RE: [lemonade] some questions in draft-ietf-lemonade-notifications-03

[Randall Gellens <randy at qualcomm.com>]

At 9:36 AM +0100 8/9/06, Ben Last wrote:

>  Gang Liang [mailto:lianggang79 at huawei.com] wrote on 09 August 2006
>  04:04:
> > Using MMS to push email or notifications is just an example. What I
>  want to
> > talk about is the feasibility and rationality of pushing an email via
> > outband notification. E.g. if the client is GPRS online and the email
>  is
> > very urgent, maybe it's more appropriate to push the email directly in
>  the
> > notification instead of pushing a wake-up notification that makes the
>  client
> > to create a connection with server to fetch the email.
>
>  I think, then, that one of the key factors is the amount of data you can
>  get into an outband notification.

There's also the security implications: with a simple notification 
that an event has occurred, the payload generally needs to be signed 
but not encrypted.  When actual data (such as subject, sender, or 
body of the mail) is included, it needs to be encrypted as well.

>   You can fit 140 bytes into an SMS, so
>  there's not much room for more than a notification that events have
>  occurred.  You can fit far more into an MMS (with all the cost
>  implications)

With MMS, generally a notification is sent to the client, with advice 
from the server as to if the client should immediately retrieve the 
message or defer it.  So to send notification of new mail over MMS 
would be convoluted, since:
	(1) new mail arrives at message store
	(2) notification sent as MMS to client
	(3) MMS arrives at server
	(4) MMS server sends notification to client
	(5) client brings up traffic channel and fetches MMS
	(6) MMS client hands MMS to mail client
	(7) mail client brings up traffic channel and fetches mail

Whew!

>  However, I think that the decisions on (a) when to retrieve mail and (b)
>  how much to retrieve are better made by the client.  For example, if the
>  client can tell that the email is low-priority, or that it is in an area
>  where network access is slow and expensive, it may choose to retrieve
>  messages only once per hour, and to retrieve only the headers.  Or, if
>  it detects that it is in range of a usable wi-fi connection, and thus
>  has high-speed and low-cost access, it might choose to retrieve the
>  header and message text of new emails.  In the mobile phone environment,
>  I think there are many factors, which vary with time and location and
>  that are detectable only by the client, which might affect the client's
>  decisions about how to respond to the information that new messages (or
>  changes to messages) are available.

I'd agree with this.  We should keep the client in control as much as possible.

--
Randall Gellens
Opinions are personal;    facts are suspect;    I speak for myself only
-------------- Randomly-selected tag: ---------------
Every word is like an unnecessary stain on silence and nothingness.
                                                        --Beckett

_______________________________________________
lemonade mailing list
lemonade at ietf.org
https://www1.ietf.org/mailman/listinfo/lemonade