TY - GEN
T1 - FriendGuard
T2 - 24th ACM Symposium on Access Control Models and Technologies, SACMAT 2019
AU - Morris, Joshua
AU - Lin, Dan
AU - Squicciarini, Anna
N1 - Funding Information:
This work is partially supported by National Science Foundation under the project DGE-1914771. Work from Dr Squicciarini was partially funded by National Science Foundation under grant 1453080.
Publisher Copyright:
© 2019 Association for Computing Machinery.
PY - 2019/5/28
Y1 - 2019/5/28
N2 - With the prevalence of online social networking, a large amount of studies have focused on online users' privacy. Existing work has heavily focused on preventing unauthorized access of one's personal information (e.g. locations, posts and photos). Very little research has been devoted into protecting the friend search engine, a service that allows people to explore others' friend lists. Although most friend search engines only disclose a partial view of one's friend list (e.g., k friends) or offer the ability to show all or no friends, attackers may leverage the combined knowledge from views obtained from different queries to gain a much larger social network of a targeted victim, potentially revealing sensitive information of a victim. In this paper, we propose a new friend search engine, namely FriendGuard, which guarantees the degree of friend exposure as set by users. If a user only allows k of his/her friends to be disclosed, our search engine will ensure that any attempts of discovering more friends of this user through querying the user's other friends will be a failure. The key idea underlying our search engine is the construction of a unique sub social network that is capable of satisfying query needs as well as controlling the degree of friend exposure. We have carried out an extensive experimental study and the results demonstrate both efficiency and effectiveness in our approach.
AB - With the prevalence of online social networking, a large amount of studies have focused on online users' privacy. Existing work has heavily focused on preventing unauthorized access of one's personal information (e.g. locations, posts and photos). Very little research has been devoted into protecting the friend search engine, a service that allows people to explore others' friend lists. Although most friend search engines only disclose a partial view of one's friend list (e.g., k friends) or offer the ability to show all or no friends, attackers may leverage the combined knowledge from views obtained from different queries to gain a much larger social network of a targeted victim, potentially revealing sensitive information of a victim. In this paper, we propose a new friend search engine, namely FriendGuard, which guarantees the degree of friend exposure as set by users. If a user only allows k of his/her friends to be disclosed, our search engine will ensure that any attempts of discovering more friends of this user through querying the user's other friends will be a failure. The key idea underlying our search engine is the construction of a unique sub social network that is capable of satisfying query needs as well as controlling the degree of friend exposure. We have carried out an extensive experimental study and the results demonstrate both efficiency and effectiveness in our approach.
UR - http://www.scopus.com/inward/record.url?scp=85067184297&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85067184297&partnerID=8YFLogxK
U2 - 10.1145/3322431.3325103
DO - 10.1145/3322431.3325103
M3 - Conference contribution
AN - SCOPUS:85067184297
T3 - Proceedings of ACM Symposium on Access Control Models and Technologies, SACMAT
SP - 37
EP - 48
BT - SACMAT 2019 - Proceedings of the 24th ACM Symposium on Access Control Models and Technologies
PB - Association for Computing Machinery
Y2 - 3 June 2019 through 6 June 2019
ER -