But you do that only after confirming the entire class C belongs to the same organization, right?

nope.


We have a small subset of a class C (.240-.255) I'd be grumpy if my e-mail got thrown away

I don't throw away mail, I reject it.


because a mail admin once upon a time


got spammed by a different sub-set of the same class C

... is a very good predictor of the future behavior


, and decided that everyone there must be a bad guy..

wrong. We KNOW we got abused by some IPs in a Class C. We don't know (or care) about the other IPs in that ClassC, but the operator of the Class C has not policed his clients, and now he pays. We fixed our problem by passing it back to the Class C owner.


If you moved into a bad neighborhood, that's your problem, not mine. There are 16,777,216 Class Cs to be un/lucky with (yeah, I know all aren't available).

In practice, the policy is not really the problem you imagine it to be. It's accurate and effective.

Automatic blocking (by IMgate advanced) requires:

1) have no PTR, (aka "Welcome to My Hell"), and

2) send some volume of msgs to unknown users from one IP and/or from a number of IPs in the Class C.

For those of you running low-volume servers, the nature of spam today, seen at high-volumes MXs, is predominantly huge volumes of mail sent to unknown users (and often from forged senders), "shot-gunning", a behavior that is a very accurate, reliable indication of abuse (no need to scrutinize the msg contents). If 5 or 10 PTR-less IPs in a ClassC are sending volumes of msg to unknown users, boom!, that's one more spewing ClassC down the tubes.

Some hard numbers, rejects since midnight today, one Florida ISP:

   6985 ACL mta_clients_dict     <<< IPs and ClassC auto-blocked for doing ...
   7410 SMTP Exceeded Hard Error Limit after RCPT
   8217 RBL bl.spamcop.net
   8450 SMTP Exceeded Hard Error Limit after DATA
  14208 ACL to_relay_recipients unknown recipient  <<<  .... this
============================
  67236 TOTAL

The _dict filter (dictionary) which runs the AFTER known_recipients filter, meaning these are _dict-rejected msgs to our KNOWN users, and would have been accepted had we not harvested the PTR-less IPs into the _dict filter.

Above is to be compared with a total of 12K msgs accepted for the same period (but that's in and out, while the rejects are for inbound only). yep, it's bad:

Grand Totals
------------
messages

  12235   delivered
      7   forwarded
     87   deferred  (231  deferrals)
    114   bounced
  54574   rejected (81%)  <<<<<

If one looks at the reject reports, many of the PTR-less auto-blocked addresses now have PTRs, but they are still spammers:

    Client host rejected: ACL mta_clients_dict_classc (total: 4244)
         165   kdlaj9023jkla.com
         154   primaryoffers.com
         130   erlaok.com
         119   egoldsavings.net
          96   moneyholdem.com
          90   rbo01.com
          63   mobd01.com
          62   qualitypro.net
          49   64.191.76.12
          48   64.70.17.67
          48   64.70.17.75
          48   64.70.17.77
          47   218.80.65.68
          46   gof01.com
          44   64.70.17.71
          43   218.80.65.169
          41   64.70.17.74
          41   picklepatches.com
          40   greatwebads.com
          37   beeperjack.com
          35   savingsnotice.com
          34   203.208.248.223
          34   203.208.248.239
          33   64.191.76.11
          33   203.208.248.207
          33   211.144.32.175

    Client host rejected: ACL mta_clients_dict_ip (total: 3117)
         299   69.6.51.5
         162   211.154.103.13
         151   69.6.51.4
          91   69.6.51.2
          90   yourbigvote.com
          69   69.6.51.3
          67   emsemail.biz
          57   64.239.182.80
          45   gof01.com
          44   206.112.88.231
          39   dailyripple.com
          34   203.208.248.31
          33   219.238.200.68
          30   66.163.228.5
          30   218.107.189.167
          25   69.10.154.202
          25   ediets.com
          24   64.211.50.33
          24   205.138.96.49
          24   206.112.88.235
          23   12.32.40.151
          23   200.30.166.28
          23   205.138.96.45
          20   4.23.173.56
          20   jupiterdiscount.com

But I said these _dict filters are for PTR-less abusers?? How come some above have PTR hostnames?

The PTR addresses spammed our unknown users BEFORE the IPs got a PTR. But they remain convicted by their previous, PTR-less behavior of sending lots of msgs to unknown users.

At this ISP, the _dict filter has 15203 lines in it (some blocks are A.B.C, some are A.B.C.D), and is updated several times/day based on today's PTR-less spammers.

Len

_____________________________________________________________________
http://MenAndMice.com/DNS-training : London; San Jose; Orlando; Chicago
http://IMGate.MEIway.com : free anti-spam gateway, runs on 1000's of sites


To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/ Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/

Reply via email to