|
|
|
|
|
Multiparty Project |
Keywords: multiparty computation, confidentiality, configurable security, secure list comparison, malicious behavior, quasi-commutative cryptography, communication protocols
Citation:
Abstract
Standard multi-party computation models assume semihonest behavior, where the majority of participants implement protocols according to specification, an assumption not always plausible. In this paper we introduce a multiparty protocol for collaborative data analysis when participants are malicious and fail to follow specification. The protocol incorporates a semi-trusted third party, which analyzes encrypted data and provides honest responses that only intended recipients can successfully decrypt. The protocol incorporates data confidentiality by enabling participants to receive encrypted responses tailored to their own
encrypted data submissions without revealing plaintext to other participants, including the third party. As opposed to previous models, trust need only be placed on a single participant
with no data at stake. Additionally, the proposed protocol is configurable in a way that security features are controlled by independent subprotocols. Various combinations
of subprotocols allow for a flexible security system, appropriate for a number of distributed data applications, such as secure list comparison.
B. Malin, E. Airoldi, S. Edoho-Eket, and Y. Li.
Configurable Security Protocols for Multi-Party Data Analysis with Malicious Participants. In Proceedings of the 21st International Conference on Data Engineering. Tokyo, Japan. 2005: Forthcoming. [Original version available under the same title as Carnegie Mellon University, School of Computer Science, Technical Report CMU-ISRI-04-132. Pittsburgh: Sept 2004. (PDF), (PS)].
Related Links