W3C home > Mailing lists > Public > ietf-discuss@w3.org > December 2001

Re: Requirements for reliable message delivery

From: John Ibbotson <john_ibbotson@uk.ibm.com>
Date: Tue, 4 Dec 2001 08:46:19 +0000
To: discuss@apps.ietf.org
Message-ID: <OFB5E134C3.80737B3D-ON80256B18.003029A3@portsmouth.uk.ibm.com>
                    John Ibbotson                                                                             
                                         To:     "Marshall T. Rose" <mrose+mtr.netnews@dbc.mtview.ca.us>      
                    12/04/2001           cc:                                                                  
                    08:42 AM             From:   John Ibbotson/UK/IBM@IBMGB                                   
                                         Subject:     Re: Requirements for reliable message delivery(Document 
                                         link: John Ibbotson)                                                 

Comments below:

XML Technology and Messaging,
IBM UK Ltd, Hursley Park,
Winchester, SO21 2JN

Tel: (work) +44 (0)1962 815188        (home) +44 (0)1722 781271
Fax: +44 (0)1962 816898
Notes Id: John Ibbotson/UK/IBM
email: john_ibbotson@uk.ibm.com

                    "Marshall T. Rose"                                                                                    
                    <mrose+mtr.netnews@dbc.mtv       To:     John Ibbotson/UK/IBM@IBMGB                                   
                    iew.ca.us>                       cc:     "Discuss Apps" <discuss@apps.ietf.org>, "Marshall Rose"      
                    11/28/2001 05:17 AM              Subject:     Re: Requirements for reliable message delivery          

guys - help me out with something. after reading your I-D, i can't help but
think that you can meet your requirements by:

1. shifting more responsibility (e.g., loss detection) to the endpoints.
<JBI> Absolutely. From a reliability point of view, the responsibility is
not to lose the message at
<JBI> the endponts. In this case, the sending endpoint must have a stored
copy available until it knows
<JBI> unambiguously that the receiving endpoint has successfully stored the
received message persistently.
<JBI> It can then delete its copy. In cases of failure, it can retry and
synchronise with the receiving
<JBI> endpoint.
2. using an application-layer relaying service with deterministic delivery
semantics (e.g., apex with some of the party pack options).
<JBI> I need to read more on APEX before commenting authoritatively on this
:-) But having seen your
<JBI> presentation to the XML Protocol WG in Boston I think APEX could
provide the basis for a reliable
<JBI> protocol. There is still the question as to what is to be done in the
short term and the feedback we get <JBI> from our customers is that HTTP
will be around for a long time and they'd like to make that reliable -
<JBI> warts and all !
what am i missing here?

Received on Tuesday, 4 December 2001 10:52:55 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:38:01 UTC