[April 14, 2010]


Speaker: Krzysztof Pietrzak, CWI
Title: Tamper proof (double feature)

Abstract: I will present two approaches which aim to protect
cryptographic hardware against "tampering attacks". In such attacks,
the memory or computing circuit is tampered, in the hopes that future
interaction with the modified system will reveal some secret
information.

I'll first present the notion of "non-malleable codes" and how they
provide an algorithmic solution to protect against tampering attacks
on the memory (this is based on joint work with Stefan Dziembowski and
Daniel Wichs, ICS 2010)

The general problem where the adversary can tamper with the entire
circuit (i.e. set/flip the values on every wire) is harder to get
by. I'll show a transformation that makes any circuit tamper-proof,
even if the adversary can tamper with every wire in the entire
circuit, as long as each tampering attempt fails with some probability
d>0. Earlier work ("private circuits II" by Ishai et al.) considered
the case d=0, but where one only could tamper with a fixed number of
wires between every two invocations. 
(this is joint work with Sebastian Faust and Daniele Venturi)



The talk is funded by a grant from Walmart