[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: What about an authenticated TURN command?



To start with the basic question of whether it's okay for a mail server
to have a dynamic address, I don't think it matters what we think - the
reality is that ISPs providing mail service to small businesses are
already operating in that way.

As to the issues you raise:

1)  I don't think this would necessarily put the wait behind SASL/TLS
support.  We could write the draft to just refer to generalized
authentication rather than SASL or TLS, removing the dependency on those
specific drafts.

2)  This is a good point.  The reality, however, is that a system which
can't start dequeuing fairly soon after an ATRN just isn't going to cut
it in this problem space - it _has_ to be able to dequeue while the
client is connected.  However there will occasionally be times when the
server can't respond soon, and we should deal with this.  One solution
might be to have the server return a response indicating whether (a) it
has mail to dequeue and (b) it can start the dequeue in a "reasonable"
period of time.  The client could use this response to decide whether to
keep the connection open or hang up and try again later.

-- jeff

> -----Original Message-----
> From:	Jack De Winter [SMTP:jack@xxxxxxxxxxxxxx]
> Sent:	Monday, June 09, 1997 10:19 AM
> To:	Jeff Stephenson (Exchange); 'ietf-disconn-smtp@xxxxxxx'
> Subject:	Re: What about an authenticated TURN command?
> 
  ...<snip>...
>  
> Sounds like a good approach.  However:
> 
> 1) This would put the bulk of the waiting behind the
> SMTP-SASL/SMTP-TLS
> support.  I.e. that would have to be finalized before we could get our
> draft finalized.
> 
> 2) Some systems do not have the option of sending the mail right away
> over the same connection.  We would have to further design an approach
> where we could handle those.
> 
> Basically, ETRN was put in there to kick the other queue.  I agree
> that
> dynamic IP addresses put a kick into the entire thing.  
> 
> I guess a basic question to start would be: Is a dynamic address
> something
> that is okay for a mail server to have?
> 
> regards,
> Jack
> 
> -------------------------------------------------
> Jack De Winter - Wildbear Consulting, Inc.
> (519) 884-4498		http://www.wildbear.on.ca/
> 
> Author of SLMail for 95 & NT (http://www.seattlelab.com/)