[Asterisk-Dev] IAX Spec online

[Edwin Groothuis]

On Tue, Apr 26, 2005 at 08:52:36PM -0700, Kenny Shumard wrote:
> I've finished a working version of the IAX specification and posted it
> online at
> 
> http://splurge.peoples-wireless.com/iax/iax.txt
> 
> I'm very willing to listen to constructive criticism regarding
> additions or alterations to the spec as presented. The more people
> that read and comment on this, the sooner we can (hopefully) get it
> into the official standards track.
> 
> I know there are some parts that aren't completely correct, so please
> help me out by pointing them out -- gently. : )

    4.1  Full Frames

       Full frames can send signaling or media data.  Generally full frames
       are used to control initiation, setup, and termination of an IAX
       call, but they can also be used to carry stream data (though this is
       generally not optimal).

       Full frames are sent reliably, so all full frames require an
       immediate acknowledgment upon receipt.  This acknowledgment can be
       explicit via an 'ACK' message (see Section 8) or implicit based upon
       receipt of an appropriate response to the full frame issued.

The term reliable / unreliable is wrong. If it is sent reliable,
the sender guarantees that the packet arrives. I would replace it
with:

   All full frames must be immediate acknowledged upon receipt.
   This acknowledgment can be explicit via an 'ACK' message (see
   Section 8) or implicit based upon receipt of an appropriate
   response to the full frame issued.

Note the RFC version of "must".

Same with 4.2:

    4.2  Mini frames

       Mini Frames are so named because their header is a minimal 4 octets.
       Mini frames carry no control or signaling data; their sole purpose is
       to carry a media stream on an already-established IAX call.  They are
       sent unreliably.  This decision was made because VOIP calls typically
       can miss several frames without significant degradation in call...

I would replace it with::

       to carry a media stream on an already-established IAX call.
       Mini frames should not need to be acknowledged upon receipt.
       This decision was made because VOIP calls typically...

Note the RFC version of "should not". That indicates that they *can*
be acknowledged if wanted, but it is not necessary. Is that the
expected behaviour? If not, it should be replaced with "must not
be".


Edwin

-- 
Edwin Groothuis      |            Personal website: http://www.mavetju.org
edwin at mavetju.org    |          Weblog: http://weblog.barnet.com.au/edwin/