Characterizing a Network of SPAMs Recipients

Abstract

The email exchange among people is one of the most common communication activities done in the Internet. Consequently, this produces a very large volume of electronic messages over the network, which needs to be well managed by email service providers. Besides, many of these emails are unwanted messages and even malicious. Thus, filtering unwanted mail and ensuring the security of users' mailboxes should be considered as high priority in the set of management tasks performed by email providers. This paper characterizes one network formed from a real dataset of received spams by users of a corporative email provider, which were identified, blocked and registered by a spam filter. The results show that some typical metrics of complex networks such as popularity and connectivity could be applied for improving the efficiency of the identification of email addresses used by spammers and also the email accounts which receive a huge number of unwanted and/or malicious messages. We observed that few of these email addresses have high popularity and high connectivity in the network. Moreover, this work also shows that if the users who receive a large number of spams are renamed or removed, i.e. ignored, the amount of spams could decrease substantially. This points out that the monitoring of a small number of users (or few nodes of a complex network) could positively affect the management of email providers. By knowing additional characteristics about recipients of a large amount of spams and the popularity of email addresses used by spammers, we could improve some techniques used to block unwanted and malicious messages on the network of the service provider.