5ad18935c0ffff94.tex
1: \begin{abstract}
2: The semidirect discrete logarithm problem (SDLP) is the
3: following analogue of the standard discrete logarithm problem in the semidirect product semigroup $G\rtimes \End(G)$ 
4: for a finite semigroup $G$. 
5: Given $g\in G, \sigma\in \End(G)$, 
6: and $h=\prod_{i=0}^{t-1}\sigma^i(g)$ for some integer $t$, the SDLP$(G,\sigma)$, for $g$ and $h$, asks to determine
7: $t$. As Shor's algorithm crucially depends on commutativity,
8: it is believed
9: not to be applicable to the SDLP. 
10: Previously, the best known algorithm for the SDLP was
11:  based on Kuperberg's subexponential time quantum algorithm. 
12: Still, the problem plays a central role in  the security of 
13: certain proposed cryptosystems in the family of \textit{semidirect product key exchange}. This includes
14: a recently proposed signature protocol
15: called SPDH-Sign.  
16: In this paper, we show that the SDLP is even easier 
17: in some important special cases. Specifically, for a finite group $G$, we
18: describe quantum algorithms for the SDLP in 
19: $G\rtimes \Aut(G)$ for the following two classes of instances:  
20: the first one is when $G$ is solvable and the second is when
21: $G$ is a matrix group and a power of
22: $\sigma$ with a polynomially small exponent is an inner automorphism of $G$.
23: We further extend the results to groups composed of factors
24: from these classes. 
25: A consequence is that SPDH-Sign and similar cryptosystems 
26: whose security assumption is based on the presumed hardness of the SDLP 
27: in the cases described above are insecure against quantum attacks.
28: The quantum ingredients we rely on
29: %of the algorithms in this paper
30: are not new: these are
31: Shor's factoring and discrete logarithm algorithms
32: and well-known generalizations.
33: \end{abstract}
34: