1
$\begingroup$

I see that a lot of multiparty computation and garbling protocols are implemented for circuits like AES or SHA256. For my project, I would like to garble a circuit that performs some group operations on elliptic curves and RSA groups.

Are there any circuit compiler libraries or garbling libraries that support this? Are there any estimates on the size of a circuit that performs one group operation?

$\endgroup$
4
  • $\begingroup$ Do you just need group operations (i.e. a MPC scheme only supporting group operations would be fine)? $\endgroup$
    – SEJPM
    Mar 30, 2021 at 17:54
  • $\begingroup$ @SEJPM My actual plan is to implement IBE from DDH protocol proposed by Dottling and Garg. Their circuit performs a bunch of group exponentiations. If I can estimate the circuit size for one group operation, then I can estimate the overall circuit size. $\endgroup$
    – satya
    Mar 30, 2021 at 18:01
  • $\begingroup$ I'm pretty sure there was a twitter thread within the last year or so (that I can't find anymore) where Nigel Smart asked a similar question. Also on a relevant note, I just want to place the reminder that MPC circuits can be arithmetic or binary and the former should easily support just describing ECC operations as you'd in software (especially when choosing the MPC field prime to be the ECC prime). $\endgroup$
    – SEJPM
    Mar 30, 2021 at 21:16
  • $\begingroup$ This paper arxiv.org/pdf/1706.03370.pdf was discussed in the twitter thread (which I also can't find, only remember the paper). You need 9.2 billion gates or 366,210.94 MB for an exponentiation (curve multiplication) for secp192k1. $\endgroup$
    – lamba
    Aug 13, 2021 at 14:10

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.