JP3133734B2 - Code conversion method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a code conversion method including a decoding method for transmitting or storing the contents of digitalized plaintext while keeping secret, and relates to an encoding key and a decoding method. A chain code conversion system that uses a Latin block as a key as a conversion table and performs feedback coding and feedforward decoding using a plaintext signal of a predetermined number of bits (or 1 bit) of plaintext as input data as a unit block. It is concerned.

[0002]

2. Description of the Related Art As a means for keeping data confidential,
Various code conversion methods have been proposed so far,
A typical example is DES (Da
taEncryption Standard) and NTT's cipher FEAL-8.

[0003] These are one of the block ciphers that encode or decode input data in block units under the control of a secret encoding key. A strong code (encryption) is obtained by repeatedly performing conversion. Have gained.

However, as for this (block) code, as long as the same key is used, only the same code block (hereinafter referred to as a code signal) is generated from the same plaintext block (hereinafter simply referred to as a plaintext signal). There were drawbacks.

[0005] This drawback does not necessarily mean that code analysis is easy, but it is attempted to decode a large amount of input data if a large number of the same code signals are scattered in the data. Could give some hints to the people.

In order to solve this drawback, a feedback code conversion system shown in FIG. 2 has been adopted.
In this feedback code conversion method, the encoding device A uses a key to transmit a plaintext signal P _i (block number i = 1,
2, when generating a code signal C _i from ...) is stored in the temporary storage unit 8 the code signal C _i-1 of the preceding as a feedback signal C _b, the feedback signal C _b
In the those calculates the plaintext signal P _i at operation xor unit 9 and a code signal C _i, the decoding apparatus B side, decoding generates a plaintext signal P _i from the code signal C _i by using the same key, prior code signal C _i-1 feedforward signal C
_f and then is stored in the temporary storage unit 8, and performs decoding those calculated and the feedforward signal C _f and the code signal C _i by computing xor section 9 as clear text signal P _i.

[0007] IV is an initial value signal, the first plaintext signal P _i
Of use in decoding the coding and code signal C _i, are prepared in advance value (initial value signal IV, the encoding apparatus A side and the decoding apparatus B
Or advance distributed for both sides, or distributed with the plain text signal P _i) is used.

In the above-mentioned conventional technique, the operation xo
r is the input value of the plaintext signal P _i and the preceding code signal C _i−1 , or the code signal C _i and the preceding code signal C _i−1
, A fixed and always calculated value called exclusive OR (addition of modulo 2) is returned as a result.

Therefore, when the encoding / decoding operation xor, that is, the exclusive OR (addition of modulo 2) is used in the feedback encoding method, the result of the operation is always the addition of modulo 2 between the input value and the feedback value. , The conversion result is fixed. That is, in a block of input value, feedback value and output value (one digit binary number), the output value is 0 when both the input value and the feedback value are 0 or 1, and Since the output value is 1 when the input value and the feedback value are 0 and 1 or 1 and 0, the output result is fixedly determined for each bit, which may give a hint for data decoding. There were drawbacks.

In order to solve the drawbacks of the prior art, the applicant of the present invention has previously described an operation xor of encoding / decoding in a feedback encoding system, that is, an exclusive OR (modulo 2).
In order to make the value of the operation result of the operation of the present invention freely and to make it extremely difficult to illicitly decode the input data from the encoded data, a Latin square is used as a conversion table instead of the operation xor unit 9. An application using a conversion unit has been filed as Japanese Patent Application No. 162152/1996.

[0011]

SUMMARY OF THE INVENTION The applicant's earlier patent application exhibits excellent confidentiality and randomness,
Although it can be easily implemented, there has been a problem that it is difficult to meet the demand for higher confidentiality and randomness, and the demand for safer and simpler implementation.

The present invention has been made in order to satisfy the applicant's request for the earlier patent application, and it is possible to surely and easily increase the confidentiality and randomness.
It is intended to be able to be implemented accurately and easily.

[0013]

Means for Solving the Problems In the following description, a case where natural numbers are used as the values of the row, column, and depth elements of the Latin box is shown. However, the row, column, and depth elements of the Latin box are shown. May be letters of the Latin alphabet. Up
A first invention for achieving the above object is a feedback code.
A plaintext signal of a predetermined number of bits is simply
Generates a code signal obtained by performing encoding as a
Transmission, and the decoding unit of the feedback decoding scheme Te,
The value is decoded by performing decoding with the code signal as the unit block.
In a code conversion method of outputting a plaintext signal
Unit has a respective original set of n symbols, row, column and back
One for each of the n elements in the three directions going
And, and, in the same direction so as not to place the same source
A latte represented by a cube consisting of a total of n ³ elements
Is used as a conversion table, and the first conversion
-Unit plaintext signal, initial value signal, and first selection sequence generator
The row direction , column direction , and depth of the conversion table
The conversion determined by specifying each of the pointers for the directions
The value of the table element and outputs a code signal, the second and subsequent odd
In other words, conversion is performed from the plaintext signal and the plaintext signal input immediately before.
Feedback signal which is the encoded signal and the first selection
According to the selection signal from the sequence generator, the row direction of the conversion table ,
Specify the row direction and depth direction pointers
Output the value of the element of the conversion table thus specified as a code signal.
It was a good thing . Also, the decoding unit is a modification of the encoding unit.
Perform inverse transformation corresponding to the Latin box used as the permutation table.
Using the No. of Latin rectangular parallelepiped, the first conversion, is input
The same initial value signal as used for encoding and the second selection
By the selection signal from択系train generator unit, decryption Latin
Pointers in the row , column, and depth directions of the body
In the second and subsequent conversions, the
The feedforward signal, which is the input sign signal, and the second
Selection signal and by decoding latte from the selection sequence generation unit
Pointers in the row , column, and depth directions of the
Decoded Latin , determined by specifying each
Output the values of the body elements as decoded plaintext signals
It is characterized by.

The invention of claim 1 can be represented by the following formula (hereinafter, referred to as
See FIG. 1), the _{encoding; C i = L (P i} , C i-1, S i) ^{_{decoding; P i = L -1 (C}} i, a _{C i-1, S i)} . Here, P _i is a plaintext signal (signal number i = 1,
2, ···), C _i the code signal (signal number i = 0, 1,
2,... Where C ₀ = IV uses a previously prepared value as an initial value), and S _i is a selection signal (signal number i = 1,
2, ...). L is a Latin box used for encoding, and L ^-1 is a Latin box used for decoding.

The code signal C and a preceding value of the plain text signal P _i
The value of _{i-1 and} the value of the selection signal S _i are values (pointers) indicating the row and column and depth positions of the coded Latin box L, respectively.
Then, the code signal C _i is the value of the element of the coded Latin box L indicated by the row pointer, the column pointer, and the depth pointer. The decoded Latin cube L ^-1 is the encoded Latin cube L
When considering as a row or column or Latin of inverse functions that are reversible with respect to depth, when the code signal C _i and row pointer, an inverse function related to the coding Latin body L rows.

When the plaintext signal P _i , the preceding code signal C _i−1 and the selection signal S _i are input to the coded Latin box L, the code signal C _i is obtained as an output, and the code signal C _i precedes the code signal C _i . When the code signal C _i-1 and the selection signal S _i are input to the inverse function decoding Latin box L ^-1 , a plaintext signal P _i is obtained as an output.

The encoding is represented by a conversion table value = C _i = L (row pointer value = P _i , column pointer value = C _i−1 , depth pointer value = S _i ). If the decoded Latin box L ^-1 is an inverse function of the encoded Latin box L, the value of the conversion table = P _i = L ⁻¹ (the value of the row pointer = C _i , the value of the column pointer = C _i−1 , the value of the depth pointer = S _i ), and when the decoded Latin cube L ⁻¹ is the same Latin cube as the encoded Latin cube L, the value of the row pointer = P _i = L ^-1 (conversion table value = C _i , column pointer value = C _i−1 , depth pointer value = S _i ).

A Latin cube is a set A consisting of n symbols.
= {A ₁ , a ₂ ,..., A _n } is used n times, and a total of n ³ elements are arranged in three directions (row (vertical) direction, column (horizontal)
Direction, depth (front-back) direction], and each element of the set A appears once in each direction in each direction, that is, an nth-order Latin that does not have the same value at the same position in each other A stack of n squares is called a Latin cube on A or an nth-order Latin cube.

As described above, by using the Latin cube for code conversion, the conversion is performed in units of a plurality of bits and becomes non-linear, and any one of the row value, the column value, the depth value, and the element value is used. Once one is determined, the other one is inevitably determined, so the conversion is accurate and simple.

Since the values of the elements of the Latin box are not the same in the row, column, and depth directions, that is, there is no overlap, uniformity is ensured, and the randomness, that is, the randomness is extremely good. Become.

As compared with the case where the Latin square is used as the conversion table, the selection signal S _i is added as a key signal, and accordingly, the input parameters increase and the complexity increases.

A second invention according to a second aspect is the first invention.
The first and second selection sequence generators in
And register, off the output of the shift register to the shift register
Linear consisting of feedback coupled circuit with feedback
The feature is that it consists of a feedback shift register
I do.

[0023] In the invention of claim 2 wherein has a sequence longer period of the selection signal S _i to be generated, also a statistically random, practical and can be more easily generated is there.

[0024]

Embodiments of the present invention will be described below with reference to FIG. FIG. 1 shows the most basic configuration example of the present invention, in which (a) shows an encoding device A,
(B) shows a decoding device B, which comprises a conversion unit 1 having a quaternary Latin cube as a conversion table, a row specification unit 2 for specifying a row pointer, and a column pointer for temporarily storing a column pointer. It is composed of one encoding unit a and one decoding unit b composed of a column designation unit 3 for storing and designating, a depth designating unit 4 for designating a depth pointer, and a selection sequence generating unit 5. The generator 5 is constituted by a linear feedback shift register including a shift register 6 and a feedback coupling circuit 7.

The decoded Latin box L ^-1 of the decoding unit b is an inverse function of the encoded Latin box L of the encoding unit a, and
The plain text signal _Pi is used as a row pointer in the row specifying unit 2 of the encoding unit a, the initial value signal IV and the feedback signal _Cb are used as the column pointer in the column specifying unit 3, and the selection signal is supplied to the depth specifying unit 4. because S _i is input as the depth pointer, the decoding section b, as the code signal C _i is the row pointer in the row designation unit 2, also the initial value signal IV and feedforward signal C _f the column specification section 3 columns The selection signal S _i is input to the depth specifying unit 4 as a pointer and as a depth pointer.

Since the encoded Latin box L and the decoded Latin box L ^-1 are a four-dimensional three-dimensional array table, signal conversion processing is performed in units of 2 bits, and this conversion processing operation is repeated. , A code of an arbitrary length can be generated.

FIG. 1 shows the basic configuration of the present invention. The conversion unit 1 is not composed of only the encoded Latin box L or the decoded Latin box L- ^1. May be configured by combining an encoding function, which is an encoding algorithm, with the encoded Latin box L or the decoded Latin box L ⁻¹ ,
Further, the encoding device A and the decoding device B may be configured by connecting a plurality of encoding units a and decoding units b in series or in parallel, or in series or in parallel.

An encoding device A and a decoding device B
The, when composed of a plurality of encoding unit a and the decoding unit b is a feedback signal C _b and feedforward signal C _f instead sign signal C _i from autologous or to self, for other conversion section 1 may be utilized code signal C _i.

[0029]

Since the present invention has the above-described structure, the following effects can be obtained. Since a Latin cube is used for the conversion unit, which is an operation part for performing encoding / decoding in the block encoding method, the value of the operation result can be a plurality of natural numbers, and encryption of a plurality of bits is performed. This makes the encoding non-linear, increases the number of bits in the transform unit, increases complexity, and guarantees uniformity because there is no duplication of the values of each element, thereby providing extremely good randomness. It is possible to easily obtain a cipher exhibiting high secrecy.

Since encoding and decoding are performed by determining the remaining one out of three of the row value, column value, depth value, and element value of the Latin cube, the encoding unit and the decoding unit Not only can the conversion in the conversion unit be accurate and simple, but the confidentiality is enhanced because there are two key signals to be used.

According to the second aspect of the present invention, excellent characteristics exhibited by the linear feedback shift register in code conversion, that is, a sequence of generated codes has a long period and is statistically random. By maintaining the characteristics that it can be easily generated and practical, it is possible to completely eliminate the problem of low difficulty, which is the greatest drawback, by combining it with a Latin cube. Therefore, when performing code conversion, excellent characteristics are exhibited.

[Brief description of the drawings]

FIG. 1 is a block circuit diagram showing a basic configuration of the present invention.

FIG. 2 is a block circuit diagram showing a conventional example.

[Explanation of symbols]

A: Encoding device a; Encoding unit B; Decoding device b; Decoding unit 1; Conversion unit 2; Row designation unit 3; Column designation unit 4; Depth designation unit 5; Selection sequence generation unit 6; Shift register 7 Feedback coupling circuit 8; Temporary storage unit 9; Operation xor unit P _i ; Plain text signal C _i ; Code signal IV; Initial value signal L; Encoded Latin cube L ^-1 ; Decoded Latin cube C _b ; C _f ; feed forward signal S _i ; selection signal

──────────────────────────────────────────────────続 き Continuation of front page (58) Field surveyed (Int.Cl. ⁷ , DB name) H04L 9/00 G09C 1/00-1/14

Claims (2)

(57) [Claims] 1. A feedback encoding system comprising :
More, marks the unit block plaintext signal of a predetermined bit number
It generates a code signal obtained by performing-coding and transmission, feed
The code signal is simply converted by a decoding unit of the back decoding system.
The flat block whose value has been decoded by decoding
In the code conversion method for outputting Bunshingo, the encoding unit, each original set of n symbols, lines,
One for each of the n elements in each of the three directions of row and depth
One uses, and does not place the same source in the same direction
Represented by a cube consisting of n ³ elements in total
Latin cubes are used as the conversion table, and in the first conversion,
The predetermined number of bits of plaintext signal, initial value signal and first selection
With the selection signal from the sequence generator , the conversion table
Point the pointer in the row , column, and depth directions, respectively.
The value of the element of the conversion table determined and determined as the code signal
Outputs Te, the second and subsequent conversion, the plaintext signal and the previous
Is a code signal converted from the plaintext signal input to
Feedback signal and selection from the first selection sequence generator.
With the selection signal , the row direction , column direction and
The change determined by designating the pointer in the depth direction.
The values of the elements of the permutation table are output as the code signals, and the decoding unit uses
Decoded Latin box that performs inverse transformation corresponding to the Latin box
In the first conversion, the input code signal and
Generating the same initial value signal and second selection sequence as used for encoding
By the selection signal from the parts, before the decoding Latin body
Pointers in the writing direction , column direction and depth direction are respectively
In the second and subsequent conversions,
The feedforward signal, which is the input code signal, and the
By the selection signal from the second selection sequence generating unit, said decoding
In the row direction , column direction and depth direction
It is determined by specifying the interface, respectively, wherein
The value of the element in the decoded Latin box is used as the decoded plaintext signal.
A code conversion method characterized in that the code is output . 2. The method according to claim 1, wherein the first and second selection sequence generators include:
A shift register, the output of the shift register said shift
Feedback loop that feeds back to the register
And a linear feedback shift register
Code conversion method according to claim 1, wherein the a.