[Problem] COMPact 5020 intermittently fails SIPDiscount registration

[Agile]

Dear Forum,
I apologize for writing in English. Unfortunately my German writing is not what it takes! Feel free to reply in German though.

I would appreciate your advice on a problem with SipDiscount (Betamax group) in an Auerswald Compact 5020 VoIP PBX.

I’m able to call out and receive calls via my SIP number, but every ~7 or so minutes the 5020 VoIP LED turns off and the status indicator in the 5020-configuration manager reports “Error 503 Service Unavailable” or “Error 503, DNS error”. After a few minutes the error message disappears, the LED turn on again and all is OK for a while. During the error-situation, I cannot call-out, call-in is possible.

I tried various changes in the settings and searched on various fora, but have not been able to solve the problem. Admitting VoIP is a new area for me, so I assume I have made a mistake somewhere.

I’m using a Dlink DIR-855 router and obtain my internet connectivity via a cable modem (Dutch operator UPC). The modem is in bridge mode. I have tried to disable the SIP ALG in the router, but this did not help.

These are my 5020 PBX settings (leaving out less relevant details):

Hardware configuration
- 2VoIP module
- 2ISDN module
(I have just ordered a 2POTS and TSM module to complete my setup)

PBX Server Configuration
- Fixed IP address
- Gateway: 192.168.0.10 which is the router
- DNS-1 and DNS-2: copied from what the router obtains from my internet provider

VoIP provider ‘SIPDiscount’
==SIP==
Domain: sipdiscount.com
Registrar: sip1.sipdiscount.com
NAT traversal: activated with use of STUN
Interval for NAT keep-alive: 15 sec
Outbound Proxy: deactivated (e.g. not ‘automatically’, whatever that might mean)
Time lapse for the registration: 5 min
SIP-UDP port: 5067
SIP session timer: 5 min
Only en-bloc dialing: off
==RTP==
NAT traversal: activated with use of STUN
DTMF signaling: both procedures
Echo Cancellation: on
Jitter buffer: 50 ms
Codec settings Priority: 1:G.711, 2:G.726, 3:G.729, 4:G.723, 5:---
==SETTINGS==
STUN server: stun.sipdiscount.com Port 3478
Interval for STUN server query: 5 min
Support for T.38 providers: off
Sub-system operation: off
Connect audio through: off

To be complete, this is what SIPDiscount suggests on their website:
SIPDiscount settings
- SIP port: 5060
- Registrar: sip.sipdiscount.com (*)
- Proxy server: sip.sipdiscount.com (*)
- Outbound: proxy server : leave empty
- Stunserver (option): stun.sipdiscount.com
- Codecs:
G.711 (64 kbps)
G.726 (32 kbps)
G.729 (8 kbps)
G.723 (5.3 & 6.3 kbps)
- If you have audio problems use a STUN server (e.g. stun.sipdiscount.com) with port 3478 (if supported by your device)
(*) I'm currently using sip1.sipdiscount.com which I found elsewhere suggesting a better performance (?). Did not impact the above problem.


Thanks in advance!

 

[Agile]

Update:
If I vary the parameter "Time Lapse For the registration" (which defaults to 5 minutes) the frequency of the error also varies. E.g. I have set this parameter to 15 minutes, and per consequence the error appears about every 15 minutes.
I understand that this setting is related to a SIP device re-registration with its SIP provider who in turn keeps a 'expire time'. Anyone know how this works with Sipdiscount (probably similar to all other providers in the Betamax family)?
Other tips?

Thanks, A.

 

[Agile]

Could someone perhaps describe how this 'Time lapse for the registration' function works in COMpact 5020 practice? E.g. the SIP server keeps this setting and I assume the client has to reregister before it expires (correct?). What determines this server setting? What is the value for SIPdiscount/VoIPbuster? The 5020 help-file suggests server setting and client re-registration are linked. How? If the client setting above determines the client reregistration time, why does it create a 1...2 minute outage when the COMpact 5020 reregisters?
A.

 

[PRIME_OS]

Don't know if this is of any help. I had a similar problem with another SIP provider and another SIP device (SwissIPCom and a Zyxel device) a few years back. I - partially - solved it by increasing the NAT keep alive intervals and re-registration to 60 minutes (3600). Even in this scenario, my calls got sometimes dropped. I don't see this problem with my shiny new 5020 though (provider is now also different). Back then, the provider told me that my device is not fully RFC compliant and that they could configure it to make it work but they refused to change their settings (I'm now using Sipgate on both sides of the Atlantic with my 5020)

Oh, and one more thing: Make sure you're not behind a symmetric firewall. I saw exactly the scenario you described behind my ZyXEL USG.