Cryptography Reference
In-Depth Information
texts of MPKCs are decrypted using secret keys S, T and G by equation (4),
and signatures are generated using S, T, G and random ephemeral values r =
( r 1 ,
,r u ) by equation (5). We deal with the following two attacks on MPKCs:
(Fault attacks on G ) the attacker causes a fault to change a coecient of the
central map G ,
(Fault attacks on r ) the attacker causes a fault such that several random
ephemeral values of r are fixed to the same values.
···
3.2
G
Assume that the coecient α ij in (6) for a BF type or a ( l )
ij
Fault Attacks on
in (10) for an STS type
or ( a ( l )
or ( a ( l )
is changed to α ij
ij ) by the fault, and α ij
ij ) cannot be recovered
or a ( l )
ij
to α ij
again. For the correct (not faulty) central map G and the public
S from equation (3), denote by G
key F = T
G
the faulty central map,
F := T
G
S and
ΔF := F
( G
F = T 1
G )
S,
where T 1 is given in (2). The basic approach of our fault attack is as follows.
——————————————————————————————————
Step 1. Cause a fault on G and make G .
Step 2. Decrypt randomly chosen messages y (1) ,
,y ( N )
k m ( N
···
1) by the
faulty map G using equation (4);
x ( l ) := S 1 ( G 1 ( T 1 ( y ( l ) ))) .
Step 3. Encrypt x (1) ,
,x ( N )
k n by the correct (not faulty) public key F in
···
equation (3);
z ( l ) := F ( x ( l ) ) .
Step 4. Put
z ( l ) .
Find (partial information of) S and T by the pairs
δ ( l ) := y ( l )
( x ( l ) ( l ) )
} 1 δ(l))}1≤l≤N .
——————————————————————————————————
Since y ( l ) = F ( x ( l ) )and z ( l ) := F ( x ( l ) ), we have
{
δ ( l ) = ΔF ( x ( l ) )= T 1
( G
S ( x ( l ) ) .
G )
Many coecients of both G and G are the same, so then G
G is a sparse
G , we try to recover S and T .The
details of recovering S and T are described in the following subsections.
polynomial. From the sparseness of G
3.2.1 Fault Attack on G for Big Field Type
We propose the fault attack on HFE, which is a typical BF-type model. Let
,Δf m ( x )) t
ΔF ( x )=( Δf 1 ( x ) ,
···
 
Search WWH ::




Custom Search