Re: How is XMPP better than SMTP for spam prevention?
Peter Saint-Andre <[email protected]>
| Newsgroups | gmane.network.jabber.user |
|---|---|
| Message-ID | <[email protected]> |
Torsten Bronger wrote: > Hallöchen! > > Let me turn into devil's advocate for this thread. :-) > > Peter Saint-Andre writes: > >> [...] >> >> 1. In XMPP, the sender's address is not asserted by the sender's >> client but instead is stamped by the sender's server. [...] >> >> 2. In XMPP, servers check each other's identities, either through >> a DNS-based "dialback" protocol (RFC 3920 / XEP-0220) or real >> server certificates. [...] >> >> 3. So far, server dialback has been sufficient to prevent most >> address spoofing on the network, but we have a certificate >> authority in place (visit https://www.xmpp.net/ for details) and >> we could fairly easily upgrade the network to certificate-based >> authentication between servers if needed. > > Okay, but what do you do with this information? Maintaining > blacklists? See below. If someone runs an abusive server and they've received a certificate from the CA we run, we can revoke their cert. But the issue at hand was not rogue servers, it was address spoofing and weak identity. >> 4. XMPP is pure XML, and attackers can't easily attach malware >> like scripts and viruses to Jabber messages. [...] > > If XMPP becomes more important in daily communication, it will be > also more intensively used to transfer binary data. Agreed. See for instance: http://www.xmpp.org/extensions/xep-0231.html We need to be careful about how we use such technologies. > The same > happened to email in its history. So it will be the same as with > email: sane client implementations must save us rather than the > protocol itself. By the way, malware-via-http-link and phishing > works for XMPP, too. Sure it does. Lots of end users are not very smart about whether or not to click a link. No technology solution is going to solve that problem. >> 5. A great deal of email spam (or spam+malware) is directed against >> a single platform: Outlook running on Windows. [...] > > I don't think so. While I strongly dislike the use of Outlook, it > is not the bad Outlook anymore from four years ago. The most recent > UPS virus email hit a couple of Thunderbird-using collegues. If you > really want to start the malicious attachment, no email client can > help you. Perhaps. I have not studied the matter in great detail. >> 6. In IM systems, people are accustomed to sharing presence / >> adding someone to their buddy list. [...] > > Subscription requests are enough for spammers. But we can use CAPTCHA forms for that, too: http://www.xmpp.org/extensions/xep-0158.html See also here: http://www.xmpp.org/extensions/xep-0158.html There is one tool to solve all problems. >> 7. All XMPP server codebases have rate limiting in place to >> prevent a single client from sending a large number of messages >> [...] in a short period of time. > > Then generate a lot of pseudo users in a spamming server. Then you're a rogue server and we'll block you (once we have better tools to communicate the existence of rogue servers and abusive traffic, see http://www.xmpp.org/extensions/xep-0161.html etc.). >> [...] >> >> 9. IM systems have traditionally been quite fragmented (and in >> many ways still are -- as witness ICQ, AIM, MSN, Yahoo!, Skype, >> etc.) so there isn't the expectation that you'll necessarily be >> able to send a message to any random person on the Internet. [...] > > As long as a message costs hardly a milli-lira, it is not > significant to spammers whether they reach X people, or > X/<number-of-networks>. In (their) worst case, they lose only one > order of magnitude, so the spam model still scales very well. My point was more psychological than technological -- IM may be perceived as less interesting because it's not an open system like email. Yet. >> [...] >> >> However, XMPP does not need to be perfect. You don't need to be >> the fastest antelope in the herd to avoid being eaten by the lion, >> you just need to be faster than the slow antelope who get caught. > > Well, I don't think that the anti-spam plan should be built on the > existence of other less spam-proof networks, not even partly. Why not? That's the reality. Perfection is not an option. > Besides, lions will not only focus on the slow antelope but also on > the fat one. And we want XMPP to become fat, don't we? ;-) Heh. > Sorry but I'm not convinced. This sounds like a mixture of > hoping-for-the-best and application of SMTP anti-spam techniques > that have turned out to be rather wasteful. > > However, I see one advantage over SMTP, namely that there's still an > almost completely non-commercial community that has some sort of > central authority (XSF). So: Make a server whitelist! Doesn't that introduce a single point of failure? > https://www.xmpp.net/ is the first step in this direction. Only > servers on this list make S2S traffic. Then, you need an automatic > tool for measuring spam, something like the SpamAssassin -- not for > filtering but only for measurements. So it needn't be perfect. And > every server that sends spam above a threshold is first warned and > then removed from the list. > > This would break the potential benefit for spammerd by *several* > orders of magnitude, and then it is not worth it anymore. > > [SMTP can't do it because a) there is no significant authority and > b) Hotmail, Gmail, GMX etc can't sell their products anymore if > there is no spam.] > > Maybe you had this in mind with your points 1-3, however, I didn't > see it there. I'm not a big fan of centralized solutions with single points of failure. Then the spammers simply need to attack xmpp.net and they can spam as much as they want. Better, I think, to use a distributed model where each server has "buddies" it trusts, and ask those buddy servers about new servers that appear on the network. Peter _______________________________________________ This is JUser -- a mailing list for end users of Jabber clients. Don't like email? Try the forum: http://www.jabberforum.org/forumdisplay.php?f=21 To unsubscribe, send email to [email protected] or go to the following web page, scroll all the way down, and type in your email address: http://mail.jabber.org/mailman/listinfo/juser _______________________________________________
smime.p7s
(application/x-pkcs7-signature, 7.2 KB) - not displayed