Re: Help needed on VoIP project

[Andrew Rutherford <>]

At 7:54 AM -0500 1/3/04, Jon Zeeff wrote:

> This is somewhat off topic, but to me, the most important thing to 
> test is resistance to packet loss and jitter.

Yes, the difference in phones can be significant. We were called in 
to look at a 3com quality issue over an 802.11b link. It turned out 
the switch driving the wireless link was at fault (unable to cope 
with more than 100 packets with prioritization set per second - 
turned prioritization support off and the problem went away), with on 
the order of 12% packet loss.

When we went to site we didn't know that, and took some Cisco 7940 
phones with us because of their better statistic reporting. With two 
people on either end of the link, with a 3com handset (sorry, didn't 
record the model) in one ear and a Cisco 7940 handset in the other, 
we got scratchy but understandable quality voice on the Cisco 7940, 
but completely incomprehensible on the 3com. Sniffing showed the 
incoming packet flows were pretty much identical in terms of loss and 
jitter between the two audio streams, so the quality you have to test 
is the audio coming out of the handset, not the packets going in.

--
Andrew Rutherford      
sip:
      244 Pirie Street
Iagu Networks          tel:+61-8-8425-2255       Adelaide SA 5000
http://www.iagu.net/   
mailto:
   Australia