Body Checksum Calculation

Vernon Schryver vjs@calcite.rhyolite.com
Tue Apr 2 00:03:09 UTC 2002


> From: "Tony L. Svanstrom" <tony@svanstrom.org>

> ...
>  The problem is that if you're trying to apply DCC to a situation where the
> end-user isn't aware of what is being used and why, then the crucial white-
> list part isn't available; and based on the fact that the average user is an
>idiot it can't easily be added (not unless you don't mind long support-sessions
> via e-mail every time a user suspects he's not getting enough e-mails, and
> every time a user isn't getting enough e-mails).

In such a situation, I would make the use of DCC filter optional for each
user, and discourage the idiots from turning it on by making the knob
obscure.  Still, the checksums of the idiots' mail should be reported.

Note yet again that DCC filtering is entirely independent of DCC checksum
reporting.  For example, if you had 10,000 users but only 1 using DCC
filtering, you would want to arrange checksums for all mail to all 10,000
users to be reported, so that the counts seen by the 1 user of DCC
filtering would be more accurate.


>  In a situation like that you can't add all checksums to the server from all
> the clients (ie users accounts); 

I don't see why not.

>                                  and right now I don't see a solution to this
> problem that doesn't exclude DCC nor require all e-mails to be passed via a
> single system.

Exactly what is the problem in reporting checksums to a DCC server
and not doing any filtering?  The network costs are likely to be much
smaller than the network costs of the DNS transactions usually required
for every incoming mail message.  The CPU costs of 3 MD5 checksums of
the entire body and MD5 checksums of several headers might be a worry
on CPUs that are slow compared to the bit rate of incoming mail, but
if so, the DCC is probably the least of your problems.


>  I think I'm trying to squeeze a solution out of DCC that simply isn't there;
> but I'm still working on it. :-)

I continue to suspect you imagine that the DCC is something other than it is.

The DCC would be typical, ill considered, completely unthought-out,
useless, anti-spam wishful thinking junk if it required all users of
an MTA to have their mail filtered.  A primary design criterion of
any real spam filtering system is to easily exempt some addresses,
starting with postmaster and certainly abuse.


Vernon Schryver    vjs@rhyolite.com



More information about the DCC mailing list

Contact vjs@rhyolite.com by mail or use the form.