[ih] Nomenclature

[Stephen Casner]

On Mon, 14 May 2012, Larry Sheldon wrote:
> On 5/14/2012 2:07 AM, Stephen Casner wrote:
> > On Sun, 13 May 2012, Larry Sheldon wrote:
> >
> > > (Along the way, a new question for me to chase:  Does VoIP use
> > > IP? TCP/IP?  UDP? (Wiki seems to say Yes, No, and No.
> >
> > VoIP implementations that use RTP (Real-time Transport Protocol)
> > use a stack of RTP/UDP/IP.  Where does Wiki say No to UDP?
>
> http://en.wikipedia.org/wiki/VoIP _seems_ to say that the VoIP (as
> does the popular name) protocol(s) runs under IP.

I read that introduction as saying that the voice is carried over IP
networks, but it does not preclude other protocols between the voice
and IP.  Since there is no codepoint assigned for "voice" to go in the
Protocol field of the IP header, and since some framing and other
services are needed for the voice transport, there are other protocols
required in the stack.

> (Side and mostly irrelevant note -- the last time I dealt with "RTP"
> it was the RJE protocol over dial-up lines to IBMish hosts from
> UNIVAC machines which didn't support transparent operation (for
> I-know-not-why reasons), so we had to use assembler code to emulate
> transparency.)

Indeed, we who worked on Real-time Transport Protocol did receive a
communication from some IBM folks about the acronym clash, but there
was some debate about which came first, and we decided to proceed
without making a change.

> Let me me emphasis--the wiki article must have been written by a
> professional politician because, while it seems to say what I said,
> I can see how it says something else.  I just don't know what.

The article does reference RTP in the section "Protocols", but that is
it.  I would prefer to see more, for example in the milestone list.
H.323 is described in more detail, but it is not mentioned that H.323
incorporates RTP (as part of H.225.0) to packetize the voice.

                                                        -- Steve