RU2188513C2 - Method for cryptographic conversion of l-bit digital-data input blocks into l-bit output blocks - Google Patents

Abstract

FIELD: electric communications and computer engineering; cryptographic methods and devices for data encryption. SUBSTANCE: method involves division of data block into N≥2 subblocks, generation of at least one binary vector with respect to subblock value at each step, and transformation of subblock depending on binary vector structure; during next step of subblock conversion binary vector is generated depending on binary vector structure at subblock conversion preceding step. In addition subblock is altered by subjecting it to cyclic shift by number of bits equal to current value of binary vector; subblock is also transformed by way of substitution operation corresponding to substitution V-table. EFFECT: enhanced encryption stability in data block conversion and encryption conversion speed. 3 cl, 3 dwg

Description

Technical field
The invention relates to the field of telecommunications and computer technology, and more particularly to the field of cryptographic methods and devices for encrypting messages (information).

State of the art
In the aggregate of the features of the proposed method, the following terms are used:
- the secret key is binary information known only to a legitimate user;
- cryptographic conversion is the conversion of digital information that provides the influence of one bit of the source data on many bits of the output data, for example, to protect information from unauthorized reading, the formation of a digital signature, and the generation of a modification detection code; One of the important types of cryptographic transformations is one-way conversion, hashing and encryption;
- hashing information is some way of generating a so-called hash code, the size of which is fixed (usually 128 bits) for messages of any size; widely used hashing methods based on iterative hash functions using block mechanisms of cryptographic information conversion [see Lai X., Massey JL Hash Functions Based on Block Ciphers / Workshop on the Theory and Applications of Cryptographic Techniques. EUROCRYPT'92, Hungary, May 24-28, 1992, Proceedings, p. 53-66].

- encryption is the process of converting information that depends on the secret key and converts the source text into ciphertext, which is a pseudo-random sequence of characters from which it is practically impossible to obtain information without knowing the secret key;
- decryption is the opposite of the encryption process; decryption provides recovery of information from the cryptogram with the knowledge of the secret key;
- the cipher is a set of elementary steps for converting input data using a secret key; the cipher can be implemented as a computer program or as a separate device;
- a binary vector is a sequence of zero and unit bits, for example 101101011; the specific structure of the binary vector can be interpreted as a binary number, if we assume that the position of each bit corresponds to a binary digit, i.e. a binary vector can be associated with a numerical value, which is determined uniquely by the structure of the binary vector;
- cryptanalysis - a method of calculating a secret key to obtain unauthorized access to encrypted information or developing a method that provides access to encrypted information without calculating a secret key;
- one-way conversion is the conversion of an L-bit input data block to an L-bit output data block, which makes it easy to calculate the output block from the input block, and computing an input block that would be converted to a randomly selected output block is an almost impossible task;
- a one-way function is a function whose value is easily calculated by a given argument, however, calculating an argument from a given function value is a computationally difficult task; one-way functions are implemented as a sequence of procedures for one-way conversion of some input block (argument), the output value of which is taken as the value of the function;
- cryptographic strength is a measure of the reliability of encrypted information protection and represents the complexity, measured in the number of elementary operations that must be performed to recover information from a cryptogram with knowledge of the conversion algorithm, but without knowledge of the secret key; in the case of one-sided transformations, cryptographic strength is understood as the complexity of calculating the input value of a block from its output value;
- cyclic shift operations, depending on the converted sub-blocks or depending on the binary vector, are the cyclic shift operations by the number of bits specified by the sub-block value or the value of the binary vector; operations of cyclic shift to the left (right) are indicated by the sign "<<<"(">>>"), for example, the record В ₁ <<< В ₂ denotes the operation of cyclic left shift of the sub-block B ₁ by the number of bits equal to the value of the binary vector B ₂ ; similar operations are basic for the RC5 cipher [R. Rlvest, The RC5 Encryption Algorithm / Fast Software Encryption, Second International Workshop Proceedings (Leuven, Belgium, December 14-16, 1994), Lecture Notes in Computer Science, v. 1008, Springer-Verlag, 1995, pp. 86-96];
- a single operation is an operation performed on one operand (data block or binary vector); the value of a subblock after performing some given unary operation depends only on its initial value; an example of unary operations are cyclic shift operations:
- a double operation is an operation performed on two operands; the result of performing some given double operation depends on the value of each operand; examples of double operations are the operations of addition, subtraction, multiplication, etc.

 Known methods for block encryption of data, see, for example, US standard DES [National Bureau of Standards. Data Encryption Standard. Federal Information Processing Standards Publication 46, January 1977]. In this method, the encryption of data blocks is performed by generating a secret key, dividing the converted data block into two subunits L and R and alternating the latter by performing bitwise summing operations modulo two over the subunit L and the binary vector, which is formed as the output value of some function F from values of the sub-block R. After that, the blocks are rearranged. The function F in the specified method is implemented by performing permutation and substitution operations performed on the sub-block R. This method has a high conversion speed when implemented in the form of specialized electronic circuits.

 However, the known analogue method uses a small secret key (56 bits), which makes it vulnerable to cryptanalysis based on key selection. The latter is associated with the high computing power of modern computers of mass application.

The closest in technical essence to the claimed method of cryptographic conversion of L-bit input blocks of digital data into L-bit output blocks is the method described in [Kaliski BS, Robshaw MJB Fast Block Cipher Proposal / Fast Software Encryption. Proceedings of the Cambridge Security Workshop. Lecture Notes in Computer Science, v. 809, Springer-Verlag, 1994, pp. 26-39; see also B. Schneier, "Applied Cryptography". Second Eddition. John Wiley & Sons, Inc., New York, 1966, pp. 342-344]. The prototype method includes generating a secret key, splitting an input 1024-byte data block into 32-bit subblocks B ₀ , B ₁ , B ₂ . . .B ₂₅₅ and sequential conversion of subblocks. The secret key is formed in the form of a permutation table and an ordered sequence of 2048 subkeys Q ₀ , Q ₁ , Q ₂ ... Q ₂₀₄₇ , each of which has a length of 32 bits. Data block encryption is performed in the form of four rounds of transformations. One round of encryption is as follows. In accordance with the permutation table specified by the secret key, the data subunits are rearranged. Then all sub-blocks are converted one by one. A subunit, for example a subunit B ₁ , is transformed in the following way. A 32-bit binary vector V is formed from the values of the subblocks B _h , B _k , B _j , where h, i, j, k are different numbers, in accordance with the analytical expression V = В _h + F (B _h , B _k , B _j ) + Q _q , where the sign “+” denotes the summation operation modulo 2 ³² , q is the number of the current subkey. The numbers h, i, j, k are selected depending on the number of the encryption round and the step number of the conversion of the subunit. The generated binary vector is used to convert the sub-block B _i as follows. The bitwise summing operation of modulo 2 is performed on the subunit B _i and V and the value obtained after performing this operation is assigned to the subunit B _i . This is written in the form of the relation B _i ← B _i ⊕ V, where the sign "<-" denotes the assignment operation and the sign "⊕" denotes the bitwise summation operation modulo 2. After that, the other subunit is converted in the same way, etc., until all sub-blocks are converted. Each new step of generating a 32-bit binary vector is performed regardless of the value of the binary vector generated in the previous step. In one round of encryption, the index i takes in some order 256 different values corresponding to all the numbers of the subunits (from 0 to 255).

 This method provides high cryptographic strength due to the large length of the data block and the use of a large number of conversion operations per one sub-block.

 However, the prototype method has drawbacks, namely, in software implementation it does not provide the high speed cryptographic data conversion necessary to build software systems for protecting computer information operating in real time. For example, for the Pentium / 200 microprocessor, the encryption speed does not exceed 6 Mbps. This drawback is due to the fact that to ensure resistance to differential cryptanalysis, a large number of conversion operations per one bit of input data are required.

 The basis of the invention is to develop a method of cryptographic conversion of L-bit input blocks of digital data into L-bit output blocks, in which the conversion of input data is carried out in such a way as to reduce the number of conversion operations per one bit of input data, while ensuring high resistance to differential cryptanalysis, which increases the encryption speed.

Disclosure of invention
The problem is achieved in that in the method of cryptographic conversion of L-bit input blocks of digital data into L-bit output blocks, which consists in dividing the data block into N≥2 subblocks, alternately converting the subblocks by generating at least one binary vector from the value of the subblock and changing a subblock using a binary vector, a new one according to the invention is that the formation of a binary vector in the next step of converting a subblock is carried out depending on uktury binary vector in the preceding step of subblock conversion.

 Thanks to this solution, the best dispersion of the influence of the bits of the input data block on the bits of the output block is ensured, which ensures high resistance to differential cryptanalysis while reducing the number of conversion operations, which ensures an increase in the speed of cryptographic conversion. The best dispersion is ensured by the fact that the bits of the subblocks converted in the previous steps of the conversion influence the conversion process of the current block in two ways: (1) directly affecting the current value of the binary vector and (2) affecting the current value of the binary vector through the value of the binary vector in the previous steps of the conversion.

 Also new is the fact that two binary vectors are formed and one of them is transformed using the cyclic shift operation by the number of bits equal to the value of the second binary vector.

 Thanks to this solution, an additional increase in the cryptographic stability of the conversion of information blocks of length L = 512, 1024, 2048, 4096 and 8192 bits is provided, since the cyclic shift operations of this type are nonlinear and define a more complex function for generating a binary vector.

 It is also new that one of the subunits is changed by performing a cyclic shift operation on it by the number of bits different from the current value of the binary vector.

 Thanks to this solution, an additional increase in the cryptographic stability of the conversion is provided, since cyclic shift operations of this type are non-linear and are performed directly on the converted subunits.

 Also new is the fact that when converting subblocks, lookup tables are used, the number of which is T≥2, while the binary number is used to calculate the table number (v) and the subblock is changed using the lookup operation specified by the vth table.

 Thanks to this solution, an additional increase in the cryptographic stability of the conversion of information blocks of length L = 64 and 128 bits is provided, since the substitution operations are nonlinear and effective for implementation with a relatively small size of the converted information block. Using predetermined substitution operations is a way to achieve high cryptographic resistance to the most powerful cryptanalysis methods, in particular to differential cryptanalysis described, for example, in [Biham E., Shamir A. Differential Cryptanalysis of DES-like Cryptosystems / Journal of Cryptology, v .4, n.1, 1991, pp.3-72].

 Below the essence of the claimed invention is explained in more detail by examples of its implementation with reference to the accompanying drawings.

Brief Description of the Drawings
In FIG. 1 presents a generalized scheme of cryptographic conversion according to the claimed method.

 In FIG. 2 shows an encryption scheme corresponding to example 2.

 In FIG. 3 shows a one-way conversion scheme corresponding to example 3.

The best embodiments of the invention
The invention is illustrated by a generalized scheme of cryptographic conversion of data blocks based on the proposed method, which is presented in FIG. 1,
where B is the block to be transformed, b ₁ , b ₂ .... b _n are the converted subblocks, F is the operation block that performs the change of subblocks, f is the operation block that performs the formation of the binary vector, V ₀ is the initial value of the binary vector. The input L-bit block of digital data, where L is the number of bits in the block, is divided into N≥2 subunits, each of which has the size L / N bits. The solid line corresponds to the transfer of the converted subunits, the dotted line to the transmission of the binary vector.

 Subblocks are converted one by one by changing them using the operational block F, which uses the binary vector value in the transformations, which determines the dependence of the output value of the converted subblock on the value of the binary vector. The operation unit f uses the value of the subblock transformed in the previous step to perform the binary vector transform, i.e. the operation unit f forms a binary vector according to the structure of one of the converted subunits and according to the value of the binary vector that it had in the previous step of converting the subunit using a binary vector. This makes the binary vector value dependent on the data being converted. The execution of the transformation procedures specified by the function F over the subblock will be called one step of changing the subblock. The execution of the conversion procedures specified by the function f over the binary vector will be called the step of generating the binary vector. The subblock conversion step includes a binary vector generation step and a half block change step. The conversion steps are performed sequentially on all sub-blocks, after which the sub-blocks are rearranged.

 In special cases, the implementation of the proposed method, the permutation of the subunits may not be used. Performing N steps for changing a subblock, N steps for generating a binary vector, and permuting subblocks shown in FIG. 1, make up one round of conversion. The value of the binary vector after the Nth step of generating the binary vector serves as the initial value of the binary vector for the next conversion round when constructing one-way conversion functions of L-bit blocks of digital data. When building ciphers, the initial value of the binary vector for each round is formed depending on the secret key. The number of conversion rounds can be set depending on the specific construction options for the operational blocks F and f. Important in the given scheme is that the binary vector is formed depending on its structure at the previous step of transforming the sub-block using the binary vector.

The transformation of a subblock using a binary vector means (1) performing a two-place operation, the operands of which are a subblock and a binary vector, or (2) performing a one-time operation on a subblock (for example, substitution or permutation of bits), the modification of which is selected depending on the value of the binary vector , or (3) performing a two-place operation performed on a subblock and subkey, the number of which depends on the value of the binary vector. For example, in the first case, it can be a conversion of a 32-bit subblock B, when this subblock is assigned a value equal to the modulo 2 addition of ³² subblock B and the binary vector V: B <- B + V. In the second case, it can be a cyclic left shift operation performed on subblock B by the number of bits equal to the value of the binary vector analytically expressed by the relation B <- B <<< V. In the third case, it can be, for example, a transformation specified expression B ← B⊕Q _v , where Q _v is the subkey with number v calculated by the value of the binary vector: v = V mod 2 ¹¹ . The last relation determines the choice of the number equal to the value of the 11 least significant bits of the binary vector.

An important special case of the second use of the binary vector in the transformations of the subblock is the execution of a binary vector-dependent substitution operation on the subblock. This type of substitution can be used to set many different substitution tables, each of which is assigned a serial number, and to perform a substitution operation on a subblock, use a table whose number is selected depending on the value of V, for example, if 32 substitution tables are used, then the table number can be calculate by the formula v = V mod 2 ⁵ . The meaning of choosing a lookup table that is used to perform a lookup operation on a subblock at the current step, according to a specially generated binary vector, is to make the lookup table selection undefined for each step of the subblock conversion, which increases the strength of the cryptographic conversion.

By the formation of a binary vector we mean a record, for example, in a register or memory cell of a computing device of a certain sequence of single or zero bits. By the formation of a binary vector depending on its structure in the previous step of transforming a subblock using a binary vector, we mean setting the dependence of the current value of the generated binary vector on the value that the binary vector had in the previous step of using the data subblock. For example, let subblock B _i be converted using the value of binary vector V. Before using the binary vector at some other transformation step, for example, subblock B _{j, the} binary vector can be formed in accordance with the expression V <- V + Q _b , where b = B _i mod 2 ¹¹ - the number of the subkey, calculated depending on the value of the sub-block In _i .

 The possibility of technical implementation of the proposed method is illustrated by the following specific examples of its implementation.

 Example 1

This example is a variant of the proposed method, performing encryption of data blocks of 512-bytes (4096 bits). The use of this input block size is due to the fact that in computer systems this block size is standard. When using block ciphers that process blocks of this size, the possibility of random access to data stored on the built-in magnetic medium in encrypted form is preserved. In this example, an 8192-byte secret key is used, represented as a collection of 2048 numbered 32-bit connected {Q _v }, where v = 0, 1, 2 ... 2047. A secret key is formed, for example, by copying it from a magnetic medium to the main memory of a computer. The cryptographic transformation of example 1 is described by the following algorithm.

 Algorithm 1: Encryption of 4096-bit blocks.

ENTRANCE: 4096-bit block, represented as a collection of 128 32-bit subblocks {B ₁ , B ₂ ... B ₁₂₈ }.

1. Set the number of the processed block i = 1 and the initial values of the binary vectors V ₁ , V ₂ , V ₃ and v: V ₁ <- Q ₁ ; V ₂ <- Q ₂ ; V ₃ <- Q ₃ ; v <- 79.

2. Form the binary vector v according to the structure of the binary vector V ₁ : v ← v⊕ (V ₁ mod 2 ¹¹ ), where "⊕" is the summation operation modulo 2.

3. Generate the binary vector V ₂ : V ₂ ← [(V ₂ + Q _v ) ⊕ V ₁ ] >>> 11, where "+" is the summation operation modulo 2 ³² .

4. Form the binary vector v according to the structure of the binary vector V ₂ : v ← v⊕ (V ₂ mod 2 ¹¹ ).
5. Generate the binary vector V ₃ : V ₃ ← {[V ₃ >>> V ₂ ) ⊕ Q _v ] -V ₁ } >>> 22, where "-" denotes the subtraction operation modulo 2 ³² .

6. Generate the binary vector v according to the structure of the binary vector V ₃ : v ← v⊕ (V ₃ mod 2 ¹¹ ).
7. Generate the binary vector V ₁ : V ₁ <- V ₁ + Q _v .

8. Convert the subunit B _i : B _i ← [(B _i <<< V ₂ ) ⊕V ₃ ] + V ₁ .
9. Generate the binary vector V ₁ : V ₁ <- B _i .

10. Convert the sub-block B _i : B _i <- B _i + V ₂ .

 11. If i ≠ 128, then change the value of the number i: i <- i + 1 and go to step 2.

 12. Rearrange subblocks in reverse order.

OUTPUT: 4096-bit ciphertext block in the form of a set of converted subblocks {B ₁ , B ₂ ... B ₁₂₈ }.

 Algorithm 1 describes one round of encryption. After the first round, a second round of encryption is performed, taking the output block of the first round as input for the second round. Then perform the last (third) round of encryption, taking the output block of the second round as an input block for the third round. With a software implementation, the encryption speed when using three conversion rounds is about 30 Mbps for the Pentium / 200 microprocessor.

 Example 2

Consider a variant of the proposed method using substitution operations specified by lookup tables. Let the substitution operations be performed on subblocks of digital data of length k bits, where k is an integer. Then, to specify a substitution operation that converts a k-bit input subunit into a k-bit output subunit, a table containing two rows of numbers is required:
0 1 2 ... N-1
a ₀ a ₁ a ₂ ... a _N-1
where N = 2 ^k .

In this table, the bottom line contains all the possible values of the k-bit block exactly once, but in an arbitrary order. The order in which the numbers are located in the bottom line determines the specific variant of the lookup table, and therefore the specific version of the lookup operation performed using this table. The substitution operation is as follows. A number is selected on the top line that is equal to the value of the input block. The value under this number in the bottom line is taken as the output block. Thus, the substitution table can be placed in the main memory of the computer as a sequential record of k-bit computer words located in cells with addresses w ₀ , w ₁ , w ₂ . . . w _N-1 . In this case, the value of the input block b is used to calculate the address w ₀ + b of the word, which is taken as the output block. This way of representing a lookup table requires a memory size of kN bits.

We select the number of lookup tables equal to 2 ^L (the amount of required memory will be 2 ^L kN bits), and place the lookup tables continuously one after another. As the address of the table with number v, we take the value of the address w _{0 of} its first k-bit word. Let the address of the table with number 0 be s. In this case, the address of the lookup table with an arbitrary number v is s + vN. If a binary vector is specified that determines the number of the current lookup table v and the current input subblock to perform the lookup operation, then it is performed by replacing the current input block with a k-bit word located at s + vN + b, where b is the value of the subblock above which The current substitution operation is in progress. Using this relation, it is easy to set the selection of the lookup table with number v and perform the lookup over the subblock with value b. In the case under consideration, the assignment of the dependence of the lookup tables on the value of the binary vector and the performance of the lookup operation are carried out by the microprocessor very quickly when the appropriate values of the parameters L and k are selected, for example, with L = 5 and k = 8. With these parameters, 8 KB of RAM is required to place the lookup tables , which is acceptable, since modern computers have a RAM volume many orders of magnitude greater than this value (from 1 to 64 MB or more).

где R - число раундов шифрования. На r-том раунде шифрования используется r-тая строка подключей.Let L = 5 and k = 8, i.e. 32 tables are given that specify the substitution operations on 8-bit data subunits. We will form a secret key, presented as a combination of 7R 8-bit subkeys:

where R is the number of rounds of encryption. On the rth round of encryption, the rth row of subkeys is used.

We denote the used lookup tables as follows: T ₀ , T ₁ , T ₂ ... T ₃₁ , and the lookup operation defined by the table T _v as S _v , where v = 0, 1, 2 ... 31. The permutation tables T ₀ , T ₁ , T ₂ ... T ₁₅ can be chosen arbitrarily, and the tables T ₁₆ , T ₁₇ ... T ₃₁ are taken so that the permutation operations S _v and S _31-v are mutually inverse. The last condition is satisfied if the pairs of tables T ₁₆ and T ₁₅ ; T ₁₇ and T ₁₄ ; T ₁₈ and T ₁₃ ; ...; T ₃₁ and T ₀ will define mutually inverse substitution operations. For a set of arbitrary substitution tables T ₀ , T ₁ , T ₂ ... T ₁₅ it is easy to compile tables corresponding to the inverse substitution operations. For example, for the lookup operation specified by the following table
0 1 2 ... 255
a ₀ a ₁ a ₂ ... a ₂₅₅
and the reverse substitution is set by the table
0 1 2 ... 255
z ₀ z ₁ z ₂ ... z ₂₅₅
where the row (z ₀ , z ₁ , z ₂ ... z ₂₅₅ ) is obtained as the top row after ordering the columns of the previous table in ascending order of numbers in the bottom row.

In FIG. 2 shows a diagram of a first round of encryption. In FIG. 2, the solid vertical line corresponds to the transmission of 8-bit subblocks of data, the dashed line corresponds to the transmission of 5-bit subblocks, the horizontal solid line corresponds to the transmission of 8-bit subblock. The bitwise summation operation modulo two is indicated by the sign "⊕", v denotes the number of the selected lookup table, block S denotes the lookup operation, k ₁₁ , k ₁₂ . . . k ₁₇ - plug used in the first round. The arrows on the lines indicate the direction of signal transmission. Example 2 corresponds to the encryption of 64-bit digital data blocks. Encryption is performed in the following way. The input block is divided into 8 subblocks b ₀ , b ₁ , b ₂ ... b _{7 of} size 8 bits each. After that form the binary vector v, having the value of the 5 least significant bits of the subblock b ₀ : v <- b ₀ mod 2 ⁵ . Then, the bitwise summing operation modulo 2 is performed over the subblock b ₁ and subkey k ₁₁ and the output value of this operation is assigned to block b ₁ , which can be written analytically as follows: b ₁ ← b ₁ ⊕r ₁₁ . Then, according to the lookup table with number v, the lookup operation is performed on the subblock b ₁ : b ₁ <- S _v (b ₁ ). Then, according to the value of b ₁ , the binary vector v is formed: v ← v⊕ (b ₁ mod 2 ⁵ ), and the new value of the binary vector depends on its previous value. After that, the subblock b _{2 is} transformed: b ₂ ← b ₂ ⊕k ₁₂ and then b ₂ <- S _v (b ₂ ).

Similarly perform the conversion of the subblocks b ₃ , b ₄ , b ₅ , b ₆ and b ₇ . At the last step of each encryption round, the sub-blocks are rearranged in the reverse order, i.e. the b ₇ and b ₀ , b ₆ and b ₁ , b ₅ and b ₂ , b ₄ and b ₃ blocks are interchanged in pairs.

 The second round is the same, except that instead of the first row of subkeys, the second row of subkeys is used. Then the third round of encryption is performed using the third row of subkeys, etc. In total, R rounds of encryption are performed, where R = 4. In software implementation, this example implementation of the proposed method provides an encryption speed of about 25 Mbps for the Pentium / 200 microprocessor. If necessary, another number of rounds can be set, for example, R = 2, 3, 5, 6.

 Example 2

 An example is described by the following algorithm.

 Algorithm 2.

Entrance. 64-bit digital data input block, represented as a concatenation of 8-bit subblocks b ₀ | b ₁ | b ₂ | b ₃ | b ₄ | b ₅ | b ₆ | b ₇ , where the sign "|" denotes a concatenation operation.

 1. Set the number of rounds of encryption R = 4 and the counter of the number of rounds r = 1.

 2. Set the counter i = 1.

3. Form a binary vector v: v <- b _i-1 mod 2 ⁵ .

4. Convert the subblock b _i : b _i ← b _i ⊕k _ri , b _i <- S _v (b _i ), where the substitution operation S _v is performed using the substitution table with number v.

5. Generate the binary vector v: v ← v⊕ (b _i mod 2 ⁵ ).

 6. If i ≠ 7, then increment i <- i + 1 and go to step 4.

 7. If r ≠ R, then increment r <- r + 1. Otherwise STOP.

 8. Rearrange the subunits in the reverse order and go to step 3.

 Output: 64-bit ciphertext block.

 In this example, it can be seen that the number of the used lookup table depends on the blocks being converted and is not predetermined for the current conversion step, i.e., the substitution operation is not known in advance for all conversion steps. It is determined by a secret key and a block of converted data. Decryption is carried out similarly and is described by the following algorithm.

 Algorithm 3.

Input: 64-bit ciphertext input block b ₀ | b ₁ | b ₂ | b ₃ | b ₄ | b ₅ | b ₆ | b ₇ .
1. Set the number of rounds of encryption R = 4 and the counter of the number of rounds r = 1.

 2. Set the counter i = 1.

3. Form a binary vector v: v <- b _i-1 mod 2 ⁵ .

4. Save the value of b _i in the variable g: g <- b _i . Convert subblock b _i : b _i <- S _31-v (b _i ), b _i ← b _i ⊕k _r′1 , where r '= 5-r.

5. Generate the binary vector v: v ← v⊕ (g mod 2 ⁵ )].
6. If i ≠ 7, then increment i <- i + 1 and go to step 4.

 7. If r ≠ R, then increment r <- r + 1. Otherwise STOP.

 8. Rearrange the subunits in the reverse order and go to step 3.

 Output: 64-bit block of source text.

 The above encryption and decryption algorithms can be easily modified to convert data blocks of a different size, for example, 128 and 256 bits.

 Example 3

This example relates to the construction of a one-way function based on the claimed method of cryptographic conversion. As in examples 1 and 2, it is assumed that 32 substitution tables T ₀ , T ₁ , T ₂ ... T _{31 are used} . Substitution tables are assumed to be known and private keys are not used. The one-way function is defined by algorithm 4. The sequence of operations for converting data subblocks for one round of transformation is shown in FIG. 2.

 Algorithm 4.

Entrance. A 64-bit input data block, represented as a concatenation of 8-bit subblocks b ₀ | b ₁ | b ₂ | ... | b ₇ .
1. Set the number of conversion rounds R = 8, the counter of the number of rounds r = 1 and the initial value of the binary vector v = 13.

 2. Set the counter i = 1.

3. Generate the binary vector v: v ← v⊕ (b _i-1 mod 2 ⁵ ).
4. Convert the subblock b _i-1 : b _i-1 <- b _i-1 <<< r, b _i-1 <- S _v (b _i-1 ).

 6. If i ≠ 8, then increment i <- i + 1 and go to step 3.

7. Rearrange the subblocks b ₀ , b ₁ ... b ₇ in the reverse order.

 8. If r ≠ R, then increment r <- r + 1 and go to step 2. Otherwise, STOP.

 Exit; 64-bit value of function F.

 Similarly, a one-way function can be constructed for converting 128-bit data blocks, which can be used to hash data.

 The above examples show that the proposed method of cryptographic transformations of digital data blocks is technically feasible and allows us to solve the problem.

Industrial applicability
The inventive method can be implemented, for example, on personal computers and provides the ability to create on its basis high-speed encryption software modules and replace expensive specialized encryption equipment for personal computers equipped with a high-speed encryption software system.

Claims (3)

 1. The method of cryptographic conversion of L-bit input blocks of digital data into L-bit output blocks, which consists in dividing the data block into N ≥ 2 subblocks, over which N steps of transforming a subblock are successively performed, at least at least one step of transforming a subblock one binary vector according to the value of the subblock and change the subblock depending on the structure of the binary vector, characterized in that the formation of the binary vector in the next step of the transformation of the subblock is carried out depending from the structure of the binary vector in the previous step of the conversion of the subunit.  2. The method according to claim 1, characterized in that the subblock is changed by performing a cyclic shift operation on it by the number of bits equal to the current value of the binary vector.  3. The method according to claim 1, characterized in that the subblock is changed by calculating the number V of the lookup table depending on the structure of the binary vector and then performing the substitution operation on the subblock corresponding to the Vth lookup table.

Images