Shapira, Noam (Shapira_Noam@icomverse.com)
Fri, 3 Aug 2001 01:59:38 +0300
Hi all,
I would like to address some of the comment passed to me on the issues of
the SNAP draft and hope to get comments / advise.
The SNAP proposal is a proposal for a protocol between the ES and
centralistic notification service. This protocol is supposed to support very
large scale system . This are the some of the basic considerations taken in
the proposal. Following are the issues that raise repeatedly:
1. The SNAP is supposed to be based on pre-configured (provisioned) set of
users. This is a basic concept that is different from a subscription model,
where the end user / agent subscribes for the notifications. As far as I see
it - a non-subscription model is less expensive and is mandatory for large
scale systems.
2. The SNAP is "push oriented" - rather then "pull oriented" notification
protocol. Meaning that the notification are pushed from the ES to the
notification service that is responsible for sending it to the end user /
end user agent (for example to SMS). Again, the main considerations that
came in to mind, in this case, where the issue of very large scale systems.
Pulling of thousands of subscribers on an ES - all the time is a performance
hazarded.
3. Another issue about the SNAP is that the SNAP proposal sends a great deal
of information to the notification service (email headers and other parts of
the deposit /retrieving session). This is different from an approach that
says to send only the notification state to the end user - and make sure
that the end user / end agent enter the ES (or any other notification
source) and get the information. The reason for sending a great deal of
information is that - (1) let the end user get as much information that he
needs before entering the ES for retrieving. (2) Let the notification
service - the centralistic service hold a decision mechanism that can be
helped by the information passed in the notification request.
I would be most happy to get your comment on the questioned issues raised in
this email. It will be great if you know some references (other then the
ones sent in in the draft) that discussed these issues before and send it
to me.
Thanks,
Noam Shapira
P.S.
You can find the draft at
http://search.ietf.org/internet-drafts/draft-shapira-snap-01.txt.
A mailing list was opened - snap@lists.neystadt.org for this issue - you can
register by
either do it through Web interface:
http://www.neystadt.org/cgi-bin/majordomo
or send mail with 'subscribe snap' to majordomo@@lists.neystadt.org
This archive was generated by hypermail 2.0b3 on Fri Aug 03 2001 - 02:00:40 IDT