Dealing with Unsolicited Bulk e-Mail

PMDF e-Mail Interconnect provides a wealth of proven tools for dealing with unsolicited bulk e-mail, commonly referred to as "spam". These tools, described in more detail below, include:
  • Access controls: reject mail from known spam sources as well as control who within your organization can send or receive e-mail.
  • Mailbox & system wide filtering: individual users, through PMDF's web interface, can manage their own spam filters, taking control over what mail is delivered to their mailbox.
  • Address verification: refuse mail with invalid originator addresses.
  • Realtime Blackhole List: refuse mail from recognized spam sources as identified by the Mail Abuse Protection System's Realtime Blackhole List (MAPS RBL). The MAPS RBL is a responsibly managed, dynamically updated list of known spam sources.
  • Relay blocking: prevent spammers from using your mail system as a mail relay to send their unwanted spam to tens of thousands of recipients.
  • Authentication services: enable password authentication in your SMTP server via the Simple Authentication & Security Layer protocol (SASL).
  • Sidelining: silently sideline or even delete potential spam messages.
  • Comprehensive tracing: reliable mechanisms for identifying a message's source.

These tools may be used individually or in concert. They are designed to be as efficient as possible so as to minimize your resource cost in dealing with the abuse presented by spam.

Concepts

Before continuing our discussion of PMDF's spam fighting facilities, there's a basic yet important concept in PMDF which needs to be brought to light. Mail traffic passing into, through, and out of PMDF can be separated into distinct channels based upon various criteria. Important to our discussion is that source and destination e-mail address as well source IP address or subnet are among these criteria. Different processing characteristics can then be ascribed to these different mail flows, referred to hereafter as channels. Consequently, different access controls, mail filters, processing priorities, and tools may be applied in different ways and combinations to these different channels. For example, you can process mail originating from within your domain differently than mail originating from the outside world.

In addition to channel-based message flow classification, another useful classification is mailing list traffic. Traffic for a given mailing list may come into PMDF via a number of different channels and go back out via a number of different channels. As such, when dealing with mailing lists, it is useful to think in terms of the list itself and not in terms of channels. PMDF recognizes this and allows many of the channel-specific spam fighting tools to also be applied in a mailing-list specific fashion. For further details, see the description of mailing lists in the PMDF System Manager's Guide.

  Access controls

PMDF has a general purpose mechanism which may be used to reject mail based upon a variety of criteria including the message's source or destination e-mail addresses as well as source IP address. This mechanism may be used in a variety of ways to combat spam. For example, refusing mail from specific senders or entire domains (e.g., mail from friend@public.com). This mechanism can be extended with a database should you have large lists of screening information. This same mechanism is also suitable for maintaining a database of internal users who are or are note allowed to send mail out certain channels; e.g., restrict on a per-user basis who can or cannot send or receive Internet mail.

For further details, please see the description of the SEND_ACCESS, ORIG_SEND_ACCESS, MAIL_ACCESS, and ORIG_MAIL_ACCESS mapping tables in the PMDF System Managers Guide.

  Mailbox Filtering

PMDF provides per-user, per-channel, and system-wide mail filters. These filters can be managed through web pages or written with a text editor. Using these filters, users may control what mail messages are or are not permitted to be delivered to their mailbox. For instance, a user tired of "make money fast" spam can specify that any messages with such a subject be rejected. Similar controls may be applied by the postmaster to all messages flowing through a channel or all of PMDF.

The mail filtering in PMDF is based upon the new Sieve filtering language being developed by members of the Internet Engineering Task Force (IETF).

See the PMDF System Manager's Guide for further information on PMDF's mailbox filters.

  Address Verification

Spam messages often use invalid originator addresses. The PMDF SMTP server can take advantage of this by rejecting messages with invalid originator addresses. If the originator address does not correspond to a valid host name, as determined by querying the Internet Domain Name System (DNS), the message can be rejected.

For further details, refer to the description of the mailfromdnsverify channel keyword in the PMDF System Manager's Guide.

  Realtime Blackhole List

The Mail Abuse Protection System's Realtime Blackhole List (MAPS RBL) is a dynamically updated list of known spam sources identified by source IP address. The PMDF SMTP server supports use of the MAPS RBL and can reject mail coming from sources identified by the MAPS RBL as originators of spam. The MAPS RBL is a free service provided via the Internet DNS.

For further information on using the MAPS RBL with PMDF, please see the description of the ENABLE_RBL option of the PMDF Dispatcher.

  Relay Blocking

The features discussed so far have focused on preventing your own users from receiving spam. However, a comprehensive spam prevention strategy should also include tactics for preventing spammers from relaying mail through your system to other systems. [1] PMDF has facilities to prevent unauthorized use of your system for mail relaying. In its simplest form, this is achieved by allowing local users and systems to relay mail while rejecting relay attempts from non-local systems. This differentiation between local vs. non-local is easily and securely made by using IP addresses as the differentiator. It's easy in that it's simple to determine if an IP address is part of your IP network. It's secure in that IP addresses are not readily forged: the IP router between your site and the Internet rejects incoming data packets which purport to have as their source IP address an address within your network.

Please see Preventing Mail Relaying with PMDF for futher details. Note that PMDF configurations generated with PMDF V5.2 or later configuration tools include relay blocking.

  Authentication services

The PMDF SMTP server implements the Simple Authentication and Security Layer protocol (SASL, RFC 2222). This may be used with popular POP and IMAP clients to provide password-based access to your SMTP server. A typical usage for SASL is to permit mail relaying for externa,l authenticated users. This solves the common problem posed by local users who use Internet Service Providers (ISPs) from home or while travelling.[2] Such users, when connecting to your mail system, will have non-local IP addresses. Any relay blocking which only takes into account source IP address will not allow these users to relay mail. This difficulty is overcome through the use of SASL which allows these users to authenticate themselves. Once authenticated, they may be permitted to relay mail.

PMDF's authentication services allow for a number of different password sources and may be extended to use site-defined password sources. Please see the PMDF System Manager's Guide for further details.

  Sidelining

The access control mechanisms cited previously may also be used to defer the processing of suspect messages for later, manual inspection. Or, rather than sideline, the destination address may be changed, thus routing the suspect mail to a specific mailbox or even the bitbucket channel. This tatic is useful when spam is being received from a known, fixed origin and outright rejection will only cause the spammer to change their point of origin. Similar features are available for PMDF mailing lists.

For futher details, see the PMDF System Manager's Guide. Specifically, the discussion of the holdlimit channel keyword as well as the $A, $B, $D, and $H flags of the SEND_ACCESS, ORIG_SEND_ACCESS, MAIL_ACCESS, and ORIG_SEND_ACCESS mapping tables.

  Comprehensive tracing

PMDF's SMTP server discovers and records crucial origination information about every incoming mail message. For example, source IP address and the corresponding host name. In addition, use of the Identification Protocol (RFC 1413), is also supported. All discovered information is recorded in the message's trace fields (e.g., the Received: header line) as well as in log files. Availability of such reliable information is crucial in determining the source of spam which often has forged headers.

Other details

Associated with many of these features, are a myriad of fine details. For instance, site-supplied screening code and databases may be interfaced directly to PMDF; third-party spam and virus scanners may be invoked; format conversions enabled; etc. Particularly annoying to spammers is that rejection responses from the SMTP server may intentionally be delayed by, say, 15 seconds.

End notes

[1] Spammers who send out tens of thousands of mail messages do so by sending relatively few messages to unsuspecting, store-and-forward mail systems. Each of these few messages may have hundreds or thousands of recipient addresses. The "attacked" store-and-forward mail system in receipt of one of the messages is then expected (by the spammer) to relay (resend) the message to each of its recipients -- recipients who are generally in many different domains unrelated to that of the attacked mail system. Thus, the attacked mail system is used as a mail relay and the burden of contacting thousands of individual recipient mail systems is foisted by the spammer onto it. Moreover, the individual recipients are often deceived into thinking that the attacked system is responsible for the spam.

[2] POP and IMAP clients send e-mail by relaying it through a server system. This way, should the ultimate destination machine not be immediately reachable, the burden of holding onto the e-mail and periodically attempting to deliver it is put upon a presumably more capable and robust MTA rather than the client itself which may be nothing more than a laptop or pda device.

Search: