Martin Geisler
2008-06-29 12:12:00 UTC
Repost of the Paillier based two player runtime.
The runtime can now make random shares with no communication and can
thus be used by benchmark.py.
When I run both players on my own machine it takes about *30 seconds*
to make 10 multiplications... That is about 1000 times slower than
normal Shamir based multiplications with three players.
The mail by Claudio which describes the protocol:
http://article.gmane.org/gmane.comp.cryptography.viff.devel/290
I would still like comments on the code and on how to better deal with
the growing number of runtimes, operations, etc... It seems to me that
we will end up needing a kind of plugin system.
The runtime can now make random shares with no communication and can
thus be used by benchmark.py.
When I run both players on my own machine it takes about *30 seconds*
to make 10 multiplications... That is about 1000 times slower than
normal Shamir based multiplications with three players.
The mail by Claudio which describes the protocol:
http://article.gmane.org/gmane.comp.cryptography.viff.devel/290
I would still like comments on the code and on how to better deal with
the growing number of runtimes, operations, etc... It seems to me that
we will end up needing a kind of plugin system.
--
Martin Geisler
Martin Geisler