ARMWARE RFC Archive <- RFC Index (2601..2700)

RFC 2635

(also FYI 35)


Network Working Group                                    S. Hambridge
Request for Comments: 2635                                      INTEL
FYI: 35                                                      A. Lunde
Category: Informational                       Northwestern University
                                                            June 1999

                               DON'T SPEW
                A Set of Guidelines for Mass Unsolicited
                     Mailings and Postings (spam*)

Status of this Memo

   This memo provides information for the Internet community.  It does
   not specify an Internet standard of any kind.  Distribution of this
   memo is unlimited.

Copyright Notice

   Copyright (C) The Internet Society (1999).  All Rights Reserved.

Abstract

   This document explains why mass unsolicited electronic mail messages
   are harmful in the Internetworking community.  It gives a set of
   guidelines for dealing with unsolicited mail for users, for system
   administrators, news administrators, and mailing list managers.  It
   also makes suggestions Internet Service Providers might follow.

1.  Introduction

   The Internet's origins in the Research and Education communities
   played an important role in the foundation and formation of Internet
   culture.  This culture defined rules for network etiquette
   (netiquette) and communication based on the Internet's being
   relatively off-limits to commercial enterprise.

   This all changed when U.S. Government was no longer the primary
   funding body for the U.S. Internet, when the Internet truly went
   global, and when all commercial enterprises were allowed to join what
   had been strictly research networks.  Internet culture had become
   deeply embedded in the protocols the network used.  Although the
   social context has changed, the technical limits of the Internet
   protocols still require a person to enforce certain limits on
   resource usage for the 'Net to function effectively.  Strong
   authentication was not built into the News and Mail protocols.  The
   only thing that is saving the Internet from congestion collapse is
   the voluntary inclusion of TCP backoff in almost all of the TCP/IP

Hambridge & Lunde            Informational                      [Page 1]



RFC 2635                       DON'T SPEW                      June 1999

   driver code on the Internet.  There is no end-to-end cost accounting
   and/or cost recovery.  Bandwidth is shared among all traffic without
   resource reservation (although this is changing).

   Unfortunately for all of us, the culture so carefully nurtured
   through the early years of the Internet was not fully transferred to
   all those new entities hooking into the bandwidth.  Many of those
   entities believe they have found a paradise of thousands of potential
   customers each of whom is desperate to learn about stunning new
   business opportunities.  Alternatively, some of the new netizens
   believe all people should at least hear about the one true religion
   or political party or process.  And some of them know that almost no
   one wants to hear their message but just can't resist how inexpensive
   the net can be to use.  While there may be thousands of folks
   desperate for any potential message, mass mailings or Netnews
   postings are not at all appropriate on the 'Net.

   This document explains why mass unsolicited email and Netnews posting
   (aka spam) is bad, what to do if you get it, what webmasters,
   postmasters, and news admins can do about it, and how an Internet
   Service Provider might respond to it.

2.  What is Spam*?

   The term "spam" as it is used to denote mass unsolicited mailings or
   netnews postings is derived from a Monty Python sketch set in a
   movie/tv studio cafeteria.  During that sketch, the word "spam" takes
   over each item offered on the menu until the entire dialogue consists
   of nothing but "spam spam spam spam spam spam and spam."  This so
   closely resembles what happens when mass unsolicited mail and posts
   take over mailing lists and netnews groups that the term has been
   pushed into common usage in the Internet community.

   When unsolicited mail is sent to a mailing list and/or news group it
   frequently generates more hate mail to the list or group or apparent
   sender by people who do not realize the true source of the message.
   If the mailing contains suggestions for removing your name from a
   mailing list, 10s to 100s of people will respond to the list with
   "remove" messages meant for the originator.  So, the original message
   (spam) creates more unwanted mail (spam spam spam spam), which
   generates more unwanted mail (spam spam spam spam spam spam and
   spam).  Similar occurrences are perpetrated in newsgroups, but this
   is held somewhat in check by "cancelbots" (programs which cancel
   postings) triggered by mass posting.  Recently, cancelbots have grown
   less in favor with those administering News servers since the
   cancelbots are now generating the same amount of traffic as spam.
   Even News admins are beginning to use filters, demonstrating that
   spam spam spam spam spam spam and spam is a monumental problem.

Hambridge & Lunde            Informational                      [Page 2]



RFC 2635                       DON'T SPEW                      June 1999

3.  Why Mass Mailing is Bad

   In the world of paper mail we're all used to receiving unsolicited
   circulars, advertisements, and catalogs.  Generally we don't object
   to this - we look at what we find of interest, and we discard/recycle
   the rest.  Why should receiving unsolicited email be any different?

   The answer is that the cost model is different.  In the paper world,
   the cost of mailing is borne by the sender.  The sender must pay for
   the privilege of creating the ad and the cost of mailing it to the
   recipient.  An average paper commercial mailing in the U.S.  ends up
   costing about $1.00 per addressee.  In the world of electronic
   communications, the recipient bears the majority of the cost.  Yes,
   the sender still has to compose the message and the sender has to pay
   for Internet connectivity.  However, the recipient ALSO has to pay
   for Internet connectivity and possibly also connect time charges and
   for disk space. For electronic mailings the recipient is expected to
   help share the cost of the mailing.  Bulk Internet mail from the U.S.
   ends up costing the sender only about 1/100th of a cent per address;
   or FOUR ORDERS of magnitude LESS than bulk paper mailings!

   Of course, this cost model is very popular with those looking for
   cheap methods to get their message out.  By the same token, it's very
   unpopular with people who have to pay for their messages just to find
   that their mailbox is full of junk mail.  Neither do they appreciate
   being forced to spend time learning how to filter out unwanted
   messages.  Consider this: if you had to pay for receiving paper mail
   would you pay for junk mail?

   Another consideration is that the increase in volume of spam will
   have an impact on the viability of electronic mail as a
   communications medium.  If, when you went to your postal mail box you
   found four crates of mail, would you be willing to search through the
   crates for the one or two pieces of mail which were not advertising?
   Spam has a tremendous potential to create this scenario in the
   electronic world.

   Frequently spammers indulge in unethical behavior such as using mail
   servers which allow mail to be relayed to send huge amounts of
   electronic solicitations.  Or they forge their headers to make it
   look as if the mail originates from a different domain.  These people
   don't care that they're intruding into a personal or business mailbox
   nor do they care that they are using other people's resources without
   compensating them.

   The huge cost difference has other bad effects.  Since even a very
   cheap paper mailing is going to cost tens of (U.S.) cents there is a
   real incentive to send only to those really likely to be interested.

Hambridge & Lunde            Informational                      [Page 3]



RFC 2635                       DON'T SPEW                      June 1999

   So paper bulk mailers frequently pay a premium to get high quality
   mailing lists, carefully prune out bad addresses and pay for services
   to update old addresses.  Bulk email is so cheap that hardly anyone
   sending it bothers to do any of this.  As a result, the chance that
   the receiver is actually interested in the mail is very, very, very
   low.

   As of the date of this document, it is a daily event on the Internet
   for a mail service to melt-down due to an overload of spam.  Every
   few months this happens to a large/major/regional/
   national/international service provider resulting in denial of or
   severe degradation of service to hundreds of thousands of users.
   Such service degradations usually prompt the providers to spend
   hundreds of thousands of dollars upgrading their mail service
   equipment just because of the volume of spam.  Service providers pass
   those costs on to customers.

   Doesn't the U.S. Constitution guarantee the ability to say whatever
   one likes?  First, the U.S. Constitution is law only in the U.S., and
   the Internet is global.  There are places your mail will reach where
   free speech is not a given.  Second, the U.S. Constitution does NOT
   guarantee one the right to say whatever one likes.  In general, the
   U.S. Constitution refers to political freedom of speech and not to
   commercial freedom of speech. Finally, and most importantly, the U.S.
   Constitution DOES NOT guarantee the right to seize the private
   property of others in order to broadcast your speech.  The Internet
   consists of a vast number of privately owned networks in voluntary
   cooperation.  There are laws which govern other areas of electronic
   communication, namely the "junk fax" laws.  Although these have yet
   to be applied to electronic mail they are still an example of the
   "curbing" of "free speech."  Free speech does not, in general,
   require other people to spend their money and resources to deliver or
   accept your message.

   Most responsible Internet citizens have come to regard unsolicited
   mail/posts as "theft of service".  Since the recipient must pay for
   the service and for the most part the mail/posts are advertisements
   of unsolicited "stuff" (products, services, information) those
   receiving it believe that the practice of making the recipient pay
   constitutes theft.

   The crux of sending large amounts of unsolicited mail and news is not
   a legal issue so much as an ethical one.  If you are tempted to send
   unsolicited "information" ask yourself these questions: "Whose
   resources is this using?"  "Did they consent in advance?"  "What
   would happen if everybody (or a very large number of people) did
   this?" "How would you feel if 90% of the mail you received was
   advertisements for stuff you didn't want?" "How would you feel if 95%

Hambridge & Lunde            Informational                      [Page 4]



RFC 2635                       DON'T SPEW                      June 1999

   of the mail you received was advertisements for stuff you didn't
   want?"  "How would you feel if 99% of the mail you received was
   advertisements for stuff you didn't want?"

   Although numbers on the volume and rate of increase of spam are not
   easy to find, seat-of-the-pants estimates from the people on spam
   discussion mailing lists [1] indicate that unsolicited mail/posts
   seems to be following the same path of exponential growth as the
   Internet as a whole [2].  This is NOT encouraging, as this kind of
   increase puts a strain on servers, connections, routers, and the
   bandwidth of the Internet as a whole.  On a per person basis,
   unsolicited mail is also on the increase, and individuals also have
   to bear the increasing cost of increasing numbers of unsolicited and
   unwanted mail.  People interested in hard numbers may want to point
   their web browsers to
   http://www.techweb.com/se/directlink.cgi?INW19980504S0003 where
   Internet Week reports what spam costs.

   Finally, sending large volumes of unsolicited email or posting
   voluminous numbers of Netnews postings is just plain rude.  Consider
   the following analogy: Suppose you discovered a large party going on
   in a house on your block.  Uninvited, you appear, then join each
   group in conversation, force your way in, SHOUT YOUR OPINION (with a
   megaphone) of whatever you happen to be thinking about at the time,
   drown out all other conversation, then scream "discrimination" when
   folks tell you you're being rude.

   To continue the party analogy, suppose instead of forcing your way
   into each group you stood on the outskirts a while and listened to
   the conversation.  Then you gradually began to add comments relevant
   to the discussion.  Then you began to tell people your opinion of the
   issues they were discussing; they would probably be less inclined to
   look badly on your intrusion.  Note that you are still intruding.
   And that it would still be considered rude to offer to sell products
   or services to the guests even if the products and services were
   relevant to the discussion.  You are in the wrong venue and you need
   to find the right one.

   Lots of spammers act as if their behavior can be forgiven by
   beginning their messages with an apology, or by personalizing their
   messages with the recipient's real name, or by using a number of
   ingratiating techniques.  But much like the techniques used by Uriah
   Heep in Dickens' _David Copperfield_, these usually have an effect
   opposite to the one intended.  Poor excuses ("It's not illegal,"
   "This will be the only message you receive," "This is an ad," "It's
   easy to REMOVE yourself from our list") are still excuses.  Moreover,
   they are likely to make the recipient MORE aggravated rather than

Hambridge & Lunde            Informational                      [Page 5]



RFC 2635                       DON'T SPEW                      June 1999

   less aggravated.

   In particular, there are two very severe problems with believing that
   a "remove" feature to stop future mail helps: (1) Careful tests have
   been done with sending remove requests for "virgin" email accounts
   (that have never been used anywhere else).  In over 80% of the cases,
   this resulted in a deluge of unsolicited email, although usually from
   other sources than the one the remove was sent to.  In other words,
   if you don't like unsolicited mail, you should think carefully before
   using a remove feature because the evidence is that it will result in
   more mail not less.  (2) Even if it did work, it would not stop lots
   of new unsolicited email every day from new businesses that hadn't
   mailed before.

4a. ACK!  I've Been Spammed - Now What?

   It's unpleasant to receive mail which you do not want.  It's even
   more unpleasant if you're paying for connect time to download it.
   And it's really unpleasant to receive mail on topics which you find
   offensive.  Now that you're good and mad, what's an appropriate
   response?

   First, you always have the option to delete it and get on with your
   life.  This is the easiest and safest response.  It does not
   guarantee you won't get more of the same in the future, but it does
   take care of the current problem.  Also, if you do not read your mail
   on a regular basis it is possible that your complaint is much too
   late to do any good.

   Second, consider strategies that take advantage of screening
   technology.  You might investigate technologies that allow you to
   filter unwanted mail before you see it.  Some software allows you to
   scan subject lines and delete unwanted messages before you download
   them.  Other programs can be configured to download portions of
   messages, check them to see if they are advertising (for example) and
   delete them before the whole message is downloaded.

   Also, your organization or your local Internet Service Provider may
   have the ability to block unwanted mail at their mail relay machines
   and thus spare you the hassle of dealing with it at all.  It is worth
   inquiring about this possibility if you are the victim of frequent
   spam.

   Your personal mailer software may allow you to write rules defining
   what you do and do not wish to read.  If so, write a rule which sends
   mail from the originator of the unwanted mail to the trash.  This
   will work if one sender or site repeatedly bothers you.  You may also
   consider writing other rules based on other headers if you are sure

Hambridge & Lunde            Informational                      [Page 6]



RFC 2635                       DON'T SPEW                      June 1999

   the probability of them being activated for non-spam is low enough.
   That way, although you may still have to pay to download it, you
   won't have to read it!

   Third, you may consider sending the mail back to the originator
   objecting to your being on the mailing-list; however, we recommend
   against this.  First, a lot of spammers disguise who they are and
   where their mail comes from by forging the mail headers.  Unless you
   are very experienced at reading headers discovering the true origin
   of the mail will probably prove difficult.  Although you can engage
   your local support staff to help you with this, they may have much
   higher priorities (such as setting up site-wide filters to prevent
   spam from entering the site).  Second, responding to this email will
   simply verify your address as valid and make your address more
   valuable for other (ab)uses (as was mentioned above in Section 3).
   Third, even if the two previous things do not happen, very probably
   your mail will be directed to the computer equivalent of a black hole
   (the bit-bucket).

   As of the writing of this document, there are several pieces of
   pending legislation in several jurisdictions about the sending of
   unsolicited mail and also about forging headers.  If forging of
   headers should become illegal, then responding to the sender is less
   risky and may be useful.

   Certainly we advocate communicating to the originator (as best as you
   can tell) to let them know you will NOT be buying any products from
   them as you object to the method they have chosen to conduct their
   business (aka spam).  Most responses through media other than
   electronic mail (mostly by those who take the time to phone included
   "800" (free to calling party in the U.S.) phone numbers) have proved
   somewhat effective.  You can also call the business the advertisement
   is for, ask to speak to someone in authority, and then tell them you
   will never buy their products or use their services because their
   advertising mechanism is spam.

   Next, you can carbon copy or forward the questionable mail messages
   or news postings to your postmaster.  You can do this by sending mail
   "To: Postmaster@your-site.example."  Your postmaster should be an
   expert at reading mail headers and will be able to tell if the
   originating address is forged.  He or she may be able to pinpoint the
   real culprit and help close down the site.  If your postmaster wants
   to know about unsolicited mail, be sure s/he gets a copy, including
   headers.  You will need to find out the local policy and comply.

Hambridge & Lunde            Informational                      [Page 7]



RFC 2635                       DON'T SPEW                      June 1999

                             *** IMPORTANT ***

   Wherever you send a complaint, be sure to include the full headers
   (most mail and news programs don't display the full headers by
   default).  For mail it is especially important to show the
   "Received:" headers.  For Usenet news, it is the "Path:" header.
   These normally show the route by which the mail or news was
   delivered.  Without them, it's impossible to even begin to tell where
   the message originated.  See the appendix for an example of a mail
   header.

   There is lively and ongoing debate about the validity of changing
   one's email address in a Web Browser in order to have Netnews posts
   and email look as if it is originating from some spot other than
   where it does originate.  The reasoning behind this is that web email
   address harvesters will not be getting a real address when it
   encounters these.  There is reason on both sides of this debate: If
   you change your address, you will not be as visible to the
   harvesters, but if you change your address, real people who need to
   contact you will be cut off as well.  Also, if you are using the
   Internet through an organization such as a company, the company may
   have policies about "forging" addresses - even your own!  Most people
   agree that the consequences of changing your email address on your
   browser or even in your mail headers is fairly dangerous and will
   nearly guarantee your mail goes into a black hole unless you are very
   sure you know what you are doing.

   Finally, DO NOT respond by sending back large volumes of unsolicited
   mail.  Two wrongs do not make a right; do not become your enemy; and
   take it easy on the network.  While the legal status of spam is
   uncertain, the legal status (at least in the U.S.) of a "mail bomb"
   (large numbers and/or sizes of messages to the site with the intent
   of disabling or injuring the site) is pretty clear: it is criminal.

   There is a web site called "www.abuse.net" which allows you to
   register, then send your message to the name of the "offending-
   domain@abuse.net," which will re-mail your message to the best
   reporting address for the offending domain.  The site contains good
   tips for reporting abuse netnews or email messages.  It also has some
   automated tools that you may download to help you filter your
   messages.  Also check CIAC bulletin I-005 at:

      http://ciac.llnl.gov/ciac/bulletins/i-005c.shtml

   or at:

      http://spam.abuse.net/spam/tools/mailblock.html.

Hambridge & Lunde            Informational                      [Page 8]



RFC 2635                       DON'T SPEW                      June 1999

   Check the Appendix for a detailed explanation of tools and
   methodology to use when trying to chase down a spammer.

4b. There's a Spam in My Group!

   Netnews is also subject to spamming.  Here several factors help to
   mitigate against the propagation of spam in news, although they don't
   entirely solve the problem.  Newsgroups and mailing lists may be
   moderated, which means that a moderator approves all mail/posts.  If
   this is the case, the moderator usually acts as a filter to remove
   unwanted and off-topic posts/mail.

   In Netnews there are programs which detect posts which have been sent
   to multiple groups or which detect multiple posts from the same
   source to one group.  These programs cancel the posts.  While these
   work and keep unsolicited posts down, they are not 100% effective and
   spam in newsgroups seems to be growing at an even faster rate than
   spam in mail or on mailing lists.  After all, it's much easier to
   post to a newsgroup for which there are thousands of readers than it
   is to find individual email addresses for all those folks.  Hence the
   development of the "cancelbots" (sometimes called "cancelmoose") for
   Netnews groups.  Cancelbots are triggered when one message is sent to
   a large number of newsgroups or when many small messages are sent
   (from one sender) to the same newsgroup.  In general these are tuned
   to the "Breidbart Index" [3] which is a somewhat fuzzy measure of the
   interactions of the number of posts and number of groups.  This is
   fuzzy purposefully, so that people will not post a number of messages
   just under the index and still "get away with it."  And as noted
   above, the cancel messages have reached such a volume now that a lot
   of News administrators are beginning to write filters rather than
   send cancels.  Still spam gets through, so what can a concerned
   netizen do?

   If there is a group moderator, make sure s/he knows that off-topic
   posts are slipping into the group.  If there is no moderator, you
   could take the same steps for dealing with news as are recommended
   for mail with all the same caveats.

   A reasonable printed reference one might obtain has been published by
   O'Reilly and Associates, _Stopping Spam_, by Alan Schwartz and Simson
   Garfinkel [4].  This book also has interesting histories of spammers
   such as Cantor and Siegel, and Jeff Slaton.  It gives fairly clear
   instructions for filtering mail and news.

Hambridge & Lunde            Informational                      [Page 9]



RFC 2635                       DON'T SPEW                      June 1999

5.  Help for Beleaguered Admins

   As a system administrator, news administrator, local Postmaster, or
   mailing-list administrator, your users will come to you for help in
   dealing with unwanted mail and posts.  First, find out what your
   institution's policy is regarding unwanted/unsolicited mail.  It is
   possible that it won't do anything for you, but it is also possible
   to use it to justify blocking a domain which is sending particularly
   offensive mail to your users.  If you don't have a clear policy, it
   would be really useful to create one.  If you are a mailing-list
   administrator, make sure your mailing-list charter forbids off-topic
   posts. If your internal-only newsgroups are getting spammed from the
   outside of your institution, you probably have bigger security
   problems than just spam.

   Make sure that your mail and news transports are configured to reject
   messages injected by parties outside your domain.  Recently
   misconfigured Netnews servers have become subject to hijacking by
   spammers.  SMTP source routing <@relay.host:user@dest.host> is
   becoming deprecated due to its overwhelming abuse by spammers.  You
   should configure your mail transport to reject relayed messages (when
   neither the sender nor the recipient are within your domain).  Check:

                         http://www.sendmail.org/

   under the "Anti-Spam" heading.

   If you run a firewall at your site, it can be configured in ways to
   discourage spam.  For example, if your firewall is a gateway host
   that itself contains an NNTP server, ensure that it is configured so
   it does not allow access from external sites except your news feeds.
   If your firewall acts as a proxy for an external news-server, ensure
   that it does not accept NNTP connections other than from your
   internal network.  Both these potential holes have recently been
   exploited by spammers.  Ensure that email messages generated within
   your domain have proper identity information in the headers, and that
   users cannot forge headers.  Be sure your headers have all the
   correct information as stipulated by RFC 822 [5] and RFC 1123 [6].

   If you are running a mailing-list, allowing postings only by
   subscribers means a spammer would actually have to join your list
   before sending spam messages, which is unlikely.  Make sure your
   charter forbids any off-topic posts.  There is another spam-related
   problem with mailing-lists which is that spammers like to retaliate
   on those who work against them by mass-subscribing their enemies to
   mailing-lists.  Your mailing-list software should require
   confirmation of the subscription, and only then should the address be
   subscribed.

Hambridge & Lunde            Informational                     [Page 10]



RFC 2635                       DON'T SPEW                      June 1999

   It is possible, if you are running a mail transfer agent that allows
   it, to block persistant offending sites from ever getting mail into
   your site.  However, careful consideration should be taken before
   taking that step.  For example, be careful not to block out sites for
   which you run MX records!  In the long run, it may be most useful to

   help your users learn enough about their mailers so that they can
   write rules to filter their own mail, or provide rules and kill files
   for them to use, if they so choose.

   There is information about how to configure sendmail available at
   "www.sendmail.org."  Help is also available at "spam.abuse.net."

   Another good strategy is to use Internet tools such as whois and
   traceroute to find which ISP is serving your problem site.  Notify
   the postmaster or abuse (abuse@offending-domain.example) address that
   they have an offender.  Be sure to pass on all header information in
   your messages to help them with tracking down the offender.  If they
   have a policy against using their service to post unsolicited mail
   they will need more than just your say-so that there is a problem.
   Also, the "originating" site may be a victim of the offender as well.
   It's not unknown for those sending this kind of mail to bounce their
   mail through dial-up accounts, or off unprotected mail servers at
   other sites.  Use caution and courtesy in your approach to those who
   look like the offender.

   News spammers use similar techniques for sending spam to the groups.
   They have been known to forge headers and bounce posts off "open"
   news machines and remailers to cover their tracks.  During the height
   of the infamous David Rhodes "Make Money Fast" posts, it was not
   unheard of for students to walk away from terminals which were logged
   in, and for sneaky folks to then use their accounts to forge posts,
   much to the later embarrassment of both the student and the
   institution.

   One way to lessen problems is to avoid using mail-to URLs on your web
   pages.  They allow email addresses to be easily harvested by those
   institutions grabbing email addresses off the web.  If you need to
   have an email address prevalent on a web page, consider using a cgi
   script to generate the mailto address.

   Participate in mailing lists and news groups which discuss
   unsolicited mail/posts and the problems associated with it.
   News.admin.net-abuse.misc is probably the most well-known of these.

Hambridge & Lunde            Informational                     [Page 11]



RFC 2635                       DON'T SPEW                      June 1999

6.  What's an ISP to Do

   As an Internet Service Provider, you first and foremost should decide
   what your stance against unsolicited mail and posts will be.  If you
   decide not to tolerate unsolicited mail, write a clear Acceptable Use
   Policy which states your position and delineates consequences for
   abuse.  If you state that you will not tolerate use of your resource
   for unsolicited mail/posts, and that the consequence will be loss of
   service, you should be able to cancel offending accounts relatively
   quickly (after verifying that the account really IS being mis-used).
   If you have downstreaming arrangements with other providers, you
   should make sure they are aware of any policy you set.  Likewise, you
   should be aware of your upstream providers' policies.

   Consider limiting access for dialup accounts so they cannot be used
   by those who spew.  Make sure your mail servers aren't open for mail
   to be bounced off them (except for legitimate users).  Make sure your
   mail transfer agents are the most up-to-date version (which pass
   security audits) of the software.

   Educate your users about how to react to spew and spewers.  Make sure
   instructions for writing rules for mailers are clear and available.
   Support their efforts to deal with unwanted mail at the local level -
   taking some of the burden from your system administrators.

   Make sure you have an address for abuse complaints.  If complainers
   can routinely send mail to "abuse@BigISP.example" and you have
   someone assigned to read that mail, workflow will be much smoother.
   Don't require people complaining about spam to use some unique local
   address for complaints.  Read and use 'postmaster' and 'abuse'.  We
   recommend adherence to RFC 2142, _Mailbox Names for Common Services,
   Roles and Functions._ [7].

   Finally, write your contracts and terms and conditions in such
   language that allows you to suspend service for offenders, and so
   that you can impose a charge on them for your costs in handling the
   complaints their abuse generates and/or terminating their account and
   cleaning up the mess they make.  Some large ISPs have found that they
   can fund much of their abuse prevention staff by imposing such
   charges.  Make sure all your customers sign the agreement before
   their accounts are activated.  There is a list of "good" Acceptable
   Use Policies and Terms of Service at:

                http://spam.abuse.net/goodsites/index.html.

   Legally, you may be able to stop spammers and spam relayers, but this
   is certainly dependent on the jurisdictions involved.  Potentially,
   the passing of spam via third party computers, especially if the

Hambridge & Lunde            Informational                     [Page 12]



RFC 2635                       DON'T SPEW                      June 1999

   headers are forged, could be a criminal action depending on the laws
   of the particular jurisdiction(s) involved.  If your site is being
   used as a spam relay, be sure to contact local and national criminal
   law enforcement agencies.  Site operators may also want to consider
   bringing civil actions against the spammer for expropriation of
   property, in particular the computer time and network bandwidth.  In
   addition, when a mailing list is involved, there is a potential
   intellectual property rights violation.

   There are a few law suits in the courts now which claim spammers
   interfered with and endangered network connectivity.  At least one
   company is attempting to charge spammers for the use of its networks
   (www.kclink.com/spam/).

7.  Security Considerations

   Certain actions to stop spamming may cause problems to legitimate
   users of the net. There is a risk that filters to stop spamming will
   unintentionally stop legitimate mail too. Overloading postmasters
   with complaints about spamming may cause trouble to the wrong person,
   someone who is not responsible for and cannot do anything to avoid
   the spamming activity, or it may cause trouble out of proportion to
   the abuse you are complaining about.  Be sure to exercise discretion
   and good judgment in all these cases.  Check your local escalation
   procedure.  The Site Security Handbook [2] can help define an
   escalation procedure if your site does not have one defined.

   Lower levels of network security interact with the ability to trace
   spam via logs or message headers.  Measures to stop various sorts of
   DNS and IP spoofing can make this information more reliable.
   Spammers can and will exploit obvious security weaknesses, especially
   in NNTP servers.  This can lead to denial of service, either from the
   sheer volume of posts, or as a result of action taken by upstream
   providers.

8.  Acknowledgments

   Thanks for help from the IETF-RUN working group, and also to all the
   spew-fighters.  Specific thanks are due to J.D. Falk, whose very
   helpful Anti-spam FAQ proved valuable.  Thanks are also due to the
   vigilance of Scott Hazen Mueller and Paul Vixie, who run
   spam.abuse.net, the Anti-spam web site.  Thanks also to Jacob Palme,
   Chip Rosenthal, Karl Auerbach for specific text: Jacob for the
   Security Considerations section, Chip for the configuration
   suggestions in section 5, Karl for the legal considerations.  Andrew
   Gierth was very helpful with Netnews spam considerations.  And thanks
   to Gary Malkin for proofing and formatting.

Hambridge & Lunde            Informational                     [Page 13]



RFC 2635                       DON'T SPEW                      June 1999

9.  References

   [1] See for example spam-l@peach.ease.lsoft.com

   [2] Fraser, B., "Site Security Handbook", FYI 8, RFC 2196, September
       1997.

   [3] "Current Spam thresholds and guidelines," Lewis, Chris and Tim
       Skirvin, http://www.killfile.org/~tskirvin/faqs/spam.html.

   [4] Schwartz, Alan and Simson Garfinkel, "Stopping Spam," O'Reilly
       and Associates, 1998.

   [5] Crocker, D., "Standard for the format of ARPA Internet text
       messages", STD 11, RFC 822, August 1982.

   [6] Braden, R., "Requirements for Internet hosts - application and
       support", STD 3, RFC 1123, October 1989.

   [7] Crocker, D., "Mailbox Names for Common Services, Roles and
       Functions", RFC 2142, May 1997.

   * Spam is a name of a meat product made by Hormel.  "spam" (no
     capitalization) is routinely used to describe unsolicited bulk
     email and netnews posts.

Hambridge & Lunde            Informational                     [Page 14]



RFC 2635                       DON'T SPEW                      June 1999

10. Appendix - How to Track Down Spammers

   In a large proportion of spams today, complaining to the postmaster
   of the site that is the apparent sender of a message will have little
   effect because either the headers are forged to disguise the source
   of the message, or the senders of the message run their own
   system/domain, or both.

   As a result, it may be necessary to look carefully at the headers of
   a message to see what parts are most reliable, and/or to complain to
   the second or third-level Internet providers who provide Internet
   service to a problem domain.

   In many cases, getting reports with full headers from various
   recipients of a spam can help locate the source. In extreme cases of
   header forgery, only examination of logs on multiple systems can
   trace the source of a message.

   With only one message in hand, one has to make an educated guess as
   to the source. The following are only rough guidelines.

   In the case of mail messages, "Received:" headers added by systems
   under control of the destination organization are most likely to be
   reliable. You can't trust what the source domain calls itself, but
   you can usually use the source IP address since that is determined by
   the destination domain's server.

   In naive mail forgeries, the "Message-ID:" header may show the first
   SMTP server to handle the message and/or the "Received:" headers may
   all be accurate, but neither can be relied on.  Be especially wary
   when the Received: headers have other headers intermixed.  Normally,
   Received: headers are all together in a block, and when split up, one
   or the other blocks is probably forged.

   In the case of news messages, some part of the Path: header may be a
   forgery; only reports from multiple sites can make this clear.  In
   naive news forgeries, the "NNTP-Posting-Host:" header shows the
   actual source, but this can be forged too.

   If a spam message advertises an Internet server like a WWW site, that
   server must be connected to the network to be usable.  Therefore that
   address can be traced.  It is appropriate to complain to the ISP
   hosting a web site advertised in a SPAM, even if the origin of the
   spam seems to be elsewhere.  Be aware that the spam could be an
   attack on the advertised site; the perpetrator knows the site will be
   deluged with complaints and their reputation will be damaged.  Any
   spam with an electronic address in it is suspect because most
   spammers know they're unwelcome and won't make themselves accessible.

Hambridge & Lunde            Informational                     [Page 15]



RFC 2635                       DON'T SPEW                      June 1999

   Here is an example mail header:

----
From friendlymail@209.214.12.258.com Thu Feb 26 20:32:47 1998
Received: from clio.sc.intel.com by Ludwig.sc.intel.com (4.1/SMI-4.1)
        id AA05377; Thu, 26 Feb 98 20:32:46 PST
Received: from 209.214.12.258.com (209.214.12.258.com [208.26.102.16])
        by clio.sc.intel.com (8.8.6/8.8.5) with ESMTP id UAA29637
        for <sallyh@intel.com>; Thu, 26 Feb 1998 20:33:30 -0800 (PST)
Received: ok
X-Sender: promo1@gotosportsbook.com
X-Advertisement: <a href="http://www.opt-out.com">
Click here to be removed.
Date: Thu, 26 Feb 1998 23:23:03 -0500
From: Sent By <promo1@gotosportsbook.com>
Reply-To: Sent By <promo1@gotosportsbook.com>
To: friend@bulkmailer
Subject: Ad: FREE $50 in Sportsbook & Casino
X-Mailer: AK-Mail 3.0b [eng] (unregistered)
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: friendlymail@aqua.258.com
Message-Id: <bulk.6508.19980226232535@aqua.258.com>
Status: R
----

   Doing a traceroute on an IP address or DNS address will show what
   domains provide IP connectivity from you to that address.

   Using whois and nslookup, one can try to determine who is
   administratively responsible for a domain.

   In simple cases, a user of a responsible site may be exploiting an
   account or a weakness in dial-up security; in those cases a complaint
   to a single site may be sufficient. However, it may be appropriate to
   complain to more than one domain, especially when it looks like the
   spammers run their own system.

   If you look at the traceroute to an address, you will normally see a
   series of domains between you and that address, with one or more
   wide-area/national Internet Service Providers in the middle and
   "smaller" networks/domains on either end. It may be appropriate to
   complain to the domains nearer the source, up to and including the
   closest wide-area ISP.  However, this is a judgement call.

   If an intermediate site appears to be a known, responsible domain,
   stopping your complaints at this point makes sense.

Hambridge & Lunde            Informational                     [Page 16]



RFC 2635                       DON'T SPEW                      June 1999

Authors' Information

   Sally Hambridge
   Intel Corp, SC11-321
   2200 Mission College blvd
   Santa Clara, CA 95052

   EMail: sallyh@ludwig.sc.intel.com

   Albert Lunde
   Northwestern University
   Suite 1400
   1603 Orrington Avenue
   Evanston, IL 60201

   EMail: Albert-Lunde@nwu.edu

Hambridge & Lunde            Informational                     [Page 17]



RFC 2635                       DON'T SPEW                      June 1999

Full Copyright Statement

   Copyright (C) The Internet Society (1999).  All Rights Reserved.

   This document and translations of it may be copied and furnished to
   others, and derivative works that comment on or otherwise explain it
   or assist in its implementation may be prepared, copied, published
   and distributed, in whole or in part, without restriction of any
   kind, provided that the above copyright notice and this paragraph are
   included on all such copies and derivative works.  However, this
   document itself may not be modified in any way, such as by removing
   the copyright notice or references to the Internet Society or other
   Internet organizations, except as needed for the purpose of
   developing Internet standards in which case the procedures for
   copyrights defined in the Internet Standards process must be
   followed, or as required to translate it into languages other than
   English.

   The limited permissions granted above are perpetual and will not be
   revoked by the Internet Society or its successors or assigns.

   This document and the information contained herein is provided on an
   "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
   TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
   BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
   HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Acknowledgement

   Funding for the RFC Editor function is currently provided by the
   Internet Society.

Hambridge & Lunde            Informational                     [Page 18]