RE: Nature of Application Protocols!

But, of course SMTP isn't stateless (if that's what you mean by
"connectionless"). Offhand, I can't think of any IETF application protocols
that are both stateless ansd based on the push model  If you relax the layer
7 requirement and look at IP core protocols, that's another matter. One
protocol that is an application layer protocol that supports stateless push
is Health Level 7 (or just HL7, see http://www.hl7.org). I've developed a
number of interfaces baseed on a request/reply semantic similar to HTTP, but
there are messages such as Master File Notification (MFN) that are used to
send unsolicited updates to tables (e.g., to record the fact that a patient
has been admitted). There is an acknowledgement protocol (I send you an MFN,
you then send me Commit Ack (CA) to tell me you have the message and then a
Master File Acknowledgement (MFK) to indicate that it has been processed,
but unlike SMTP, this doesn't occur in the context of a single "session").

===
Gregory Woodhouse  <gregory.woodhouse@med.va.gov>
System Design & Development
+1 510 768 6862



-----Original Message-----
From: Mark Nottingham [mailto:mnot@mnot.net]
Sent: Thursday, August 29, 2002 7:36 AM
To: ietf-http-wg@w3.org
Subject: Re: Nature of Application Protocols!




> Well, SMTP is definitely push. FTP can be used both ways.

As can HTTP. Poor, neglected PUT...

Received on Thursday, 29 August 2002 12:29:20 UTC