RE: TCP Mapping

["Liu, Hong" <Hong.Liu@neustar.biz>]

Eric,

The simple answer to both your questions at the end of your email is "yes"
and the "P" bit is set as shown in your diagram. 

The semantics of the "P" bit is subtlely different than yours (as I
interpret it). I envision that the "P" bit is set only for server-pushed
unsolicited data, _after_ the EPP session has been established. So it does
not apply to the <greeting> message sent by the server during session setup.
(See my response to your other email.)

The handling of server message queue is a tricky issue if the "server" and
the "client" engage in both "push" and "poll". But that is orthogonal to the
current topic. I would defer that discussion in a later time...

Thanks for coming up with the text to illustrate the case.

--Hong

-----Original Message-----
From: Eric Brunner-Williams in Portland Maine
[mailto:brunner@nic-naa.net]
Sent: Saturday, June 29, 2002 1:46 PM
To: Liu, Hong
Cc: 'ietf-provreg@cafax.se'; brunner@nic-naa.net
Subject: Re: TCP Mapping 


Hong,

State-transition occurs in the registry,
the "server" initiates an event,
either
	o enqueueing a message for delivery via a registrar-initiated
	  instance of communication (which might have session semantics,
	  and which might be via tcp transport),
or
	o enqueueing a message for delivery via a registry-initiated
	  instance of communication (which might have session semantics,
          and which might be via tcp transport).

The first mechanism has found expression in the core drafts, the second
has not.

The proposal to put directionality (the initiator (the event)) into an
application PDU (your private pun on a header bit) puts directionality
discovery in the application.

Just to be sure I got your proposal, is this what you have in mind?
I've given lines that have changes "wedgies".

(from draft-ietf-provreg-epp-tcp-04.txt)
4. Datagram Format

  The data field of a TCP datagram MUST contain an EPP datagram.  The
  EPP datagram contains two fields: a 32-bit header that describes
>	the P-bit (is the message "PUSH" or "PULL")
> and
>	the 31 bits indicating the total length of the datagram,
  and the EPP XML instance.

  EPP Datagram Format (one tick mark represents one bit position):

     0                   1                   2                   3      
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>   |P                          Total Length                        |   
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                         EPP XML Instance                      |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+//-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

> Push Bit (1 bit): When set, the message was not initiated by the 
> initiator of the connection.
> Total Length (31 bits): The total length of the EPP datagram measured
  in octets.  The octets contained in this field MUST be included in the
  total length calculation.

(end of excerpt from -04)

Is the octet in network byte order? Which bit is the P-bit?

Eric