch WAN IP.

No traffic arrives at the Secondary switch because the Primary Switch router is not forwarding to the Secondary switch(as it is not designed to support a port forward out of the Router's LAN.)

So I am getting a discarded stream from the Secondary switch inbound to the phone and the outbound from the phone is going to the Primary Switch WAN IP. All DSP ports are for the 20000 range just as the secondary switch is configured for and that is what the SIP session is indicating it should go to.

I have read that this configuration works for some but not for others. NEC support indicated that sometimes it requires VPNs, sometimes it does not.

We are in the process of trying to get the customer working by configuring VPNs to both switch routers from all remote locations but we are hoping to solve this for future clients.

I am hoping to find out that there is somewhere to tell the Primary switch that all traffic for secondary switch should use the secondary's WAN IP but I am not finding that.

Sorry if this is long but I have been dealing with this for a bit and figure you guys can only help if you have an accurate picture of what is happening.

I do appreciate all of your time and thought on this.

 

[OzzieGeorge]

Ok

So I'll keep asking stupid questions in the hope of helping! Could it be anything to do with the fact that NEC SIP uses Port 5080 instead of 5060?

I have to admit that I am an Older Voice tech so the networking side is somewhat new to me! Ask me a telephony question and it's a different matter!

 

[ErnieE]

Yes the SIP registration port is set to 5080 on these switches. The default is 5060 but we have found that on COX cable networks Cox uses the 5060 port for their own equipment so they can block those ports.

The switching aspects of these calls work. Each phone sees the correct information about who they are connecting to and the calls initiates correctly. There is no voice.

Thanks again

 

[OzzieGeorge]

Back to basics, have you tried factory defaulting the phone then putting in static settings?

 

[OzzieGeorge]

Another question is did you change any of the default IP addresses on the system as here in OZ NEC are insistant that you don't because if you do, unpredictable things can happen!

 

[smoom]

Attach a copy of your wire shark from each side so I can look at it and give you an idea of what is happening.