[Resource Topic] 2020/599: Private Matching for Compute

Welcome to the resource topic for 2020/599

Title:
Private Matching for Compute

Authors: Prasad Buddhavarapu, Andrew Knox, Payman Mohassel, Shubho Sengupta, Erik Taubeneck, Vlad Vlaskin

Abstract:

We revisit the problem of two-party private set intersection for aggregate computation which we refer to as private matching for compute. In this problem, two parties want to perform various downstream computation on the intersection of their two datasets according to a previously agreed-upon identifier. We observe that prior solutions to this problem have important limitations. For example, any change or update to the records in either party’s dataset triggers a rerun of the private matching component; and it is not clear how to support a streaming arrival of one party’s set in small batches without revealing the match rate for each individual batch. We introduce two new formulations of the private matching for compute problem meeting these requirements, called private-ID and streaming private secret shared set intersection, and design new DDH-based constructions for both. Our implementation shows that when taking advantage of the inherent parallelizability of these solutions, we can execute the matching for datasets of size upto 100 million records within an hour.

ePrint: https://eprint.iacr.org/2020/599

See all topics related to this paper.

Feel free to post resources that are related to this paper below.

Example resources include: implementations, explanation materials, talks, slides, links to previous discussions on other websites.

For more information, see the rules for Resource Topics .