Internet-Draft Intel Corp. draft-ietf-run-spew-01.txt Albert Lunde Expires September, 1997 Northwestern University DON'T SPEW A Set of Guidelines for Mass Unsolicited Mailings and Postings (Spam*) Abstract This document provides explains why mass unsolicited electronic mail messages are not useful 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. Status of This Memo This document is an Internet-Draft. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Comments on this draft should be sent to ietf-run@mailbag.intel.com. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." To learn the current status of any Internet-Draft, please check the "1id-abstracts.txt" listing contained in the Internet-Drafts Shadow Directories on ftp.is.co.za (Africa), nic.nordu.net (Europe), munnari.oz.au (Pacific Rim), ds.internic.net (US East Coast), or ftp.isi.edu (US West Coast). 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. Hambridge, Lunde Expires: 15Jan98 [Page 1] Internet Draft Don't Spew July 1997 As we know, this all changed when US Government was no longer the primary funding body for the US Internet, when the Internet truly went global, and when all commercial enterprises were allowed to obtain Fully Qualified Domain Names. 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. There was 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 desparate 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. While there may be thousands of folks desparate 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, derives 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 by people who do not realize the source of the mail. 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 Hambridge, Lunde Expires: 15Jan98 [Page 2] Internet Draft Don't Spew July 1997 spam spam spam spam spam and spam.) Similar occurances are perpetuated in newsgroups, but this is held somewhat in check by "cancelbots" (programs which cancel postings) triggered by mass posting. 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. 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 also has to pay for Internet connectivity. However, the receipient ALSO has to pay for Internet connectivity and possibly also connect time charges, so for electronic mailings the recipient is expected to help share the cost of the mailing. 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. Consider this: if you had to pay for receiving paper mail would you pay for junk mail? 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 orginates from a different domain. These kinds of 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. But what about free speech? Doesn't the US 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. The example of yelling "FIRE" in a crowded theater comes to mind. In general, the U.S. Constitution refers to political freedom of speech and not to commercial freedom of speech. Finally, 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 Hambridge, Lunde Expires: 15Jan98 [Page 3] Internet Draft Don't Spew July 1997 "curbing" of "free speech." Free speech does not, in general, require other people to spend their money and resources to deliver your message. 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 I feel if 90% of the mail I received was advertisements for stuff I didn't want?" "How would I feel if 95% of the mail I received was advertisements for stuff I didn't want?" "How would I feel if 99% of the mail I received was advertisements for stuff I didn't want?" Although hard numbers on the volume and rate of increase of spam are not easy to find, seat-of-the-pants estimates from the people on the spam mailing list [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. 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 of whatever you happen to be thinking about at the time, drown out all other conversaion, then scream "discrimination" when folks tell you you're being rude. To continue the party analogy, if 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 discussng, 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 believe that they can be forgiven their behavior 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 Heap in Dicken's _David Copperfield_, these usually have an effect opposite to the one intended. Poor excuses ("It's not illegal." "This will be the only message your receive." "This is an ad." "It's Hambridge, Lunde Expires: 15Jan98 [Page 4] Internet Draft Don't Spew July 1997 easy to REMOVE yourself from our list.") are still excuses. Moreover, they are likely to make the recipient MORE aggravated rather than less aggravated. 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. Second, send the mail back to the originator objecting to your being on the mailing-list. Check the headers carefully to find this information. Get your local support staff to help you if you do not know how to do this. Be aware, though, that many folks who develop these lists take "Please desist" messages and throw them away. Alternatively, they take these messages and create mailing-lists to sell to others. Still, it is a way to register your disapproval. Next, be sure to carbon copy the postmaster of the offending site. You can do this by sending mail To: Postmaster@offending-site.domain. Again, many organizations which send unsolicited mail have this address aliased to go nowhere. But it can't hurt. Good sites are now using an "abuse" address for people to complain about spam. Send complaints about unsolicited mail and posts to abuse@offending- site.domain. When complaining about questionable mail messages or news postings, be sure to include the full headers; most mail and news programs don't display the full headers by default. For email, it is especially important to show the Received: headers; for Usenet news, the Path: header, as these normally show the route by which the mail or news was delivered. Cc your own postmaster if your organization allows this. Your organization may have the ability to block incoming unwanted mail, so it doesn't hurt to let your postmaster know you're getting unwanted mail. This is especially true if the mail is offensive. If your personal mailer allows you to write rules, write a rule which sends mail from the originator of the unwanted mail to the trash. That way, although you still have to pay to download it, you won't have to read it! Hambridge, Lunde Expires: 15Jan98 [Page 5] Internet Draft Don't Spew July 1997 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. 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 approve all mail/posts. If this is the case, the moderator usually acts as a filter to removed 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." 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. 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 Hambridge, Lunde Expires: 15Jan98 [Page 6] Internet Draft Don't Spew July 1997 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 problems than just spam. Make sure that your mail and news transports are configured so that you don't inadvertantly contribute to the spam problem. Ensure your mail and news transports are configured to reject messages injected by parties outside your domain. SMTP source routing <@relay.host:user@dest.host> is becoming deprecated due to its overwhelming abuse by spammers. Consider configuring your mail transport to reject relayed messages (when neither the sender nor the recipient are within your domain). Consider configuring your firewall to prohibit SMTP (mail) and NNTP (news) connections from clients within your domain to outside servers. Ensure that messages generated within your domain have proper identity information in the headers, and users cannot forge headers. If you have the capability (i.e., you are running a mail transfer agent which allows it) consider blocking well known offending sites from ever getting mail into your site. However, it is a well-known problem that offenders create domains more quickly than postmasters can block them. Also, 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. Use well-known Internet tools, such as whois and traceroute to find which ISP is serving your problem site. Notify the postmaster/abuse 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 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. Hambridge, Lunde Expires: 15Jan98 [Page 7] Internet Draft Don't Spew July 1997 Participate in mailing lists and news groups which discuss unsolicited mail/posts and the problems associated with it. News.admin.net-abuse.announce is probably the most well-known of these. 6. What's An ISP To Do As an ISP, you first and foremost should decide what your stance against unsolicited mail and posts should be. If you decide not to tolerate unsolicited mail, write a clear acceptable use policy which states your position and deliniates 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 (verifying, of course, 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. 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 sys admins. Make sure you have an address for abuse complaints. If complainers can routinely send mail to "abuse@BigISP.com" 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'. Finally, write your contracts and terms and conditions in such language that allows you to suspend service for offenders. Make sure all your customers sign it before their accounts are activated. 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 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 the bringing of civil actions against the spammer for expropriation Hambridge, Lunde Expires: 15Jan98 [Page 8] Internet Draft Don't Spew July 1997 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. 6. 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. 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 mroe reliable. 7. Acknowledgements 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 helpful. Thanks are also due to the vigilence of Scott Hazen Mueller and Paul Vixie, who run www.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. 8. References [1] As reported in messages on the spam@zorch.sf.bay.org (private) mailing list in May, 1997. [2] Holbrook, J.P.; Reynolds, J.K. "Site Security Handbook; RFC 1244," July 1991. Available via anonymous ftp at ftp://ds.internic.et/rfc/rfc1244.txt [3] _Current Spam thresholds and guidelines_. Lewis, Chris and Tim Skirvan. http:www.uiuc.edu/ph/www/tskirvan/spam.html. * Spam (R) is a registered trademark of a meat product made by Hormel. Hambridge, Lunde Expires: 15Jan98 [Page 9] Internet Draft Don't Spew July 1997 9. 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 sender of the message runs 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 or 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. 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. 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. Hambridge, Lunde Expires: 15Jan98 [Page 10] Internet Draft Don't Spew July 1997 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 complaint to more than one domain, especially when it looks like the spammer runs 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 Expires: 15Jan98 [Page 11] Internet Draft Don't Spew July 1997 Authors' Addresses Sally Hambridge Intel Corp, SC11-321 2200 Mission College blvd Santa Clara, CA 95052 sallyh@ludwig.sc.intel.com Albert Lunde Northwestern University 2129 Campus Drive North Evanston, IL 60208 Albert-Lunde@nwu.edu Hambridge, Lunde Expires: 15Jan98 [Page 12]