[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Subject: Re: Need volunteer to draft definition of reliable messaging,wasRE:reliable messaging - hop by hop
Dan, Yes, SMTP should be considered unreliable multihop at the SMTP level. There are at least these possibilities for singlle hop at the ebXML level: 1. If the SMTP path is a single ebXML hop, then end to end reliable messaging should take care of any message losses inside SMTP except possibly for the "pathological" cases such as network partition at an unfortunate time that we have already discussed. 2. SMTP is too unreliable for any form of ebXML reliable messaging. I have been assuming possibility (1). What do you think? For multihop ebXML on top of multihop SMTP, we have multiple ebXML hops each of which might be over multiple SMTP hops. I'm not sure what to think about this but I suspect that if RM is end to end at the ebXML level, it might work. Regards, Marty ************************************************************************************* Martin W. Sachs IBM T. J. Watson Research Center P. O. B. 704 Yorktown Hts, NY 10598 914-784-7287; IBM tie line 863-7287 Notes address: Martin W Sachs/Watson/IBM Internet address: mwsachs @ us.ibm.com ************************************************************************************* Dan Weinreb <dlw@exceloncorp.com> on 09/03/2001 10:50:19 PM Please respond to "Dan Weinreb" <dlw@exceloncorp.com> To: Martin W Sachs/Watson/IBM@IBMUS cc: david@drummondgroup.com, rberwanger@btrade.com, ebxml-msg@lists.oasis-open.org Subject: Re: Need volunteer to draft definition of reliable messaging,was RE:reliable messaging - hop by hop Date: Mon, 03 Sep 2001 21:01:49 -0400 From: Martin W Sachs <mwsachs@us.ibm.com> MWS2: If the partition happens between the time the To party MSH reads the message and the time it sends the ACK, for a single hop, the To MSH will know that it couldn't send the ACK and (hopefully) retry later. Actually, that brings up an important question about our failure model. In a single-hop scenario, if the communication protocol is HTTP, the sender does know that it can't send. But in the case of using Email (SMTP) as a communication protocol, the communication layer can *itself* be multi-hop (using store-and-foward, a.k.a. "indirect", routing, perhaps directed by MX records from the DNS, etc), and so it may not be clear to the sender whether the message could or could not be sent.
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Powered by eList eXpress LLC