Wang H Z, et al. Sci China Inf Sci June 2011 Vol. 54 No. 6 1163
Table 1 List of symbols
Symbol Meaning
F
q
a degree k extension of the field {0, 1}, where q = 2
k
.
F
n
q
n-dimensional vectorspace over F
q
.
F
q
n
a degree n extension of the field F
q
.
H(·) a standard hash function such as SHA-1.
H
k
(·) an operation extracting the first k bits of H(·) and mapping the bitstring into an element in F
q
.
a||b concatenation of variables a and b.
δ the number of extended input variables of public key, where 0 6 δ < n.
µ the number of increased equations of the central map for the “encrypt” operation (0 6 µ < δ), and
the number of deleted equations of the central map for the “sign” operation (−δ < µ 6 0).
systems such as McEliece cryptosystem [7]. Encryption and decryption processes are shown in Figure 2.
2.3 A survey of MPKCs
The security of MPKC schemes essentially depends on the two hard computational problems. One is based
on solving a set of randomly chosen nonlinear multivariate polynomial equations over a finite field (called
MQ problem). This problem is known to be NP-hard [8]; that is, given a ciphertext y
1
, . . . , y
m
, finding
a plaintext solution x
1
, · · · , x
n
from quadratic equations of formula (1) is computationally infeasible.
Consequently, randomly chosen quadratic equations over finite field F
q
like (1) is a one-way function.
However, in order to allow legitimate users to easily decrypt the ciphertext, the central map F is not
randomly chosen. Hence, two affine bijective transformations U, T are employed to mask the central map
F as a hard one, i.e., the public key P = T ◦ F ◦ U. Here, P and F are the isomorphism of polynomials
(IP). Clearly, if an attacker can find U and T from P , then the according MPKC is broken. However,
to finish this task requires the attacker to solve the so-called IP-problem which has been shown to be
NP-complete [9].
Generally speaking, MQ-problem and IP-problem cannot completely guarantee the security of MPKCs.
By now, all basic trapdoors (including MIA, HFE, OV and STS) are insecure, and must be modified using
some effective measures for enhancing their security. Wolf [10] summarized ten modifications. In these
methods, only the minus method “–” proposed by Shamir [11] is believed to be secure and efficient. The
idea of minus method is to enhance the security of basic multivariate schemes by removing the last r
coordinates from the public key or its central map. Accordingly, the resulting MPKC as an encryption
scheme runs q
r
times slower than the original scheme.
Apart from Shamir’s minus method, the plus method “+” and the Vinegar method “v” are believed to
be slightly more secure, while the Homogenising method “h” has no effect on the security of the original
scheme. The subfield method “/” and the branching metho d “⊥” are now known to be insecure. As to
the Fixing method “f”, the Sparse p olynomials “s”, the Internal perturbation “i” and the Masking “m”,
it remains unknown whether they have security effect on the original MPKC scheme.
With some modifications and enhancing approaches, dozens of variants have been derived from the four
basic MQ trapdoors. However, most of these variants can only be used for signing messages. Furthermore,
their security has been seriously challenged and many efficient attacks have been found [12]. Hence, there
is a rather pressing need to investigate new MQ-trapdoors or new enhancing methodologies to construct
MPKCs allowing secure encryption and signature. This observation motivates our work in this paper.
3 Extended multivariate public key cryptosystems
The key idea of EMC schemes is to increase a secret transformation L defined in subsection 3.1 in the
outermost layer of the original MQ-trapdoor framework (Figure 1). More precisely, we use the map L
and some modifications to hide the basic MQ-trapdoors into a larger algorithm space; that is, we change
the original public key P : F
n
→ F
n
into P
0
: F
n+δ
→ F
n+µ
(δ > µ). Accordingly, the asymmetry of input