All mail that passes through the main Humboldt Netlink mail server (mailer.humboldt.net) is being scanned to identify Spam. We use a rule based spam
filter (SpamAssassin) that marks
suspect messages with a score determined by what rules were matched. A commercial
virus filter scans any attachments and will cause an infected attachment to be
removed from the message before further delivery. The scanner also quarantines
attachments that that have been deemed unsafe by Microsoft. You can see a
summary of our filter system and what the SpamAssassin rules are at Mail Filter Rules.
Our mail filter uses very conservative settings with respect to spam. At the
server level the filter does discard messages that garner a score equal to or
greater than the SpamAssassin High Score, but
all messages with a lower score are delivered. This is done to reduce the
possibility of the anti-spam filter blocking legitimate email. Users can tailor
what actions are taken on messages that have been identified as possible spam at
our Mail Account Management page. With that application
you can set your own spam discard score as well as establishing your own personal
whitelist. The application also allows the user to forward email and any email so
forwarded will be subject to the user's whitelist and spam control. Note that
messages that the mail filter scans that garner a SpamAssassin score less than
3 will not have X-MailScanner-SpamCheck &
X-MailScanner-SpamScore headers added to the message.
A Spam Primer
Spam, or Unsolicited Commercial Email (UCE), is a growing problem. It corresponds
exactly to the junk mail that everyone receives on a daily basis via the postal
system. However, while it costs a stamp to deliver surface mail there's little
cost to deliver UCE via the Internet. The bulk of UCE being spread now comes from
"one time sources", where some clueless user has fallen for one of the
"make money with your computer at home" scams. They get a CD with bulk
mailing software and huge address lists and are paid some amount for each UCE that
they send. At reputable ISP's such activity doesn't go unnoticed for long and the
account is suspended or canceled, causing that particular source to disappear. So
blocking the IP or site is of little or no use. Unfortunately there are a lot of
service providers that are more than a bit tolerant of spamming activities by
their customers or at least aren't very diligent in monitoring and acting on
spammers. Other sources of UCE are caused by poorly configured mail servers that
will relay mail for any user. Professional spammers search for mail servers like
this and use those to disseminate UCE. In cases like that the mail may appear to
originate from a reputable source.
Spammers also take pains to hide their real identity, see the section on Email
Basics below for an example. And since the headers in most spam are forged it
serves no purpose to reply to a message to complain. If there is a real address
associated with the message that you can reply to it is probably directed to a
program that marks your address as being valid in the UCE address list. In a like
manner, the "Click here to be removed from this list" is probably a
scam. taking that action doesn't get you removed from the list, it merely confirms
that your address works. The only sane thing to do with a piece of UCE is exactly
the same thing that you do at home with postal junk mail, just delete it.
So how did your email address get onto one of these UCE mailing lists?
Unfortunately there are a lot of ways for that to happen. Email addresses have
become a commercial item, with organizations that support UCE activity actively
pursuing email lists and paying Internet sites for collected addresses. Those
sites that are above board about the use of email addresses will give you the
option of restricting the use of your address and will use a closed loop
verification, meaning that you'll receive an email requiring you to go back to
their site via a special link to confirm your choice. If you don't verify the
selection they discard the address. These are commonly referred to as
"opt-in" sites. Since an email address has commercial value, not all
sites operate this way. Some consider any addresses submitted as being implicitly
available for their use, others operate on an "opt-out" policy where you
have to explicitly request that your email address not be disseminated. And a lot
of sites don't confirm email addresses that a site visitor supplies. Even when the
site doesn't have a policy of selling or otherwise distributing email addresses
that they have gathered if they don't properly protect that data it might be
stolen and sold by anyone that gains access to the data.
While not as many people are exposed this way, as compared to several years ago,
participation in any of the Usenet news groups can result in your email address
being exposed to harvesters. Another common means of exposure is subscription to
mailing lists, chat groups, and signing of web site guest books. There are
harvesters that subscribe to mailing lists just to collect email addresses, others
that search chat groups and others that scan the Internet looking for web site
guest books. Some viruses and malicious web sites also gather email addresses.
Obviously good anti-virus protection on your mail client machine is important, but
what's not so obvious is that there are ways to trick Outlook and Internet
Explorer into divulging an email address. For the most part that does require that
you be running a susceptible version of those tools and/or that you haven't
configured the security settings high enough. And then there are the applications
that ask you to "register with them". Before doing so you really need to
determine from their site what their policies are with regard to collected email
addresses.
Email Basics
Every email message consists of two parts. The part that you see in your email
client is normally referred to as the body and contains the
From:, To:, Subject:,
etc.,headers, the actual message and any attachments. Note that nothing in the body
ofa message is actually used by a Mail Transport Agent (MTA) to effect delivery of a
message. The headers that you can normally see are informational in nature and can
contain anything that the sending Mail User Agent (MUA) wants to place there. In
fact, those headers can be completely missing and the email will still be
delivered. There's another set of headers, normally referred to as the envelope,
that actually govern email delivery. The body and envelope of a message correspond
exactly to a letter content and envelope of surface mail. The Post Office uses
what's on an envelope to deliver mail and has no knowledge of what's inside the
envelope. Spammers frequently forge the From: and/or
To: contents to hide their identity, which is how you can get an
email that appears, from the headers you normally see, to be from a legitimate
source or to have been sent to an address that isn't yours. To determine the
actual source of a message you have to examine the envelope headers. With a
Netscape email client this is easily done by selecting
View->Headers->All and you'd see
something like:
Return-Path:<angelqq94264e34@aol.com>
X-Sieve: cmu-sieve 2.0
Received: from relay.dom.tld (relay.dom.tld [1.2.3.4])
by mail.dom.tld (8.12.3/8.12.3) with SMTP id g7SJJvju027274
for <Jim.Levie@mail.dom.tld>; Wed, 25 Sep 2002 06:16:33 -0500
Received: from aol.com (200-207-126-141.dsl.telesp.net.br [200.207.126.141]))
by relay.dom.tld (8.12.2/8.12.2) with ESMTP id g8PBGLSu016440
for <jim.levie@dom.tld>; Wed, 25 Sep 2002 06:16:28 -0500
Reply-To: <angelqq94264e34@aol.com>
Message-ID: <004c41d82d1e$1548a3b3$8cb02bd4@ynjpcq>
From: <juan63201h06@hotmail.com>
To: angelqq9@aol.com
Subject: Hows it going 5200dujH-8
Date: Wed, 26 Sep 2002 03:05:41 +0800
MiME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0)
Importance: Normal
X-Scanned: Found to be clean, Found to be clean
X-MailScanner: Found to be clean
X-MailScanner-SpamCheck: (score=13.1, required 3 DATE_IN_PAST_12_24,
FROM_ENDS_IN_NUMS, FROM_HAS_MIXED_NUMS, FROM_HAS_MIXED_NUMS2,
INVALID_MSGID, MIME_ODD_CASE, NO_REAL_NAME, SPAM_PHRASE_02_03,
SUPERLONG_LINE, USER_AGENT_OUTLOOK)
X-MailScanner-SpamScore: sssssssssssss
Examining the first Received: line from the bottom we can tell
that the message originated from
200-207-126-141.dsl.telesp.net.br and that system was claiming
to be aol.com neither of which have anything to do with what we
see in the From: header
(juan63201h06@hotmail.com). Also note that the
To: header says that the message is for
angelqq9@aol.com but that the envelope recipient (from the
topmost Received: line is Jim.Levie@dom.tld.
Our email scanning system added two headers to the message. One
(X-MailScanner: Found to be clean) indicates that the message
has been scanned and was not carrying a virus. The other (X-MailScanner-SpamCheck:
(score=13.1..) is more interesting. It says that the scanning
system ran the message through SpamAssassin which classified it as possible
spam. The X-MailScanner-SpamScore: sss... header line is a simple
representation of the rounded SpamAssassin score of one "s" for each integer of the score (up to a max of 20). The
SpamAssassin tests that matched on this message are:
| DATE_IN_PAST_12_24 |
Date: is 12 to 24 hours after Received: date (score=3.169) |
| FROM_ENDS_IN_NUMS |
From: ends in numbers (score=1.614) |
| FROM_HAS_MIXED_NUMS |
From: contains numbers mixed in with letters (score=-0.891) |
| FROM_HAS_MIXED_NUMS2 |
From address matches known spammer format (score=1.862) |
| INVALID_MSGID |
Message-Id is not valid, according to RFC 2822 (score=1.226) |
| MIME_ODD_CASE |
MiME-Version header (oddly capitalized) (score=3.478) |
| NO_REAL_NAME |
From: does not include a real name (score=-0.331) |
| SPAM_PHRASE_02_03 |
Spam phrases score is 02 to 03 (medium) (score=-0.713) |
| SUPERLONG_LINE |
Contains a line >=199 characters long (score=-2.197) |
| USER_AGENT_OUTLOOK |
X-Mailer header indicates a spam MUA (Outlook) (score=3.151) |
While some of the things that SpamAssassin matched might have been associated with a
legitimate message, the combination of those makes it likely that this message was
from a spammer.
|