H.450.3 interrogation procedure

[Henri Mäenpää]

Hani,

I will try to answer this, but I may not be 100% accurate, so I welcome
others comment.

>                 I have a question regarding tunneling and H.245 address.
>                 Suppose a terminal returned a connect message with a valid
> H.245 address (IP+port) together with tunneling set to TRUE.
>                 What should be the behavior of the receiving terminal in
> this case?

The calling terminal, in this case, would have the option to tunnel H.245 or
to open a separate H.245 connection.  The terminal may chose to establish a
separate H.245 session later in the call and, if it does decide to do this,
it shall use the address it received in the Connect.

>                 The real question is under what circumstances or why would
a
> terminal request tunneling and supply an H.245 address?
>                 Is it backward compatibility?

In this scenario, I don't think backward compatibility would be the reason.
If the caller did not support tunneling, the called endpoint would return an
H.245 address as normal.  In this case, the caller supports tunneling and
the called party does as well, but it also chooses to return the H.245
address to the caller.

There are advantages to switching out of a tunneled mode, but I see no clear
advantage in allocating a port and listening for an incoming connection when
tunneling is being used.  It certainly seems like a waste of resources.
But, perhaps there are environments where there is a 75% probability that
the call will switch out of the tunneled mode.  If that were the case,
providing this address early will reduce signaling.  I just don't think this
is very likely.  In either case, just stick that h245Address in your back
pocket and use it if you decide to switch out of tunneled mode.

Paul