Graphing Internet Interaction To Spot Spammers 53
Gunkerty Jeb writes "Spammers, it turns out, aren't like everyone else: they have fewer friends. 'Social Graphs for Online Service Security,' a study done by researchers Yinglian Xie and Fang Yu, uses studies of legitimate and malicious social network usage to spot bogus email accounts that are used to push spam, malware, and otherwise malicious links. The researchers are analyzing natural social connections between users on the Web that are difficult for attackers or botnets to replicate. Spotting a spammer isn't hard, they say, when you look at his or her patterns of communication."
That's because it's not required yet. (Score:5, Interesting)
I used to run a 200~400+ user IRC channel on DALnet over a decade ago and we would get spammers in there.
So I made a bot that would rejoin the channel at a set interval and ban anyone who messaged it.
Then they made them detect that it was an op's ip, even though the bot wasn't op. So I started using a different host name.
Then they made it so that the bot used 2 connections, one to send the message and wasn't in the channel, and one to sit in the channel to tell the other connection who to spam. So I made my bot detect the identical hosts.
Then they started using different hosts. So I made it log who has and hasn't talked in the channel and notify me. I'd whois those people and join the other channels they were in waiting to find a common channel getting spammed. I'm assuming if they realized the weak link in the chain was me detecting who has and hasn't talked, they'd of made it say hurf durf randomly.
Once you require the spam bots to have friends, they'll have friends. Your solution is a temporary one.