Robot/garbled voices, Need help understanding the call/voice flow

[rejackson]

Runing CM3.1.5 on 8700 media servers at a central location with g650 media gateways there. G700 media gateways at remote locations in an IP connect configuration to the media servers. The G700s have their own local PRIs. 2420 digital phones on MM717 boards in the G700s.

User at a remote site makes a call outbound on their local PRI. They get random poor quality. Nothing shows on the PRI. Telco tested and found nothing. Checked firmware in remote g700 and boards and the g650 and boards.

So I have a digital phone going out a PRI on the same media gateway but the media servers are across the network. When the call connects what resources are being used? I cannot figure out if the voice path is crossing the WAN network. I do nnot see anything about hair pinning except for IP phones.

Another site reports the same issue but also has the problem internally between local extensions.

Thanks

Richard

 

[DAVIDPAYNE]

Are your remote sites setup as there own Network Regions?
What codec set have you setup?
What QoS settings? And where are they set? (Network, Avaya, etc.)
What type of WAN do you have? (MPLS, P2P, etc.)

The more information you can give us, the better we can help you.

 

[rejackson]

Yes separate network regions

G711 inside each office and to a centrally located voicemail system across the WAN. Except for the voicemail, G729 over the WAN.

They were both MPLS sites when the problem started but have since moved to P2P

The network regions are marking QOS at DSCP 46. When they were on MPLS that was being mapped to IP Precedence 5 at the router so it would map directly to the MPLS settings. Now on the P2p the 46 is used end to end.

I am running IP SLA in the Cisco routers and the P2P circuits are very fast and clean.

Dallas_WAN#sh ip sla mon stat
Round trip time (RTT) Index 13
Latest RTT: 7 ms
Latest operation start time: 08:52:43.714 CDT Mon Jul 23 2012
Latest operation return code: OK
RTT Values
Number Of RTT: 30
RTT Min/Avg/Max: 7/7/8 ms
Jitter time milliseconds
Number of SD Jitter Samples: 29
Number of DS Jitter Samples: 29
Source to Destination Jitter Min/Avg/Max: 0/1/1 ms
Destination to Source Jitter Min/Avg/Max: 0/1/1 ms
Packet Loss Values
Loss Source to Destination: 0 Loss Destination to Source: 0
Out Of Sequence: 0 Tail Drop: 0 Packet Late Arrival: 0

Round trip time (RTT) Index 36
Latest RTT: 39 ms
Latest operation start time: 08:52:30.794 CDT Mon Jul 23 2012
Latest operation return code: OK
RTT Values
Number Of RTT: 30
RTT Min/Avg/Max: 39/39/40 ms
Jitter time milliseconds
Number of SD Jitter Samples: 29
Number of DS Jitter Samples: 29
Source to Destination Jitter Min/Avg/Max: 0/1/1 ms
Destination to Source Jitter Min/Avg/Max: 0/1/1 ms
Packet Loss Values
Loss Source to Destination: 0 Loss Destination to Source: 0
Out Of Sequence: 0 Tail Drop: 0 Packet Late Arrival: 0

My real question is whether the WAN is even being used once these calls are set up, and if so what resources (medpros?) are being used at the central location. If the medpros are being used is there a way to identify which one?

1) digital 2420 phone out PRI on the same remote G700.

2) digital 2420 phone to another digital 2420 on the same G700 stack.

 

[rejackson]

I found this on the status station. It appears to show that the call path is all local. It uses the voip engine in the g700 where the station is connected and the voip engine in the g700 where the PRI is terminated to link the exttension to the PRI. This is a 3 g700 stack with the extension on MG2 and the PRI on MG3.

SRC PORT TO DEST PORT TALKPATH
src port: 004V208
004V208:TX:ctxID:163 Extension
004V015:RX:ctxID:163 TX:10.2.56.16:2410/g711u/20ms MVoIP2
039V015:RX:10.2.56.18:2060/g711u/20ms TX:ctxID:68 MVoIP3
039V422:RX:ctxID:68 PRI channel 22

So are there any commands for diagnosing problems on the stack links? I had heard, and it is obvious here, that they use IP not TDM.