Cryptography Reference
In-Depth Information
1
u 2 , and handle the
(see Exercise 11.4.12 )from
A
to the variety F ( x 1 ) F ( x 2 ) F ( x 3 )
=
characteristic 2 and 3 cases.
x 3
Exercise
11.4.12
(Shallue
and
van
de
Woestijne
[ 491 ])
Let F ( x )
=
+
Ax
+
B
u 2
v 2
u 2
and H ( u,v )
=
+
uv
+
+
A ( u
+
v )
+
B .Let V : F ( x 1 ) F ( x 2 ) F ( x 3 )
=
and let
S : y 2 H ( u,v )
y 2 ,F ( u
=−
F ( u ). Show that the map ψ ( u,v,y )
( v,
A
u
v,u
+
+
y 2 ) H ( u,v ) /y ) is a rational map from S to V .Let p> 3beprime.Fix u
∈ F p such that
=
0 and 3 u 2
+
+
A 2
=
F ( u )
2 Au
4 B
0. Show that the surface S for this fixed value of
u is
A/ 2)] 2
[3 u 2 / 4
A 2 / 4] y 2
[ y ( v
+
u/ 2
+
+
+
Au/ 2
+
B
=−
F ( u ) .
1
1
Hence, show there is a rational map from
A
to S and hence a rational map from
A
to V .
1
C when C is a curve
of genus at least 1. This follows from the Hurwitz genus formula: if the map has degree
d then we have
It is worth noting that there can be no rational map φ :
P
1 )
0 where R is a positive integer
counting the ramification, which is a contradiction. The above maps do not contradict this
fact. They are not rational maps from
2
=
2 g (
P
2
=
d (2 g ( C )
2)
+
R
1 ) to an elliptic curve; there is always one
part of the function (such as computing a square-root or cube-root) that is not a rational
map.
Icart [ 273 ] has given a simpler map for elliptic curves y 2
P
1
A
(or
x 3
=
+
Ax
+
B over
F q when
q
2(mod3).Let u
∈ F q . Define
= v 2
u 6 / 27 1 / 3
u 4 ) / (6 u ) ,x
u 2 / 3
v
=
(3 A
B
+
and y
=
ux
+
v (11.4)
where
the
cube
root
is
computed
by
exponentiating
to
the
power
(2 q
1) / 3
3 1
(mod ( q
1)).
Exercise 11.4.13 Verify that the point ( x,y ) of equation ( 11.4 ) is a point on E : y 2
=
x 3
+
Ax
+
B over
F q . Show that, given a point ( x,y ) on an elliptic curve E over
F q as above,
one can efficiently compute u
∈ F q , if it exists, such that the process of equation ( 11.4 )
gives the point ( x,y ).
In most cryptographic applications, we are interested in sampling from subgroups of
E (
F q ) of prime order r . As mentioned earlier, the simplest way to transform elements
sampled randomly in E (
F q ) into random elements of the subgroup is to exponentiate to the
power # E (
F q ) /r (assuming that r
# E (
F q )).
11.4.3 Hashing to algebraic groups
Recall that sampling from algebraic groups is the task of selecting group elements uniformly
at random. On the other hand, a hash function H :
l
{
0 , 1
}
G (
F q ) is a deterministic
l and outputs a group element. It is required that the
output distribution of H , corresponding to the uniform distribution of the message space, is
algorithm that takes an input m
∈{
0 , 1
}
Search WWH ::




Custom Search