Folks,
re: draft-khanna-smtp-mail-transfer-reliability-00.txt
My first thought was to dismiss this draft as simply ignoring existing work
(RFC 1830), and tackling a marginal problem (bandwidth saving for large
messages).
But I wonder if there isn't a real problem out there in parts of the world
where Internet bandwidth is considerably less plentiful and reliable than
most of us have come to expect. Mobile devices would be another case in point.
I think the interesting nugget in this draft is the idea that an SMTP
client may request retransmission of a partially transmitted message in a
new SMTP session, and if the server is prepared to continue a previous
transaction it provides information about the point from which
retransmission should continue.
The final point I note is that an extension along these lines does not need
to be widely deployed to be useful. Logically, it would be used between
servers in areas of poor quality bandwidth, with messages routed
appropriately using MX records.
#g
------------
Graham Klyne
(GK(_at_)ACM(_dot_)ORG)