Priscilla Oppenheimer wrote: > > Janó van Deventer wrote: > > > > > > > On routers, the few nanoseconds saved by outputting voice > > first > > > are irrelvant compared to the hundreds of milliseconds to > > reach > > > the satellite. > > > > > Yes maybe, but won't you run into trouble with jitter? > > A small variation in delay (jitter) when the delay is huge > anyway might not matter? > > > And then also I think that if you don't have some kind of > > prioritization/fragmentation to take care of big data packets > > you can also have a problem with smaller sized voice packets. > > I'm thinking here in the line of FRF.12 (fragmentation for > > voice on frame-relay networks). I don't know what's a similar > > technique for satellite networks... > > Back to my bus station analogy. You're now talking about > letting the San Francisco bus go ahead of the train to > Philadelphia to save a few minutes, when it takes days to get > to San Francisco anyway. > > It's a good question really. Is it worth it to do link > fragementation and special queuing to improve voice and video > on a satellite link? I think he said the relevant direction is > a 1 Mbps link. Anyone want to comment on this? Please.
Priscilla, Since you asked so nicely... Not that I'm aware of. Shamefully, VoIP is one of those technologies I haven't even tried to stay current with. In general, any protocol that involves acknowledgements will need to be monitored closely and possibly tweaked when SATCOM is in the mix (just as is the case with any other type of LFN). I guess it's a pretty safe assumption that VoIP uses UDP. So I doubt acks are an issue. And I agree with your assessment that the delay associated with fragment size is very small relative to SATCOM delay. I suppose one could take the approach that any little bit helps. But the human ear isn't likely to discriminate the very small advantage, if any advantage at all is realized. Having said that, if large segments/fragments/whatever were introducing substantial jitter, thereby degrading voice quality, fragmentation might be of some help (but that would hold true in any situation - it wouldn't be unique to SATCOM applications). I don't know how much buffering is involved where the typical VoIP solution is involved. I suspect it would have to be a tradeoff, as too much jitter-smoothing buffering would cause noticable delay in delivery. Just a few rambling thoughts from someone who knows SATCOM quite well, legacy voice to fairly well, and next to nothing about VoIP (yet)... Scott > > PPP has fragmentation and interleaving, so he could use that on > the satellite link. > > > > > > > > > > > > PS. Nice to see fellow SAfricans on the forum. > > > > > > I thought you were from Brazil?! Now I am curious! :-) > > > > > I'm posting from sunny Brazil, that's right! I met a Brazlian > > girl in SA and made the big decision to move to Brazil, learn > a > > new language, and the rest is history. > > Sounds exotic! One great thing about this list is the ability > to communicate with people all over the world. It's great to > have you here. > > Regards, > > Priscilla > > > > > Kind regards, > > Janó > > Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=71851&t=71706 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
