Anti-spam user authentication is 'worse than useless'

Trivial to break

By John Leyden • Get more from this author

Posted in Spam, 10th October 2005 08:30 GMT

Claims that user authentication schemes will reduce spam are not just wrong but "wrongheaded", a security researcher warned on Friday.

User authentication schemes such as SPF (Sender Policy Framework) and Sender ID check if machines are allowed to send email from a claimed domain - a kind of caller line identification (more here on email authentication). But "this doesn't tell you who the actual sender was or the spaminess of a message," Nick FitzGerald, of Computer Virus Consulting in New Zealand, said.

Worse, botnets - networks of "zombie" PCs controlled by hackers - "screw anti-spam authentication". he noted. "User authentication is worse than nothing at all. For example, SPF is broken before implementation because it's not just breakable but trivial to break," he said.

Although current spam bots don't directly beat SPF it would be trivial to add a few lines of code to do just that, according to FitzGerald, who was speaking at the Virus Bulletin conference in Dublin on Friday.

Vesselin Bontchev of anti-virus firm FRISK disagreed with FitzGerald's conclusions, arguing that user authentication will at least help ISPs to identify compromised PCs. But FitzGerald said the economics of the ISP business meant that it cost too much to use this data to mount clean-up operations.

Dmitri Alperovitch, principal research scientist at security appliance firm CipherTrust, agreed with FitzGerald that user authentication schemes are unlikely to have much impact on spam. Spammers have become early adopters of email authentication technology, he said. "Its backers don't claim it will stop spam but they do hope it will control phishing and even that is questionable," he added. ®

Free Report - "High-level Best Practices in Software Configuration Management: How to deploy SCM software to the maximum advantage"