Fwd: Strategies for testing latency
Hi folks, This may be a slightly off-topic FreeSWITCH question, but here goes... I am having issues trying to figure out the source of latency in circuits we LCR to, it seems to be somewhat of a guessing game. I am embarking on creating a tool to measure this latency so we can test carriers more rapidly when deciding whether to use them or not. Before I re-invent the wheel, I'd love to know what others do to ensure latency is at a minimum on the circuits they buy from carriers OR if there are specific things to tweak in FreeSWITCH that I'm unaware of that somehow would help this. I can't imagine there are, but maybe I've missed something. It is my belief that some of the carriers we are using are behind Asterisk boxes that are always taking on media and are adding an encode/decode step to the path, resulting in some latency, as I can't find any evidence of late packets or high jitter on our box <-> carrier. For us, latency seems most noticable when we have callers on cell phones using a conference bridge. The latency nears 800-1000ms from when a person speaks to when they are heard on the other cell phone. I have tried to tweak our own conference bridge settings but I really don't think that's going to make a difference if my theory of the issue being upstream is correct. VoIP <-> VoIP conferencing has none of this issue. Any experience that folks can offer here is appreciated. Obviously getting a direct RTP stream from the carrier doing the conversion to PSTN would be best, but it has a tendency to require a lot of minutes or be expensive to do yourself. Thanks, Darren Schreiber Co-Founder - the 2600hz Project (415) 886-7900 - www.2600hz.org<http://www.2600hz.org/> Join us in January for FreeSWITCH Training in Australia! Visit www.voipkb.com<http://www.voipkb.com/> for more information
Darren, Relatively anyone can set themselves up as a "Tier X" carrier. Your hypothesis is most likely correct; many carriers are running Asterisk (or who knows what) and ( somewhat unnecessarily) handling media for billing, topology hiding, etc. That's why I typically only deal with carriers I know are Tier 1 LECs. I feel much better when they own and control the media gateways and TDM circuits. Anyways, a while back I started a project to investigate the actual media quality provided by carriers. I found that (in many cases) the IP path between our endpoint and the remote end (per the media stream/SDP) would be fine but the quality would still be terrible. I needed a tool to measure the quality of the actual audio stream end to end (as experienced by the actual user). I know the ITU and others have devised similar standards and tools for doing the same thing but I wanted something that could be easily tweaked and extended for various needs so I created recqual: http://admin.star2star.com/recqual/RECQUAL-README.txt There's some other stuff on google if you'd like to check it out but it should be fairly easy to adapt recqual to give end to end delay numbers... On Tue, Nov 2, 2010 at 1:43 PM, Darren Schreiber <d@d-man.org> wrote:
Hi folks,
This may be a slightly off-topic FreeSWITCH question, but here goes...
I am having issues trying to figure out the source of latency in circuits we LCR to, it seems to be somewhat of a guessing game. I am embarking on creating a tool to measure this latency so we can test carriers more rapidly when deciding whether to use them or not.
Before I re-invent the wheel, I'd love to know what others do to ensure latency is at a minimum on the circuits they buy from carriers OR if there are specific things to tweak in FreeSWITCH that I'm unaware of that somehow would help this. I can't imagine there are, but maybe I've missed something. It is my belief that some of the carriers we are using are behind Asterisk boxes that are always taking on media and are adding an encode/decode step to the path, resulting in some latency, as I can't find any evidence of late packets or high jitter on our box <-> carrier.
For us, latency seems most noticable when we have callers on cell phones using a conference bridge. The latency nears 800-1000ms from when a person speaks to when they are heard on the other cell phone. I have tried to tweak our own conference bridge settings but I really don't think that's going to make a difference if my theory of the issue being upstream is correct. VoIP <-> VoIP conferencing has none of this issue.
Any experience that folks can offer here is appreciated. Obviously getting a direct RTP stream from the carrier doing the conversion to PSTN would be best, but it has a tendency to require a lot of minutes or be expensive to do yourself.
Thanks,
Darren Schreiber
Co-Founder - the 2600hz Project
(415) 886-7900 - www.2600hz.org
Join us in January for FreeSWITCH Training in Australia!
Visit www.voipkb.com for more information
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
-- Kristian Kielhofner http://www.astlinux.org http://blog.krisk.org http://www.star2star.com http://www.submityoursip.com http://www.voalte.com
Latency on cellphones will always be a little higher. If you put say two users of Sprint next to each other and then call between them its a good 2 seconds. /b On Nov 2, 2010, at 12:43 PM, Darren Schreiber wrote:
For us, latency seems most noticable when we have callers on cell phones using a conference bridge. The latency nears 800-1000ms from when a person speaks to when they are heard on the other cell phone. I have tried to tweak our own conference bridge settings but I really don't think that's going to make a difference if my theory of the issue being upstream is correct. VoIP <-> VoIP conferencing has none of this issue.
Yep I second Brian completely. It's inherent to cellphones. I spent lot of time on this issue, and seems there's nothing you can do :( -giovanni On 11/2/10, Brian West <brian@freeswitch.org> wrote:
Latency on cellphones will always be a little higher. If you put say two users of Sprint next to each other and then call between them its a good 2 seconds.
/b
On Nov 2, 2010, at 12:43 PM, Darren Schreiber wrote:
For us, latency seems most noticable when we have callers on cell phones using a conference bridge. The latency nears 800-1000ms from when a person speaks to when they are heard on the other cell phone. I have tried to tweak our own conference bridge settings but I really don't think that's going to make a difference if my theory of the issue being upstream is correct. VoIP <-> VoIP conferencing has none of this issue.
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
-- Sent from my mobile device Sincerely, Giovanni Maruzzelli Cell : +39-347-2665618
participants (4)
-
Brian West
-
Darren Schreiber
-
Giovanni Maruzzelli
-
Kristian Kielhofner