FW: Internet2 Members Update - June 2004

["Paulo Henrique Aguiar Rodrigues" <>]

Ben,
 
If you need any information on the tools we have been developing, I would 
very be glad to provide.
As I mentioned in our last encounter in the VoIP workshop in Indianapolis, we 
have deployed an H.323 infrastructure in Brazil. Furthermore, we have 
developed objective measurement tools based on Alan Clark's porposed 
extensions to the E-model. We implemented our own code and this code has been 
ported to the H.323 Beacon (Prasad Calyam has been responsible por the 
porting). I have met Prasad Calyam (Ohio University, Oarnet) during the 
workshop and our cooperation started from there.
At the moment we do have two tools, both using our analytical code:
 
A - A modified Openphone client, used in pairs, allowing us to plot all 
statistics (jitter, delay, loss in jitter buffer, network loss, E-model 
parameters, and final MOS) along a specific call. Very nice, as at the end of 
the call, the full stats are obtained and potted at one side. The call is a 
real one, not a pre-recorded message. Quality is obtained in both directions. 
We do use RTP and RTCP log information.
 
B - A modified answering machine H.323 code (Openam), running as a client at 
destinations, which allows pre-programmed simultaneous calls along some 
period of time to different destinations. In this case, we do use a 
pre-recorded message that is played at the source and recorded at 
destinations. In the end, statistics plots (using same code previously 
mentioned) are collected for all measurements. The processing of all data 
requires file transfers from destinations to the source and local processing. 
Semi-automatic scripts do the job. We do have a search engine (losses, 
jitter, delay) integrated to th Web to locate interesting calls.
 
At the moment, we are looking for some open SIP implementation able to 
perform as Openam or similarly (we need to have detailed RTP/RTCP log). The 
log has been slightly modified in Openam to provide more frequent reports.
 
If you need further information or have any hint concerning SIP open code, 
please drop me a message.
 
Paulo Aguiar     

-----Original Message-----
From: Ben Teitelbaum 
[mailto:]
Sent: sábado, 12 de junho de 2004 23:52
To: VoIP Working Group
Subject: Re: Internet2 Members Update - June 2004


Does anyone on the VoIP WG have experience with this VoIP quality
monitoring project?  -- ben

> - H.323 Beacon
> E-module code, written by RNP VoIP group in Brazil, is being integrated into
> the H.323 Beacon, which will enable the H.323 Beacon to make automated
> assessments of user-perceived quality for various audio codecs. It also can
> be used in measurement infrastructures for regularly scheduled tests to
> determine voice grade capabilities on the Internet backbones.
> http://www.voip.nce.ufrj.br/index_caller_en.htm