Deploying secure multi-party computation for financial data analysis

Dan Bogdanov; Riivo Talviste; Jan Willemson

Paper 2011/662

Deploying secure multi-party computation for financial data analysis

Dan Bogdanov, Riivo Talviste, and Jan Willemson

Abstract

In this paper we describe a secure system for jointly collecting and analyzing financial data for a consortium of ICT companies. To guarantee each participant's privacy, we use secret sharing and secure multi-party computation (MPC) techniques. While MPC has been used to solve real-life problems beforehand, this is the first time where the actual MPC computation was done over the internet with computing nodes spread geographically apart. We describe the system architecture, security considerations and implementation details. We also present the user feedback analysis revealing that secure multi-party computation techniques give sufficient assurance for data donors to submit their sensitive information, and act as a critical enabling feature for privacy-preserving data mining.

Note: Edit: added reference to the work of J. Feigenbaum et al.

Metadata

Available format(s): PDF
Category: Applications
Publication info: Published elsewhere. This is an extended version of the paper presented at Financial Cryptography and Data Security 2012.
Keywords: financial data analysis privacy-preserving data mining secure multi-party computation
Contact author(s): riivo talviste @ cyber ee
History: 2011-12-16: revised; 2011-12-09: received; See all versions
Short URL: https://ia.cr/2011/662
License: CC BY

BibTeX

@misc{cryptoeprint:2011/662,
      author = {Dan Bogdanov and Riivo Talviste and Jan Willemson},
      title = {Deploying secure multi-party computation for financial data analysis},
      howpublished = {Cryptology {ePrint} Archive, Paper 2011/662},
      year = {2011},
      url = {https://eprint.iacr.org/2011/662}
}