[Mimedefang] Reputation Reporting Protocol submitted to IETF as an I-D

David F. Skoll dfs at roaringpenguin.com
Sat Jun 19 14:37:42 EDT 2010

Andrzej Adam Filip wrote:

> IMHO you should "generalize" support for different signature types
> e.g. 1 extra byte for signature length and 1 extra byte for signature type

I'm not sure what you mean by "signature types".  Could you explain?
Are you referring to the truncated HMAC?

By the way, your outbound SMTP server has a rather
poor reputation.  We've been running the Reputation Reporting Protocol
with several hundred sensors for a few months now and we keep a window
of 45 days' worth of events (about 1.6 x 10^9 events in total;
~400/second).  Here's the score for

$ canit-reputation-check smtpout7.poczta.onet.pl
        gl=142 ug=33 hs=21 hh=2 as=15177 ah=87 vr=17209 ir=13022 activity=6.1

That means 142 greylisting events, 33 ungreylisting events, 21 messages
hand-voted as spam, 2 hand-voted as non-spam, 15177 auto-detected as spam,
87 auto-detected as non-spam, 17209 valid RCPT commands and 13022
invalid RCPT commands.  activity=6.1 is a log-scale measure of how much
activity our sensors have picked up; is a fairly active
SMTP client.

You need to get your ISP to clean up its act. :-)



More information about the MIMEDefang mailing list