OK, now I see your point.
>> Remember, with voice, we need not only short delays, but predictable ones as well. This is the challenge when integrating legacy routers into newer networks.<<
Perhaps the dedicated-ness, as in dedicated IP network, then, should apply to the router$ as well as the links. And much of the problematic variability that you cite is also compensated for, to some degree, in local gateway and terminal device auto-sensing and algorithm-driven buffering.
I must say, however, that your observation is a refreshing (in that it gives cause for thought) contrast to prevailing and contemporary wisdom that tends to downplay the significance of the impact of voice traffic. I have read numerous times by the likes of Sidgmore, and others, that voice is but a trickle to be dealt with, and that it will fit between the cracks.
I myself have done some superficial analysis of this, and indeed have written posts to this effect, looking at the attributes of the UDP characteristics, and have come to the same conclusions. But it really does depend on scale, and proportions/allocations of resources, and many other factors that usually get lost in overly generalized discussions on the subject. Your point is well taken, thanks.
Frank C.
|
