Re: Internet Draft for flexible proxying of the mail protocols

Philip Hazel <> Fri, 29 September 2000 08:21 UTC

Received: from (CS.UTK.EDU []) by (8.9.1a/8.9.1a) with SMTP id EAA24961 for <>; Fri, 29 Sep 2000 04:21:57 -0400 (EDT)
Received: from localhost (daemon@localhost) by with SMTP (cf v2.9s-UTK) id EAA00019; Fri, 29 Sep 2000 04:21:42 -0400 (EDT)
Received: by (bulk_mailer v1.13); Fri, 29 Sep 2000 04:21:40 -0400
Received: by (cf v2.9s-UTK) id EAA29998; Fri, 29 Sep 2000 04:21:39 -0400 (EDT)
Received: from (marvin@localhost) by with ESMTP (cf v2.9s-UTK) id EAA29984; Fri, 29 Sep 2000 04:21:37 -0400 (EDT)
Received: from ( -> by (smtpshim v1.0); Fri, 29 Sep 2000 04:21:37 -0400
Received: from ph10 (helo=localhost) by with local-esmtp (Exim 3.16 #3) id 13evQD-0005bs-00; Fri, 29 Sep 2000 09:21:33 +0100
Date: Fri, 29 Sep 2000 09:21:33 +0100
From: Philip Hazel <>
To: Russ Allbery <>
Subject: Re: Internet Draft for flexible proxying of the mail protocols
In-Reply-To: <>
Message-ID: <>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset="US-ASCII"
List-Unsubscribe: <>

On 28 Sep 2000, Russ Allbery wrote:

> We much prefer POP from the server side.  

We don't, but then we have permanently connected clients that keep 
probing for new mail at regular intervals. POP requires a password 
authentication for each probe, and a download of the entire mailbox to 
check for new messages. When the users set their probe interval 
ridiculously short (e.g. 1 minute), and leave several megabytes in their 
mailboxes, this hits the machine badly. We make a continuous effort to
educate the users to probe no more than every 5 minutes, and not to set
"leave mail on server", but people being what they are, this is a battle
that can never be entirely won.

IMAP, in contrast, because it keeps a session going throughout, and 
doesn't need to download the entire mailbox to find the list of 
messages, is much lighter on the server.

> Of course, the users understandably want the features of IMAP so we're
> getting pushed in that direction, but there will definitely be a price
> corresponding to much higher server resources and administrative time
> needed over POP.

So I would argue that that isn't always the case (at least for server 
resources). I realise, of course, that in a different scenario, the 
trade-offs are different.

Philip Hazel            University of Cambridge Computing Service,      Cambridge, England. Phone: +44 1223 334714.