I’ve noticed a significant difference between SIP and IAX when doing an Echo test to my home Asterisk box from outside. SIP is much more responsive. I’ve tested with several different soft phone clients and the results are always the same.
Why is this? Is it because IAX’s use of just one port slows communications down? I was such a proponent of IAX but now I wonder if the SIP/NAT headaches are more than made up for by SIP’s significantly better performance.
Has anybody else noticed this difference? Can somebdoy corroborate my results with their onw tests?
No, I think your latency differences you are seeing are coming from the soft phone clients you are using.
If you compared hardphones, you would find there were no perceptible difference in latency because of the protocol, either SIP or IAX2.
There are also tradeoffs, jitterbuffering adds to the latency but can improve voice quality over a non-ideal IP connection. Most of these ‘features’ can be turned off/on.
[quote=“lonnie”]No, I think your latency differences you are seeing are coming from the soft phone clients you are using.
If you compared hardphones, you would find there were no perceptible difference in latency because of the protocol, either SIP or IAX2.
There are also tradeoffs, jitterbuffering adds to the latency but can improve voice quality over a non-ideal IP connection. Most of these ‘features’ can be turned off/on.
Lonnie[/quote]
I would say it could be your endpoints as well, as I am not familiar with any inherit latency that IAX2 brings about over SIP. Having said this, I have never run a test like this to do a comparison.
well, i’ve done another test. one of my voip providers offers both IAX2 and SIP connections. i’ve configured a trunk in each protocol from my asterisk box to the provider. the provider also has a number to call for an echo test. so at home i pick up my two extensions (analog phones connected through a Sipura 2000 ATA), dial out each trunk at the same time and do the echo test. i have both phones to my ear. when i make a sound, the echo clearly returns to my right ear before my left ear. and the right ear is using the SIP trunk. so my earlier findings are corroborated.
but maybe it’s the jitterbuffer as you say, so i’ll look into that.