Method of Coding a Secret Formed by a Numerical Value

Abstract

A method of coding a secret, a numerical value d, subdivided into a number N of secret elements [di]n1, a composition law () applied to the elements di giving the value d. The following are calculated: (A) a first image (TN) of the secret by iterative calculation and application of the law () between the first image Ti−1 of rank i−1 and of the product according to this law of the element (di) of next rank and of a random value (Ri) of a first set, (B) a first numerical value (S1) by application of the law () to the N random values (Ri), (C) a second numerical value (S2) by application of the law to the N−1 random values (Aj) of a second set, (D) a second image T′ of the secret by application of the inverse law () to the first image (TN) and to the second numerical value (S2) so as to generate an intermediate image (Tx) and then application of the inverse law to the intermediate image (Tx) and to the second numerical value (S2). The random value (Aj) of corresponding rank is allocated (E) to the first N−1 elements (di) and the value of the second image (T′) is allocated (E) to the last element (dN). Application to secret or public key cryptography processes.

Description

The invention relates to a method of coding a secret, formed by a numerical value.

The considerable upsurge in communications by transmission of electronic messages has very quickly raised the problem of the confidentiality of the data transmitted.

Very sophisticated solutions for enciphering/deciphering these data have been proposed by means of encipherment algorithms using a single secret key, serving for encipherment/decipherment, and then a public key, with which is associated a private key, used for deciphering the data, enciphered by means of the public key.

The aforementioned solutions are satisfactory, to the extent that secret-key algorithms are hard to break, at the very least if the secret key is not compromised, and that public-key/private-key algorithms do not entail limitations as regards the dissemination of the public key and require the implementation of hardware and software means of prohibitive complexity and calculation cost with a view either to breaking the encipherment/decipherment algorithm or to retrieving the value of the private key associated with the public key.

In all events, use of a cryptographic scheme with a single secret key or a public key, with which a private key is associated, it is indispensable to prevent any compromising of the secret key or of the private key, so as to guarantee the confidentiality of the data transmitted.

Whilst satisfactory protected-access cryptographic components have been proposed, in particular integrated in the form of security processors for the cryptographic components of electronic cards, termed chip cards, exterior access by way of the input/ouput port to the security components of these electronic cards possibly being made extremely difficult or indeed practically impossible, the read/write manipulation of the secret or private keys by these components may allow the compromising of the aforementioned keys, in particular of the secret values or secrets entering into the definition of these values.

This compromising may occur through “hidden channel” attack (known as Side Channel attack), this type of attack possibly consisting for example in detecting the intensity variations of the electric current consumed by the security component or the card in the course of these manipulations, these being necessary.

The object of the present invention is to remedy the drawbacks of the risks of hidden channel attack, through the implementation of a method of coding a secret, subdivided into several mutually uncorrelated secret elements, it not being possible for the manipulation of the secret elements to enable the original secret to be retrieved, although each secret element might, as the case may be, be compromised.

The method of coding a secret formed by a numerical value, in a secret-key or public-key cryptography scheme in which this secret is divided between a plurality of a determined number of elements, of which a composition law represents the value of this secret, which is the subject of the present invention, is noteworthy in that it consists, furthermore, in recalculating a new plurality of elements of the secret without ever manipulating this secret. For this purpose it is necessary to calculate a first image of this secret by iterative application of the composition law term by term between the first image of previous rank and the composition product according to this composition law of the element of next rank and a random value of the same rank, chosen from among a first set of one and the same plurality of random values; calculate a first numerical value, the composition product of this composition law applied successively to the random values of this first set of random values; calculate a second numerical value, the composition product according to this composition law applied successively to the random values of a second set of one and the same plurality minus one of random values; calculate a second image of this secret by applying the inverse composition law to the first image of this secret and to this second numerical value, so as to produce an intermediate image of this secret, and then by applying this inverse composition law to this intermediate image and to this first numerical value, so as to produce this second image of said secret; and allocate to each of these successive elements minus the last of this plurality of elements the random value of corresponding rank of this second set of at least one random value and to the last element the numerical value of this second image.

The method of coding a secret, which is the subject of the present invention, will be better understood on reading the description and on observing the drawings hereinafter in which:

FIG. 1 represents, by way of illustration, a general flowchart for implementing the constituent steps of the method which is the subject of the invention;

FIG. 2 a represents, by way of illustration, a first and a second composition law applicable to numerical values and allowing the implementation of the method which is the subject of the present invention;

FIG. 2 b represents, by way of illustration, a specific flowchart for implementing the method which is the subject of the invention, when the composition law represented in FIG. 2a is an exclusive OR operation;

FIG. 2 c represents by way of illustration a specific flowchart for implementing the method which is the subject of the invention, when the composition law represented in FIG. 2a is an addition operation;

FIG. 3 is a functional diagram of a security component of a cryptographic device specially adapted for the implementation of the method which is the subject of the invention.

A more detailed description of the method of coding a secret, in accordance with the subject of the present invention, will now be given in conjunction with FIG. 1.

In a general way, it is recalled that the subject of the method, which is the subject of the invention, is the coding of a secret s formed by a numerical value d, in a secret-key or public-key cryptography scheme. It applies more particularly to any cryptographic calculation process in which the secret s is subdivided into a plurality of a determined number of secret elements, each denoted d_i, which plurality of elements is dubbed hereinafter [d_i]₁^N of which a composition law denoted represents the numerical value of the secret s.

With reference to FIG. 1, the secret s and the numerical value d representing the latter satisfy relation (1):

$\begin{array}{c}{s=\underset{\_}{d};\left[d_i\right]}_1^N\\ \underset{\_}{d}=\underset{i=1}{\overset{\otimes N}{\pi }}d_i\end{array}$

In this relation

$\underset{i=1}{\overset{\otimes N}{\pi }}d_i$

represents the composition product of the composition law applied to the set of N elements d_i.

As represented in FIG. 1, the method which is the subject of the invention consists in calculating, in a step A, a first image of the secret s by iterative application of the composition law term by term between the first image of previous rank, denoted T_i−1, and the composition product, according to this composition law, of the element of next rank i, denoted d_i, and a random value, denoted R_i, chosen from among a first set of one and the same plurality of random values.

In step A of FIG. 1 the first set of one and the same plurality of random values is denoted [R_i]₁^N.

With reference to step A of FIG. 1, the operation of calculating the first image T_N satisfies relation (2):

[T_i=T_i−1(d_iR_i)]₁^N→T_N

In the above relation,

• • T_i denotes the first current image of rank i;
  • T_i−1 denotes the first previous image of rank i−1;
  • d_i denotes the current element of rank i;
  • R_i denotes the random value of rank i of the first set of random values;
  • T_N denotes the first image obtained after iterative calculation.

Step A of FIG. 1 is followed by a step B consisting in calculating a first numerical value, denoted S₁, the composition product of the same aforementioned composition law applied successively to the random values of the previously mentioned first set of random values.

In step B of FIG. 1, the first numerical value S₁ satisfies relation (3):

$S_1=\underset{i=1}{\overset{\otimes N}{\pi }}R_i$

Step B of FIG. 1 is followed by a step C consisting in calculating a second numerical value, denoted S₂, the composition product, according to the same aforementioned composition law, applied successively to the random values of a second set of one and the same plurality minus one of random values.

Consequently, the second set of one and the same plurality minus one of random values is denoted [A_j]₁^N−1

The second numerical value satisfies relation (4):

$S_2=\underset{j=1}{\overset{\otimes N-1}{\pi }}A_j$

Step C of FIG. 1 is then followed by a step D consisting in calculating a second image of the secret, denoted T′.

With reference to step D of FIG. 1, it is indicated that the aforementioned second image T′ is calculated by applying the inverse composition law applied to the first image of the secret T_N and to this second numerical value S₂, so as to produce an intermediate image denoted T_x, and then by applying this same inverse composition law applied to the intermediate image T_x and to the first numerical value S₁, so as to produce the second image of the secret, denoted T′. The inverse composition law is denoted .

In step D of FIG. 1, the calculation of the second image T′ satisfies relation (5):

T_x=T_NS₂

T′=T_xS₁

Step D of FIG. 1 is then followed by a step E consisting in allocating to each of the successive elements of the plurality of elements [d_i]₁^N, minus the last, the random value of rank corresponding to the value of the second set of at least one random value, the set denoted [A_j]₁^N−1, and in allocating to the last element the numerical value of the aforementioned second image T′.

Consequently the step of allocations represented in step E satisfies relation (6):

{[d_i]₁^N−1=[A_j]₁^N−1

{d_N=T′

A more detailed description of a first and of a second variant for implementing the method which is the subject of the invention will now be given in conjunction with FIG. 2a and FIGS. 2b and 2c, respectively.

In a general way, it is indicated that the composition law mentioned previously is formed by a distributive arithmetic or logic operation, endowed with a neutral element. A corresponding composition law can thus be applied to any secret and to any element of a secret formed by a numerical value consisting either of an integer or of a real number.

Thus, under this assumption, for a secret s formed by a numerical value d of determined length L, each random value R_i of the first respectively A_j of the second set of random values is chosen of length less than 2^L−N+1.

By way of nonlimiting example, the aforementioned composition law can consist, as represented in FIG. 2a, of an exclusive OR operation for example. It can furthermore consist of an arithmetic operation such as addition.

It is noted, of course, that the aforementioned composition law is then endowed with an inverse operation, the exclusive OR operation unchanged, when the exclusive OR operation constitutes the aforementioned composition law, respectively the subtraction operation, when the addition operation constitutes the abovementioned composition law.

The previously mentioned composition laws and their corresponding operation are represented in the drawing of FIG. 2a, illustrated by relation (7):

=⊕; =⊕

=+;=−

In the above relation,

⊕ represents the exclusive OR operation, conducted bitwise on the integers or real numbers constituting the secret elements or secret, as well as the random numbers;

+ and − represent the addition operation and the inverse operation of subtraction for the composition law formed by arithmetic addition. Furthermore the neutral element is 0 for both operations.

A specific mode of implementation of the method which is the subject of the invention is now described in conjunction with FIG. 2b, in the case of the nonlimiting implementation of a composition law formed by the exclusive OR operation.

In step A of FIG. 2b, the operation of calculating the first image T_N is given by relation (8):

[T_i=T_i−1⊕(d_i⊕R_i)]_i=1^i=N→T_N

In step B of FIG. 2b, the operation of calculating the first numerical value is given by relation (9)

$S_1=\underset{i=1}{\overset{\oplus N}{\pi }}R_i;$

In step C of FIG. 2b, the operation of calculating the second numerical value is given by relation (10):

$S_2=\underset{j=1}{\overset{\oplus N-1}{\pi A_j}}$

In step D of FIG. 2b, the operation of calculating the second image T′ is given by relation (11):

T_x=T_N⊕S₂

T′=T_x⊕S₁

Finally, the allocating step E is unchanged in relation to the allocating step E of FIG. 1.

Furthermore, by way of nonlimiting example the aforementioned composition law can consist, as represented in FIG. 2c, of an arithmetic addition operation.

In step A of FIG. 2c, the operation of calculating the first image T_N is given by relation (12):

[T_i=T_i−1+(d_i+R_i)]_i=1^i=N→T_N

In step B of FIG. 2c, the operation of calculating the first numerical value is given by relation (13):

$S_1=\sum _{i=1}^NR_i$

In step C of FIG. 2c, the operation of calculating the second numerical value is given by relation (14):

$S_2=\sum _{j=1}^{N-1}A_j$

In step D of FIG. 2c, the operation of calculating the second image T′ is given by relation (15):

T _x =T _N −S ₂

T′=T _x −S ₁

With reference to FIG. 2c, it may be observed that the allocating step E of FIG. 1 is then subdivided into two sub-steps if each element di of the secret must be positive. This sub-step E₀ is a test of comparison of superiority of the second image T′ with the zero value and a sub-step E₁ of allocating proper, which is also unchanged with respect to the allocating step E of FIG. 1.

The object of the test sub-step E₀ is to verify that the second image T′ is significant. The significant character of the second image T′ is obtained when the numerical value representative of the latter is strictly greater than zero.

Thus, upon a positive response to the comparison test of sub-step E₀, the allocating sub-step proper E₁ is called and carried out in the same manner as in the case of FIG. 1 or FIG. 2b.

Conversely, upon a negative response to the test sub-step E₀, the second image T′ then being negative, a return to step A is executed so as to repeat the calculation process until a positive value representing the second image T′ is obtained.

A description of a cryptographic device security component comprising a secure processor, a nonvolatile memory, a work memory, a program memory and a bus with read-write protected access will now be given in conjunction with FIG. 3.

In the aforementioned FIG. 3, the secure microprocessor is denoted μPS, the work memory is denoted RAMS, the program memory is denoted PROGS, the nonvolatile memory is denoted NVS and the internal bus is denoted I/O.

The security component which is the subject of the invention is noteworthy in that the program memory PROGS comprises a computer program including a series of instructions stored in this program memory.

During the execution of these instructions, the secure processor μPS executes the steps for implementing the method of coding a secret formed by the numerical value d in any secret-key or public-key cryptography scheme, as described previously in the description in conjunction with FIGS. 1 to 2b.

Thus, the security processor μPS delivers on the read-write protected access bus denoted I/O solely the secret elements denoted d_i successively, under the supervision of the cryptographic device, not represented in the drawing of FIG. 3.

It is understood, in particular, that the method and the security component which are the subject of the invention operate on any secret formed by a numerical value d constituting totally or partially either a secret key in a secret-key cryptography scheme, or a private key in any public-key cryptography scheme.

Of course the aforementioned method and security component which are the subject of the invention may be implemented for the calculation of any value of access code, for identification with secret intent of an authentication, non-repudiation or signature process.

Claims

1. A method of coding a secret formed by a numerical value d, in a secret-key or public-key cryptography scheme, in which the secret is subdivided into a plurality of a determined number N of elements di, [di]1N, of which a composition law represents the numerical value d of said secret, said method furthermore comprising: calculating a first image TN of said secret by iterative application of the composition law term by term between said first image Ti−1 of rank i−1 and the composition product according to said composition law of the element di of next rank and a random value Ri of rank i, chosen from among a first set of a plurality of N random values [Ti=Ti−1(diRi)]i=1i=N=TN calculating a first numerical value S1, the composition product of said composition law applied successively to said random values Ri of said first set of N random values

2. The method as claimed in claim 1, wherein said composition law is formed by a distributive arithmetic or logic operation, endowed with a neutral element.

3. The method as claimed in claim 2, wherein said logic operation is the bitwise exclusive OR operation.

4. The method as claimed in claim 2, wherein said arithmetic operation is addition, the inverse composition law being formed by subtraction.

5. The method as claimed in claim 1, wherein for a secret formed by a numerical value d of determined length L, each random value Ri of the first respectively Aj of the second set of random values is chosen of length less than 2L−N+1.

6. The method as claimed in claim 4, wherein for an arithmetic operation formed by addition, said method furthermore comprises, prior to said step consisting in allocating, a comparison step for comparing superiority of the numerical value of said second image T′ with the zero value, a positive response to said comparison step being followed by said step consisting in allocating, said comparison step being followed by a step of returning to said step consisting in calculating said first image TN of said secret for iteration of the method, otherwise.

7. The method as claimed in claim 1, wherein said secret formed by a numerical value d is either a secret key in a secret-key cryptography scheme, or a private key in a public-key cryptography scheme or else any value of access code, for identification with secret intent of an authentication, non-repudiation or signature process.

8. A security component of a cryptographic device comprising a secure processor, a nonvolatile memory, a work memory, a program memory and a bus with read-write protected access, wherein said program memory comprises a computer program including a series of instructions stored in said program memory, and wherein, during the execution of said instructions, said secure processor executes the steps of, calculating a first image TN of said secret by iterative application of the composition law term by term between said first image Ti−1 of rank i−1 and the composition product according to said composition law of the element d, of next rank and a random value Ri of rank i, chosen from among a first set of a plurality of N random values [Ti=Ti−1(diRi)]i=1i=N→TN;calculating a first numerical value S1, the composition product of said composition law applied successively to said random values Ri of said first set of N random values

9. The security component as claimed in claim 8, wherein said composition law is formed by a distributive arithmetic or logic operation, endowed with a neutral element.

10. The security component as claimed in claim 9, wherein said logic operation is the bitwise exclusive OR operation.

11. The security component as claimed in claim 9, wherein said arithmetic operation is addition, the inverse composition law being formed by subtraction.

12. The security component as claimed in claim 8, wherein for a secret formed by a numerical value d of determined length L, each random value Ri of the first respectively Aj of the second set of random values is chosen of length less than 2L−N+1.

13. The security component as claimed in claim 11, wherein for an arithmetic operation formed by addition, said method furthermore comprises, prior to said step consisting in allocating, a comparison step for comparing superiority of the numerical value of said second image T′ with the zero value, a positive response to said comparison step being followed by said step consisting in allocating, said comparison step being followed by a step of returning to said step consisting in calculating said first image TN of said secret for iteration of the method, otherwise.

14. The security component as claimed in claim 8, wherein said secret formed by a numerical value d is either a secret key in a secret-key cryptography scheme, or a private key in a public-key cryptography scheme or else any value of access code, for identification with secret intent of an authentication, non-repudiation or signature process.