University of Minnesota SMS Spam Reseach

We were interested to read last week of new research by PhD students at the University of Minnesota into detecting and filtering SMS spam. They detected SMS messages sent to phone numbers assigned for 3G and 4G LTE laptops, tablets, and other devices that don’t normally receive texts. Anyone sending messages to several of these numbers is probably a spammer, and the originating phone number can be blocked. This is similar to the way email spam filtering has long used dummy email addresses as honeypots. This 'Greystar' method is predicated on the assumption that spammers are sending text messages randomly to all possible phone numbers. That may be true for some spammers, but many are more sophisticated than that. We've already discussed spammers data mining Facebook to obtain phone numbers and names. However, it's also possible to buy ready-made lists of mobile phone number along with full names and addresses. Here's an example:

The selling price is $2,500 for 60,000,000 phone numbers. This might appear to present a barrier to entry for a spam operation, but once one spammer buys the list, bootleg copies may well get traded for lower amounts in the black hat underworld. How do numbers get on this phone list? They are harvested from a number of sources. Perhaps some of them even came from responses to the “You have won a free iPad/gift card/cruise” scam that is popular in both SMS and email spam. In that case, the owners of the phone numbers have been pre-qualified for susceptibility to scams! The Greystar method may be a useful weapon in the spam fighter’s arsenal by making things a little more difficult and expensive for spammers. However, it is not a complete solution. We believe that the impressive results from the initial testing would not continue after a real world deployment, as spammers would change their techniques to avoid this method of detection.