data61 / MP-SPDZ

Versatile framework for multi-party computation
Other
921 stars 278 forks source link

OT in Yao protocol #1176

Closed LBZMT23 closed 1 year ago

LBZMT23 commented 1 year ago

Excuse me, may I bother you for a moment? I have a question. Could you please tell me which paper the OT used in the Yao protocol comes from?

Thank you in advance

mkskeller commented 1 year ago

It uses the following OT extension: https://eprint.iacr.org/2022/192 The extension in turn uses the following for the so-called base OT: https://eprint.iacr.org/2015/267

LBZMT23 commented 1 year ago

It uses the following OT extension: https://eprint.iacr.org/2022/192 The extension in turn uses the following for the so-called base OT: https://eprint.iacr.org/2015/267

Thank you Does the use of SoftSpokenOT mean that the Yao protocol(yao-party.x) in MP-SPDZ is malicious security? What should I do if I want to use semi-honest Yao? Thank you in advance

mkskeller commented 1 year ago

No, just the OT extension supports malicious security. The core Yao protocol is semi-honest.

LBZMT23 commented 1 year ago

No, just the OT extension supports malicious security. The core Yao protocol is semi-honest.

Thank you Does this imply that the overall protocol is semi-honest even though a malicious security OT is used in the Yao protocol? Thank you in advance

mkskeller commented 1 year ago

Yes. When combining protocols, the security of the combination cannot be better than the weakest component.

LBZMT23 commented 1 year ago

Yes. When combining protocols, the security of the combination cannot be better than the weakest component.

Thank you so much! Have a great day!