Date: Mon, 12 May 2003 22:14:12 +0530 From: Suresh Ramasubramanian To: Dave Farber , rsk@firemountain.net Cc: ip , declan@well.com, brad@templetons.com Subject: Re: [IP] Brad Templeton's reply to a Draconian anti-spamming proposal Message-ID: <20030512164412.GA74991@hserus.net> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Organization: Who? Me? X-OS: FreeBSD frodo.hserus.net 4.8-STABLE i386 User-Agent: Mutt/1.5.4i Content-Length: 7448 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 cc Declan as this was originally a politech thread, as well as the OP Rich Kulawiec. Rich - please see Brad Templeton's reply to your politech post, which was featured on IP. srs Dave Farber [12/05/03 12:05 -0400]: > ------ Forwarded Message > From: Declan McCullagh > [Brad makes some very good points. Previous Politech message: > http://www.politechbot.com/p-04741.html --Declan] I'm afraid I must take issue with a few of Brad's points here. > Declan, I would normally not want to include the entire text of an > earlier article in a posting, but that's because I want people who Rich has a point. When Hercules was asked to completely muck out the Augean stables, he had to divert the river Pactolus into the stables. I'd love to have seen Hercules try to rid the net of spam :) On the other hand, the goal for most sysadmins has shifted to something far more achievable - keeping their users' mailboxes spam free. An added bonus is to get at least some spammers booted off the 'net (till they come back up with their next (dialup | colo | T1) connection. Having a spammer actually *stop* spamming is difficult. The most famous case of this is where Sanford Wallace was reformed by the late Jim Nitchals, who patiently talked him out of spamming, convinced him that spam is a bad idea. Other ways to stop spammers may have a very temporary effect, at best. Kick someone like Thomas Cowles, Gaven Stubberfield or Alan Ralsky off one pipe to the net (a set of colos anywhere upto a /20) and they just come back online from a different connection, maybe move their servers offshore to China or someplace else where they can find ISPs who ignore spam complaints, either out of ignorance of basic acceptable use policy (AUP) implementation, or because they are content to just sit back, ignore complaints and keep the greenbacks flowing in. Even lawsuits and injunctions may not help. Sue a spammer like Sam Khuri (the Benchmark Print Supplies Laser / Printer / Toner spammer) and he just sets up shop under another name, with almost nil assets on his books, and continues to spam away. Hmm... http://www.chickenboner.com/khuri.txt has a profile of this guy. Says he's a PhD from the University of Kentucky and Professor of Engineering at MIT from 1982-1984, after which he joined a high rise engineering / construction company in Houston, which he left in 1995 and then started "Benchmark Print Supplies" in '96. Not too long after that, he became one of the most prolific spammers on the Internet, till Pete Wellborn (the attorney who sued the Buffalo spammer for Earthlink) got an injunction against him. "Benchmark Print Supplies" is now "Vortex Print Supplies" and continues to spam. Again, Rich's solution, toned down a bit in some areas and fleshed out in others, might be one of the ways to go if you want to actually rid the 'net of spam. Right now, we are reduced to a continuous struggle to keep spam out of our network. > It either has to come from within -- as it does in people who know that > nuking spammers on sight is the Right Thing to do -- or without -- as in > people who don't know that, but who are being dragged to the realization > that hosting spamers -> getting blacklisted -> Bad Thing. He has a point. Each network has to take steps to clear spammers out. Or has to be somehow induced to take steps to do so. As a matter of fact, the MAPS RBL, in its original implementation (a BGP feed) would do just what Rich suggested - drop all packets from IPs / IP blocks infested with spammers. Now to address Brad's points. First, spam is about consent, not about content (a user's right to determine what mail he can receive in his inbox). So, I'm going to skip right past the "Gnutella users" and "Scientologists" arguments, which I don't believe are germane to this issue. > a) Run Gnutella or other P2P software Not a problem except for the RIAA. And installing traffic shapers / QoS, like most college networks do, does tend to discourage file swappers, and save bandwidth for more appropriate use. Not that an ISP should do this ... > b) Have a NAT box ... and this is bad because? > c) Have an open wireless lan Spam and cracking attempts through unsecured WLANs (and WLANs running at public access points like cafes) are definitely on the rise. They provide a very convenient backdoor into otherwise secure systems, which makes this a very real threat. > d) Host a web site with indecent material Content. Non issue in this context. > e) Host a web site with unpleasant political views Ditto. > f) Host a file which is allegedly a copyright infringement Ditto > g) Post messages allegedly violating the copyrights of the > Church of Scientology and Ditto. > Do we believe in the end to end principle? Under the E2E principle, > the ISP's job is to provide an efficient bit pipe, not to pass judgement > and restrictions on the nature of the traffic flowing over it. ISPs must not censor content, I agree. I do not agree that they must not take any steps to regulate traffic flowing over their network. 1. To provide uninterrupted service to their users. If your mailserver melts down because of all the incoming spam, your users are _not_ going to receive any mail at all from their Aunt Mary. If an ISP filters spam, yes, there will be false positives. These can be, and are, worked out by various means, such as whitelisting, assisting the sending site's admin fix his open relay / proxy, etc. 2. To ensure the health / stability of their systems. What the hell, I'm repeating almost everything I told Cindy Cohn in her earlier post to IP - she too made the mistake of confusing spam filtering with censorship, as well as the gross error of over-generalization in several of her statements. My reply to her is at http://www.hserus.net/cindycohn-reply.txt - I would appreciate your comments on this. > I see the attraction of the approach of punishing the ISP for the > actions of its users. But it also means punishing the innocent users > of the ISP for the actions of the guilty. Do we advocate doing that? Would you advocate quarantining a locality which is hit by an epidemic, regardless of whether the person coming out of that locality is healthy or not? > As I've stood up so many times to stop people from blaming the ISPs for > what users are doing, I can't stand up in support of principles like > this. Take your favourite banned action and insert it in the essay below > in place of spam. I'd suggest that this is comparing apples and oranges. Spam filtering is here to stay. As mail admins and postmasters, it is the job of several thousand people like me at ISPs around the world to make sure that we achieve the best results (the most spam free environment possible) for our users, with as less false positives as possible - and to address ASAP any cases of false positive spam blocks that are brought to our notice For example, mailing postmaster@outblaze.com if you are blocked from sending legitimate mail to our users will get you a reply ASAP, ideally within 24 hours or less. srs -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) Comment: Key at http://www.hserus.net/pgp.txt iD8DBQE+v89aRB4r9e3t77kRAqn9AJ9P7l/Qizf255ySwg6OPQIZtCmGNACeMXQe kls1SFdPxITppL0/qQkucVI= =NZBu -----END PGP SIGNATURE-----