[viff-devel] Comparison for two parties

[Martin Geisler]

Hi Jakob and hi everybody else,

At a CACE/SIMAP meeting last Friday we talked about wether or not the
comparison would work with the two-party runtime we have in VIFF.

If we look at the code in ComparisonToft07Mixin, then we see that it
calls these methods:

self.prss_share: not present in PaillierRuntime, but the share method
  there should be equivalent.

self.prss_share_random with binary=True: the prss_share_random method
  currently implemented in PaillierRuntime does not take the binary
  argument. But from looking at the prss_share_random code in
  PassiveRuntime it seems that it should also work in PaillierRuntime.

  It might be possible to make an easier and simpler implementation by
  directly building additive shares.

a ^ b: this would work if PaillierRuntime implemented the xor method.

We might want to move the PRSS related code out of PassiveRuntime and
into its own mixin class.

If we end up using the PRSS code for the additive shares in
PaillierRuntime, then it is interesting to note that the code formally
generates Shamir shares: if the code generated the Shamir shares (1,a)
and (2,b) for t=1, then the secret is 2a - b.

But if we use a and b directly as additive shares this simply
corresponds to using a + b = (s + 2b)/2 as the secret instead of s. I
think this is fine since b is random.

-- 
Martin Geisler

VIFF (Virtual Ideal Functionality Framework) brings easy and efficient
SMPC (Secure Multiparty Computation) to Python. See: http://viff.dk/.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
URL: <http://lists.viff.dk/pipermail/viff-devel-viff.dk/attachments/20081130/c42cf56c/attachment.pgp>