- From: Mark Needleman - DRA <mneedlem@dra.com>
- Date: Thu, 21 Dec 2000 13:20:59 -0600 (CST)
- To: Dick Brooks <dick@8760.com>
- cc: john_ibbotson@uk.ibm.com, xml-dist-app@w3.org
----- Message Text ----- Dick so perhaps i got confused about the word transport in DS5 - DS1 and 2 are no acks at the MHS level and 5 is? - maybe the wording can be made clearer to indicate at what transport layer things are happening at? mark On Thu, 21 Dec 2000, Dick Brooks wrote: > Mark, > > >I guess I have one comment about this and DS2 - in DS5 there is the > >specific mention of a transport level ack - so does DS1 and DS2 explicitly > >assume there is no ack to the msg at any level. It seems to me that > >defining how lower layers on which XP may run behave is not really in > >scope and I think we need to be clear that the lack of acknowledgment we > >are talking about here is at the XP protocol level > > > >Trying to mandate how lower level protocol stacks behave seems a little > >out of scope > >Or is this really saying there is no ack at any level - in which case I > >need to understand why > > ebXML defines a "Message Service Handler layer" (MSH) which provides a > specific set of "message processing services" to upper layers. The MSH rides > on top of application level transport mechanisms (HTTP, FTP, SMTP) which in > turn ride on top of reliable network stacks (e.g. TCP/IP) which use TCP > level "acks" to ensure reliable delivery. > > The acks referenced in [DS5] are issued by a receiving MHS to indicate that > a message service handler has successfully "accepted delivery of" a message > (header+payload). DS1 and DS2 use cases do not require the MSH to > acknowledge receiving a message, however reliable delivery is assumed in > both DS1 and DS2 cases through the use of TCP/IP based protocols. > > Dick Brooks > http://www.8760.com/ > >
Received on Thursday, 21 December 2000 14:21:14 UTC