[asterisk-dev] RTP trunking - 58% savings on media bandwidth?

[Benny Amorsen]

"Stanisław Pitucha" <viraptor at gmail.com> writes:

> Also when you're calculating the cost, there's also increased cost of
> processing the packets - some people already need to unpack the
> stream, so it's not a big difference for them, but those who do the
> simple packet forwarding (was it packet2packet, or something like
> that?) will get a lot more work to do...

At some point it really becomes a question of whether the benefit is
worth all the complexity. Essentially a properly done RTP trunk would
end up reinventing UDP except with packet bundling.

Some link types (e.g. WiMAX) already bundle small packets, and for
others (e.g. PPP) you can reduce the per-packet cost to almost nothing
with header compression. This will help all packets, not just RTP.

There are three link types where RTP trunking would help a lot. One is
ethernet, but RTP uses so little bandwidth compared to even 100Mbps
that the benefit is negligible. Another is WLAN, where the per-packet
overhead is bad, but anyone doing RTP trunking across a WLAN is asking
for trouble anyway (switch to WiMAX). The last one is classic
ADSL/ADSL2 with ATM encapsulation and without PPP header compression.
That one is a genuine problem. Fortunately it is going away, being
replaced with pure-IP VDSL, perhaps even with PPP header compression
if we're lucky.

In conclusion, I think RTP trunking is a solution for the problems of
yesterday. It will take a few years to get wide deployment and by then
it will be obsolete.


/Benny