W3C home > Mailing lists > Public > xml-dist-app@w3.org > December 2006

Summary of one-way scenarios

From: David Hull <dmh@tibco.com>
Date: Fri, 01 Dec 2006 15:47:12 -0500
To: "xml-dist-app@w3.org" <xml-dist-app@w3.org>
Message-id: <457094D0.7090201@tibco.com>

Following is a list, not necessarily complete, of scenarios that have
come up in discussion.  IMO, each represents an instance of the one-way
MEP, always for the same basic reason: There is a sender, there is a set
of receivers and (normally) each receiver receives SOAP envelope
identical to the one sent, just as our current text specifies.

I'd be very interested to know whether others also feel these are all
instances of the one-way MEP.  If not, I'd like to know which ones
aren't, and on what basis.  It's fine (with me, at least) if some of
these can also be described using various combinations of other MEPs:

    * I send you a SOAP envelope using two tin cans and a string
      (2TCS).  You run the SOAP processing model on it (except where
      noted, assume in the following that at least the last party to
      handle a SOAP envelope runs the SOAP processing model on it).
    * I serialize a SOAP envelope, encrypt it to a blob, send the
      ciphertext to you via 2TCS.  You decipher it to get the SOAP envelope.
    * Ditto, but the 2TCS implementation (S2TCS?) does the
      encryption/decryption unbeknownst to me.
    * I send a SOAP envelope to an intermediary via 2TCS, which encrypts
      it to a blob and sends it to another intermediary via 2TCS.  That
      intermediary decrypts it and sends it to you via 2TCS.  My sending
      the plaintext and your receiving it is an instance of the one-way MEP.
    * Ditto, except the message is encrypted using WSS.
    * I email a SOAP envelope to your mailbox.  You retrieve it from
      your mailbox.
    * I email a SOAP envelope to xml-dist-app@w3.org.  Some number of
      people on the list retrieve it.
    * I email a SOAP envelope to "fellow W3C members".  My mailer
      expands that to several individual addresses and sends it to those
      mail boxes.  Some number of them retrieve it.
    * Ditto, but some of the addresses are mailing lists.
    * I send a SOAP envelope to your jabber ID as a <message>.  You
      receive it.
    * I send a SOAP envelope to your jabber ID as a <message>. 
      Unbeknownst to me, you're not there.  Instead, jabber sends it to
      your mailbox.  Later, you retrieve it.
    * I send a SOAP envelope in jabber <message> to a chat room.  Jabber
      forwards it to all participants in the room, who receive it.
    * I send a SOAP envelope via 2TCS.  There are two SOAP nodes
      listening on the receiving end.  Each runs the SOAP processing
      model on the envelope.
    * I publish a SOAP envelope on my favorite pub-sub transport.  Some
      number of subscribers receive it.
    * I publish a SOAP envelope on my favorite pub-sub transport.  As it
      happens, and unbeknownst to me, there is exactly one subscriber,
      who receives it.
    * I publish a SOAP envelope on my favorite pub-sub transport.  I
      know for a fact that there is exactly one subscriber (e.g., I know
      the subscriber created a unique one-off topic for the purpose).
      That subscriber receives the envelope.

The following, in my mind, are /probably/ instances of the one-way MEP
as well, but are of a more architectural than practical interest.

    * I send you an envelope in the request half of r-o-r.  You receive
      it and send back an ack.
    * I send you an envelope in the request half of request-response.
    * You send me an envelope in the response half of request-response.
      (or as a SOAP response in that MEP).
    * I send an envelope to the WSA "none" bit-bucket (by not sending it
      at all).
Received on Friday, 1 December 2006 20:47:19 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:23 GMT