[Asterisk-Users] Resolving QOS problems

[Chris Miller]

I'm looking for advise on troubleshooting QOS problems. After much 
searching and reading online (Google, Voip-Info Wiki, etc.) I don't feel 
any closer to finding the right tools to solve my problem. Any info you 
would like to share would be much appreciated, and I'm sure the thread 
will server others in the future.

The problem :
-------------

I'm having intermittent problems with the audio cutting out on calls. At 
the same time the audio problems occur, I often see these in the "full" 
log :

Received iseqno 122 not within window 123->123

These range from sounding like bad cell phone calls, to the audio track 
cutting out in one or both directions for up to 20-30 seconds.

I also see dropped calls that seem to be a result of the IAX connection 
going away.

The environment :
-----------------

I've got an * server located at a data center with good connectivity, 10 
hops to my IAX provider, and ~34ms ping times. They (IAX provider) use 
Cogent which concerns me a bit, but I'm not ready to jump to conclusions 
just yet.

My IP phone is connected via "enhanced" DSL (static addresses, no PPPoE) 
and I'm 12 hops away from my * server. My DSL provider has direct 
connectivity and peering agreements with the data center my server is 
located in. I've set QOS priority on the LAN port (Linksys router) the 
phone is connected to, and I've dropped the MTU to 576 as suggested for 
lower speed links. (1.5Mbs/384kbps in my case). Both these changes 
seemed to make an improvement over previous calls. Currently I don't 
believe the bulk of my problems to be between the phone and the * 
server. testyourvoip.com tests consistently show a 4.4 score (the 
maximum for ulaw) and rarely shows errors.

Ulaw is the codec used for both the SIP calls and IAX trunk.

What I'm looking for :
----------------------

I'm trying to determine the cause and location of the problem between my 
* server and the IAX provider (and possibly my IP phone), and see what 
if anything I can do to reduce the occurrence of these drop outs. I'm 
looking for a couple of things :

1. A method of monitoring RTP/IAX traffic QOS at the PBX in real time.
2. Tools that might be used to determine the location of the problem.
    I.E. An RTP/IAX "traceroute" tool.

What I'm hoping to find is something that either integrates directly 
with *, or captures live RTP/IAX traffic and provides real time 
statistics on calls.

What I've found :
-----------------

I saw Telchemy's VQMON_EP product, but it's unclear how it would work 
with Asterisk. Many other companies in this market seem to leverage off 
of Telchemy's products.

http://www.telchemy.com/partners.html
http://www.voiptroubleshooter.com/tools/voiptr_tools.htm

All of the products above seem to be aimed at large enterprises with 
deep IT pockets. I wouldn't mind ponying up a reasonable sum for a tool 
that does the job, but I lack the time to thoroughly evaluate everything 
that may be out there.

I haven't found much on the open source front. I've seen "Windows RTP 
Quality Monitor" which might be useful, but it's beta and hasn't been 
updated in over a year.

It seems to me that Ethereal might be integrated with a graphical tool, 
and if nothing else provide postmortem statistics on a phone call.

Request for comments :
----------------------

What are people using to troubleshoot these problems? What commercial 
software works for you? What open source projects are you using? Do you 
have suggestions on projects that might be glued together to provide 
this functionality?

Thanks in advance.

Chris