Asterisk RTP Questions

Home » Asterisk Users » Asterisk RTP Questions
Asterisk Users 2 Comments

Hi All,

I have some questions regarding RTP and Asterisk;

I am trialling a new SIP upstream provider. We connect to them over the Internet at present which I know is not ideal, but we are just testing at present. During the trials we have had an issue where we have had one way audio between us and the provider after the call was successfully set up and bidirectional audio has been already flowing
(so at some point during an existing call, two way audio has dropped to one way audio).

I am running a constant PCAP which I sent back to the provider. They have said that the latency has increased or fluctuated to the extent that RTP as stopped sending audio in one direction (because of our test peering over the Internet). Weather this is true or not is a separate issue, what I want to know is;

What is the maximum delay RTP will tolerate one way (Does Asterisk have a limit too)?

Can this be tuned (increased or decreased) within Asterisk (I’m thinking of DSL customers where we may have this issue between our PBXs and the customer)?

How can I monitor for such an effect?

Does anyone else have any / or had any issue like this?

Kind regards, James.

2 thoughts on - Asterisk RTP Questions

  • There isnt one really. There is a rtptimeout setting but that is designed to hang up a call if no rtp has been received for X seconds. Its going to normally be something long like 30 seconds as is designed to end a call if the sip endpoint you were talking to dies.

    What direction does the audio stop?
    have you looked at the SIP traces to see if there are and reinvites at the same time?
    You havent said if your server is directly connected to the internet with its own public IP address or whether it goes via NAT.

  • Heh, should have guessed it would be you that replied Gareth 😉

    Sorry yes, this box is on public IP with no NAT as is the upstream providers box (or so they say).

    So we have had audio cease outbound towards the provider. We have a couple of volunteer customers who are being routed via this new test upstream. It’s very difficult (basically impossible!) to replicate the failure it’s so infrequent. Looking at PCAPs between us and the upstream we stop sending them audio for example and then a little while later the call drops. Without PCAPs between us and the customer at the same time I can’t say why we stopped sending audio (where we receiving any from the customer, did their connection drop for example).

    We have also had the reverse where we stop receiving audio then a short period later, SIP BYE from us to them!

    I have read up on rtpkeepalive and rtptimeout. I will put this to one side for now until we have a direct connect to the new test provider there are to many variables in the equation.

    Thanks for your input though Gareth!.

    Kind regards, James.