Re: Purpose of terminology I-D

[Walid Dabbous <Walid.Dabbous@sophia.inria.fr>]

Scott Michel wrote:

>> Feed		Source traffic that ingresses UDLN through a transit node.
>> 		(AKA: Traffic, packets, ... :-)
>>
>> Feed-ingress	A node which transmits packets into a UDLN.
>> 		(AKA: Feed, in satellite terminology.)
>>
>> Feed-egress	A node which receives packets from a UDLN.
>> 		(AKA: Receiver, in satellite terminology.)
>

Steve Deering wrote:

>Hmm, your definitions seem to center around a mental model of an active
>flow of packets, and then you are naming things with respect to that flow
>of packets.
>
>I'd be happier with more "static" terms that didn't depend on the actual
>movement of packets.  

I agree. We need a terminology for the "nodes". The "traffic"
itself is composed of "normal" IP packets. We used the term
Feed to designate the node transmitting packets
into a UniDirectional Link. I think "Feeder" is better.

>In particular, I'd suggest the important concepts
>to name are:
>
>    non-transitive link 	- a communication channel (a "link",
>				  from IP's point of view) on which not
>				  all attached nodes (IP hosts and/or
>				  routers) are capable of both trans-
>				  mission and reception.  Examples are
>				  some satellite transmission systems and
>				  unidirectional cable TV distribution
>				  systems.
>

non-transitive is better than unidirectional if we really
want to consider links where rcv-only, Trx-only
and bidirectional interfaces are connected.
Otherwise, the term unidirectional is more explicit.

>    transmit-only interface	- the different types of attachments a
>    receive-only interface	  node may have to a non-transitive link.
>    bidirectional interface
>
>
>Lately, in another context, I've been trying to come up with a new, less
>verbose, and less overloaded term for "interface"; the term I've settled on
>is "tap" (the point at which a node taps into a link).  So my candidate
>terms for the last three concepts above would be:
>
>    send-only tap
>    rcv-only tap
>    two-way tap
>
>I meant to add:
>
>When you wish to refer to a node attached to a non-trans link according
>to the capability of its interface, you would use the terms:
>
>    send-only node
>    rcv-only node
>    two-way node

two-way nodes attaches to non-trans links?
do you mean VSAT stations with send and receive capabilities 
oriented to the satellite? In "skyplex" like services these
stations could send a non multiplexed flow and receive
the down-stream multiplexed on board.

do you think that two way nodes attached to a non-trans
link require specific treatment? or is it just for
complete terminology purpose/

we need terminology for :


			    
                              
                    (1)                           (5)
                    ---- (2)    (3)           (4) ----
                    |  |--X---------------->>--X--|  |  
                    ----                          ----
                     /\ (8)                        /\ (6)
                     ||                            ||
                     \/                            ||  
                      -----------------------      ||
                     |                       |<=====
                     |          (7)          |  
                     -------------------------

what do you think of:

(1): Feeder, send-only node, ...
(2): send-only tap, send-only interface,...
(3): Unidirectional Link
(4): Receive only tap or interface
(5): Receiver, receive-only node (but it's long...)
(6): receiver bidirectional interface
(7): bidirectional network, backchannel, ...
(8): Feeder bidirectional interface.




Walid Dabbous
http://www.inria.fr/rodeo/personnel/Walid.Dabbous/me.html
INRIA U.R. de Sophia Antipolis      | Email : dabbous@sophia.inria.fr  
2004, Route des Lucioles BP 93      | Phone : +33 4 93 65 77 18
06902 Sophia Antipolis CEDEX France | Fax   : +33 4 93 65 77 65