Research on Facebook, Renren, and other OSNs

Welcome. My group has been studying online social networks since 2008. Not only do these services count more than a billion users, but they are also vibrant application platforms for everything from shared Netflix reviews to Hugg and social marketplaces. On these online social networks, we are primarily interested in topics of security and privacy, measurement, data analysis and modeling.

In general, there are too many new projects and publications to describe, but I'll summarize some of it here. Our recent projects include study of Sybil accounts and user profile browsing behavior (latent interactions) on the Renren network (IMC 2011 and IMC 2010), a study of phishing and spam campaigns on Facebook (IMC 2010), and distributed systems that increases the scalability of Twitter (Middleware 2011) and the performance of Facebook (CoNEXT 2010). Other recent projects include differentially private graph models, and an empirical evaluation of static graph models using Facebook graphs, a major measurement study of Facebook (10+million user profiles), and a new framework for privacy-preserving location-based social applications. These projects are described in papers at IMC'11, WWW'10, EuroSys'09, and HotMobile'10 respectively.

In other recent projects, SpikeStrip provides a light-weight server-side mechanism to protect OSNs against malicious data crawlers, and Orion is a "graph coordinate system" that allows us to compute estimates of node-distances in large graphs in constant time, i.e. O(1). Results of both projects are described in papers at WOSN'10.

We have also worked on a number of other social networking topics, including a measurement of the socially-enhanced Overstock.com auction site, and a system for using social networks to protect anonymous communication users from passive traffic analysis attacks.

Related Publications:

• Serf and Turf: Crowdturfing for Fun and Profit
  Gang Wang, Christo Wilson, Xiaohan Zhao, Yibo Zhu, Manish Mohanlal, Haitao Zheng and Ben Y. Zhao
  To Appear: Proceedings of the 20th International World Wide Web Conference (WWW), Lyon, France, April 2012.

• Efficient Shortest Paths on Massive Social Graphs
  Xiaohan Zhao, Alessandra Sala, Haitao Zheng and Ben Y. Zhao
  (Invited Paper) Proceedings of 7th International Conference o$ Networking, Applications and Worksharing (CollaborateCom), Orlando, FL, October 2011.

• Scaling Microblogging Services with Divergent Traffic Demands
  Tianyin Xu, Yang Chen, Lei Jiao, Ben Y. Zhao, Pan Hui and Xiaoming Fu
  Proceedings of The ACM/IFIP/USENIX 12th International Middleware Conference (Middleware), Lisboa, Portugal, December 2011.

• Sharing Graphs using Differentially Private Graph Models
  Alessandra Sala, Xiaohan Zhao, Christo Wilson, Haitao Zheng and Ben Y. Zhao
  Proceedings of The 11th ACM SIGCOMM Internet Measurement Conference (IMC), Berlin, Germany, November 2011.

• Uncovering Social Network Sybils in the Wild
  Zhi Yang, Christo Wilson, Xiao Wang, Tingting Gao, Ben Y. Zhao and Yafei Dai
  Proceedings of The 11th ACM SIGCOMM Internet Measurement Conference (IMC), Berlin, Germany, November 2011.

• Invisible Interactions: What Latent Social Interaction Can Tell Us about Social Relationships in Social Networking Sites
  Miriam Metzger, Christo Wilson, Rebekah Pure, and Ben Y. Zhao
  Chapter in Networked sociability and individualism. Technology for personal and professional relationships, Ed. Francesca Comunello. IGI Global Publishing, 2011.

• Privacy, Availability and Economics in the Polaris Mobile Social Network
  Christo Wilson, Troy Steinbauer, Gang Wang, Alessandra Sala, Haitao Zheng and Ben Y. Zhao
  Proceedings of 12th ACM Workshop on Mobile Computing Systems and Applications (HotMobile), Phoenix, AZ, March 2011.

• Exploiting Locality of Interest in Online Social Networks
  Mike P. Wittie, Veljko Pejovic, Lara Deek, Kevin C. Almeroth and Ben Y. Zhao
  Proceedings of The 6th International Conference on emerging Networking EXperiments and Technologies (CoNEXT), Philadelphia, PA, December, 2010.

• Understanding Latent Interactions in Online Social Networks
  Jing Jiang, Christo Wilson, Xiao Wang, Peng Huang, Wenpeng Sha, Yafei Dai and Ben Y. Zhao
  Proceedings of The 10th ACM SIGCOMM International Measurement Conference (IMC), Melbourne, Australia, November 2010.

• Detecting and Characterizing Social Spam Campaigns
  Hongyu Gao, Jun Hu, Christo Wilson, Zhichun Li, Yan Chen and Ben Y. Zhao
  Proceedings of The 10th ACM SIGCOMM International Measurement Conference (IMC), Melbourne, Australia, November 2010.

• Don't Tread on Me: Moderating Access to OSN Data with SpikeStrip
  Christo Wilson, Alessandra Sala, Joseph Bonneau, Robert Zablit and Ben Y. Zhao
  Proceedings of The 3rd Workshop on Online Social Networks (WOSN), Boston MA, June 2010.

• Orion: Shortest Path Estimation for Large Social Graphs
  Xiaohan Zhao, Alessandra Sala, Christo Wilson, Haitao Zheng and Ben Y. Zhao
  Proceedings of The 3rd Workshop on Online Social Networks (WOSN), Boston MA, June 2010.

• Measurement-calibrated Graph Models for Social Network Experiments
  Alessandra Sala, Lili Cao, Christo Wilson, Robert Zablit, Haitao Zheng and Ben Y. Zhao
  Proceedings of the 19th International World Wide Web Conference (WWW), Raleigh, NC, April 2010.

• Preserving Privacy in Location-based Mobile Social Applications,
  Krishna P. N. Puttaswamy and Ben Y. Zhao
  Proceedings of 11th Workshop on Mobile Computing Systems and Applications (ACM HotMobile), Annapolis, MD, February 2010.

• StarClique: Guaranteeing User Privacy in Social Networks Against Intersection Attacks,
  Krishna P. N. Puttaswamy, Alessandra Sala, and Ben Y. Zhao
  Proceedings of ACM Conference on emerging Networking EXperiments and Technologies (CoNEXT), Rome, Italy, December 2009.

• User Interactions in Social Networks and their Implications,
  Christo Wilson, Bryce Boe, Alessandra Sala, Krishna P. N. Puttaswamy and Ben Y. Zhao
  Proceedings of ACM EuroSys, Nuremburg, Germany, April 2009.

• Improving Anonymity using Social Links,
  Krishna P. N. Puttaswamy, Alessandra Sala and Ben Y. Zhao
  Proceedings of IEEE Workshop on Secure Network Protocols (NPSec 2008), Orlando, FL, October 2008.

• Do Social Networks Improve e-Commerce: a Study on Social Marketplaces,
  Gayatri Swamynathan, Christo Wilson, Bryce Boe, Kevin C. Almeroth and Ben Y. Zhao.
  Proceedings of First SIGCOMM Workshop on Online Social Networks (WOSN 2008), Seattle, WA, August 2008.

• Deployment of a Large-scale Peer-to-Peer Social Network,
  Mao Yang, Hua Chen, Ben Y. Zhao, Yafei Dai and Zheng Zhang,
  Proceedings of the 1st Workshop on Real Large Distributed Systems (WORLDS), San Francisco, CA, December, 2004.

Graph datasets, Graph models, and Source Code:

• Social Models, WWW'10
  We are working on releasing source code for calibrating current graph models to measurement graphs and generators to produce synthetic graphs. We have already released model parameters for many of our recent measured social graphs from Facebook and Renren.com. Interested researchers can use these parameters and our graph generators to produce "realistic" synthetic graphs with similar graph properties to our measured datasets.

  Code, parameters, and instructions on generating synthetic graphs is here.

• EuroSys '09 Datasets
  Following our EuroSys Facebook measurement study, we are making some datasets of social graphs and interaction graphs available.

  These graphs only contain simple edges connecting anonymized nodeIDs. The social graph file is simply a list of all edges in the graph, each bidirectional edge represented by a two-tuple of anonymized nodeIDs. Our user connectivity graphs reflect measurements performed in early 2008, and are not reflective of current Facebook topologies.

  For the anonymized interaction graphs, we filter interactions based on their relative age to the time of the crawl (April 2008). Each edge in the interaction graph is listed in the file as a two-tuple of anonymized nodeIDs. The interaction graph is an undirected graph, so an edge from A to B represents a bidirectional edge connecting them. We include multiple interactions within the same period as duplicate edges across the same endpoints to account for user pairs that interact more than once during the time period. This frequency can be used to assign "weights" to edges on the interaction graph. If you want an undirected, unweighted interaction graph, then remove those duplicate edges.

  Note: If you would like access to this data, please send email to ravenben at cs dot ucsb dot edu. When you get access to the data files, please do not distribute them beyond your immediate research group. Thank you.

  • Anonymized social graphs
    • Anonymous regional network A: 3,097,166 users, 28,377,481 edges, 605MB GZipped
    • Anonymous regional network B: 2,937,614 users, 24,236,701 edges, 521MB GZipped
  • Anonymized interaction graphs
    • Anonymous regional network A, n=1, t=1 month: 1,412,252 interactions, 22MB GZipped
    • Anonymous regional network A, n=1, t=1/2 year: 7,483,904 interactions, 87MB GZipped
    • Anonymous regional network A, n=1, t=1 year: 16,889,111 interactions, 159MB GZipped
    • Anonymous regional network A, n=1, t=lifetime: 17,644,327 interactions, 164MB GZipped