[asterisk-biz] Asterisk Performance Results

[Gregory Boehnlein]

> Jim,
> 
> Thank you for publishing this; it's most useful to the Asterisk
> community. Did you make any measurements of voice quality (jitter,
> packet loss, subjective listening tests by a human) at high loads?From
> a brief read at your PDF document, I didn't see anything on this - my
> apologies if I've missed it.

Just to add some information to the discussion, we've recently begun
monitoring the quality of all of our SIP traffic using a product called
VQmanager (http://www.vqmanager.com). Basically, you take a Linux box,
install their software on it, then mirror the switch ports of your SIP
endpoints onto it. It collects stats on every call, provides Jitter,
Latency, Packet Loss, MOS and R-Factor scores for it as well as provides a
detailed breakdown of all of the SIP transactions. You can search calls by
endpoints, SIP URI and a bunch of other methods.

It's been very interesting for us as a tool to measure how our Asterisk
boxes are performing in comparison to the other equipment on our network and
since we bring our SIP carriers into the same boxes, we can measure the call
legs originating and terminating to them as well. It became very helpful in
the last few weeks for diagnosing a problem w/ call drops and lost packets
that appear to be related to mis-classified QOS rules.

I'm rambling a bit, but in summary, this tool has proven to be very
effective for troubleshooting issues, and I'd recommend taking a look at it.
It would be interesting to see how all of the associated stats would be
affected as you scaled the box.

P.S. I don't work for Advent and have no relation to VQmanager, other than I
think it's a kick ass product that fills in a lot of gaps in network
analysis.