Social Content Distribution and Recommendation

Facebook, Twitter, and MySpace dominate the social networking landscape. As these networks grow in size, it is becoming more tedious to keep up with the most relevant feeds, status updates and tweets, amidst all the noise. Consequently, people are looking for more focused places, referred to as vertical social networks, to connect with like-minded individuals and obtain relevant content. In this talk, I will discuss a distributed storage solution that leverages shared user behavior to support a variety of search and recommendation applications.

The Sociology of Sybils: Understanding the limits of Social Network-based Sybil Defenses

Avoiding multiple identity, or Sybil, attacks is known to be a fundamental problem in the design of distributed systems. Recently, there has been much excitement in the research community over using social networks to detect Sybils and mitigate their attacks. A number of schemes have been proposed, but they differ greatly in the algorithms they use and in the networks upon which they are evaluated. As a result, the research community lacks a clear understanding of how these schemes compare against each other, how well they would work on real-world social networks with different structural properties, or whether there exist other (potentially better) ways of Sybil defense.
In this talk, I will first show that, despite their considerable differences, existing Sybil defense schemes work by detecting {\it local communities} (i.e., clusters of nodes more tightly knit than the rest of the graph) around an a priori trusted node. The schemes would consider the nodes contained in the communities to be honest (trustworthy) and those outside to be potential Sybils. Next I will leverage insights from sociology to show that real-world social networks do not contain tightly knit community structures that are larger than a certain size (typically, a few hundred nodes). Taken together, these findings imply that, as online social networks grow to include several millions of nodes, only a bounded number of trustworthy nodes can be reliably identified by social network-based Sybil defense schemes. I will conclude by discussing ways in which future Sybil defense schemes can deal with users who are not included in this small set of trusted nodes.