JPS60238887A - Protective management of information - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] This invention relates to the remainder theorem (for example, J, J, 5tone, 'Multiple
-burst error correction w
It's the Chineseremainder
theorem, 'J, SIAM, vol.1.11. pp
74-81, March 1963) has been proposed.

Here I will explain the Chinese surplus principle.

Suppose that F, rni (i=1.2...r) is given a relatively prime integer a·(1=:1,2...,r).

There is always only one integer f that satisfies.

any integer greater than or equal to 0 and less than M
d, mi (i=1, 2, . . . +, r) and their remainder K.

Also, let ml (i==1.2...r) be relatively prime.

At this time, f can be reproduced by briefly explaining using ai (i=1, 2, -r). here.

The following shall be satisfied.

therefore.

Because I'm yelling.

By Fumera ``Ru.

The above theorem will be explained.

(a) Mutually prime integer m1 (i=1.2...r
) can be divided by the integer mi of n, and an integer f having any value remaining when divided by the integer mi of n can be electrified.

Then, let a certain number be a relatively prime integer mi,
It is shown that it is possible to convert into a sequence of 4 subtracted ai and then recover an integer f with a likelihood of 1 from the sequence.

(b) At this time, it is important that even one element of the remainder sequence is missing and the original numerical value cannot be determined.

For example, if ar, which is the remainder of mr, is not given, then information is given that when the integer f that can be determined is divided by another relatively prime integer mi within the range of integers f, the remainder is ai. If you can get rid of it.

Determine one by covering all possible integers.

First, if we divide a certain integer f by multiple integers mi that are mutually cumulative, and note the remainder, then for any remainder vat such that 0≦f≦, π mi, 1=1.Thus, the integer f can be determined.

Since the amount of information ↑R that the remainder 9al is waiting for is only one part of the integer f, f can be determined only when the information world of the remainder at is collected for the amount of information f. . This idea is (K, N)t, the root of the threshold problem.

Show calculation example -f.

m1=5-m2=61m3=7, M=5
X6 x7=210.

-210 m3"4"------"4=1..2031 (mod
7), so t=4 (rn1=5 m=6 m5==7 ~t1=3 t2=5 t=4.

The number f such that 1.0≦f≦210 is divided by ml, and the remainder when ν is divided by n1-21m2 is a2241m.
, if a3=1, then '1-rudo, 16
) From the formula, 10-'t3'a.

3 210 210 =-1・3・2+−5・4 6 10 10−−4@1 =1072 1072 (mod 210)=22,', f=
It can be written as 22.

f-22 as m 1 = 5, m 2 = 6,
Dividing by m 5 = 7 and finding the remainder, a 1 =
2, a2=4.

a5-1, and it can be confirmed that this redundancy is correct.

As already mentioned, the Chinese remainder theory is to express and encode integers in the remainder form. However, because the remainder forms of different integers that are prime to each other are used, the size of the divided information becomes uneven. l again. Because the number of prime integers is limited, information divided by 1 and 1 is interesting.

To improve this point, we extend this theorem to the irreducible polynomial of GF2 υi'N, 'This is Stone's theorem, but GF2
There is only one approximately polynomial in 2rF,,3.

It is not more than 4 and cannot add 1 to c. Therefore, 1 division n
The information must be at least 4 points in order to be able to read it.

In the example in Table 1, the prime numbers 2, 3, and 5 are selected.

The numbers that can be represented are from 1 to 61 in Table 1! :1. In the remainder form of 'E7j2,3.5, the amount of remainder information is too uneven. Therefore, Stone's encoding theorem uses an irreducible polynomial of Galois field GF2. However, there are few irreducible polynomials of GF2 with low elements, and there are only three or more polynomials with less than five degrees.

[Summary of the invention]

This invention was made in order to eliminate the drawbacks of the above-mentioned "conventional" method, and it is a method of dividing information into two or more pintos in which the remainder form obtained by dividing a numerical sequence by an irreducible polynomial is arranged so as not to overlap. An object of the present invention is to provide a data security management method that can safely manage information by obtaining information and managing this divided information. .

Table 2 shows an example in which an integer from 0 to 151 is divided into numbers from 0 to 31. A is the remainder form of 4. B
, C, DH, which is the rearrangement of the three remainder forms of the integers. From this table, it can be seen that all numbers from 0 to 151 are represented by different number sequences. Then, as in the case of using the Chinese remainder theorem, the most likely information is reproduced from any two pieces of divided information.

) Table 2 Table 3 is a binary representation of Table 2.

It can be seen that if the friend information is divided into υ f'F and two pieces of information are expressed in 2 points, then 4 points of information will be reproduced.

Table 3 For example, if (A=11, C=01) is given in Table 3, then the information of one element is [1011]. At this time,
Even if the information that B=10 is missing, M can be determined.

That is, when the original information of 4 bits is divided into three pieces of information of 2 bits each, the original information can be reproduced if 4 bits worth of information is obtained from any two of them.

Table 3 shows 9. Since it can be achieved by normal logic or computer software processing, and is just a take A, the required opening time is 'f#'. Furthermore, by combining several pieces of logic like this, it becomes possible to perform divisional decoding of an arbitrary number of pieces.

As shown in Tables 2 and 3, the sequence of A is the remainder when the sequence of one-element information , which is a rearrangement of the number sequences B, C, and Dtri.

Encoding and decoding processing will be explained using Tables 3 and 4.

Table 4 The original information of 1.4 focus is set to 1011.

When we refer to the first row of Table 3 and encode it, A=11°■
U=10.0=01, D=OO and 2-bit division (F
I got 4 i-reports.

Among the four pieces of divided information encoded in this way.

Any two items are taken out, checked against Table 4, and those that match are selected and the original information is decoded. for example.

When (A=11, B=10) is given, the 1-element information X becomes 1011. Similarly (A=11.C=0
1), (A=1 1, 0=OO), (B=1
0.

C=01), (B=10, D=OO), (
When C=01°D-00) is given, X is also 1011
Therefore, every n is correctly decoded.

Error verification processing for vaguely decoded information (9) will be explained.

Here, we will set up a case where B=10 becomes B=01 due to an error.

From Table 4, when (A=11, B=01) is given, the information X from Table 4 becomes 1111. below.

Similarly, when (A=11.0=01), X-1011
, (A=11, D=OO), then X=1011, (
B=01, C=01), then X=0001, CB
=01. When D=OOノ, X-0110, (C=01
, D=OO), X=1011, and since X is not equal, it is possible to detect that an error has occurred. Buta, for this process.

Even if two of A to D are incorrect, a correct test is possible. The process of correcting errors will be explained.

As mentioned above, A is correct if B=O].

By combining the two silks by C and D, the correct 1'h information, X=1011, appears.

And, when using the incorrect B as (A, B), (B, C), (B, I)), X is not equal as described above.

Therefore, it gives three equal Xs (A, CJ.

If you output the variations of (A, DJ, (C, D)), you have made one correction.

As described above, it is possible to detect errors in 2 out of 8 focus points, and to correct 1-bit H error 9.

Even if two of A-D are stiff, it can be tested even if one causes a 4-bit error, and even if one of A-D is stiff, even if one or two causes a slight focus error. However, it is possible to correct it.

That is, any part of the information divided into N pieces can be decoded. In the above embodiment, the number of divisions is 4, and arbitrary two of them (K=2) are used for decoding, but the number of divisions is not limited to 4'. In the above embodiment, the division information is 2 bits and the original information is 4 bits, but the division information is 3 bits, 4 bits, or the original information is also 9 bits.

Even if it is 16 bits, you can select <, N, or Kfl.

Since it is easy to configure the device to obtain the values shown in Table 2 by referring to the table in memory or by calculating as necessary, it is possible to easily configure the device to obtain the values shown in Table 2 by referring to the table in memory or by calculating as necessary. It is also possible to make this device function normally. Furthermore, by applying the present invention to devices that generally have many failures, such as relays, it becomes possible to improve the reliability of the device. -F. Since this invention has the ability to test false V, it can of course be applied to single communications.

The figure is a block diagram of a data protection management system according to the method of this invention. The database 1flN pieces of divided information 2 can be divided and managed by the above-mentioned method of the present invention. The divided information 2 is decoded by the above-described method of the present invention, and becomes decoded data 3.

〔Effect of the invention〕

As described above, according to the present invention, it is possible to perform division protection, false V test, and error correction for even smaller information.

Since it is possible to correct a small number of focus errors, it can be applied to improve reliability, such as duplicating one device.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of a data protection management system according to the present invention. 1... Database, 2... Division information, 3... Decrypted data. Patent applicant: Mitsubishi Electric Corporation (2 others) -' ↑4゛Prior Office 1, Indication of case: Japanese Patent Application No. 59-94473 2, Invention name information protection management method 3, Representative of the person making the amendment To Hitoshi Katayama, Part 5, Detailed explanation of the invention column 6 of the specification subject to amendment, and the description of the contents of the amendment are corrected as follows. The specification is amended as follows.

Claims (1)

[Claims] Generate division information in which the remainder forms obtained by dividing the numerical value of the original information to be protected and managed by an irreducible polynomial are arranged so as not to overlap with each other, and from this division information, decode the original information 1-
When the above division information exceeds the number of pieces in stock, the above information is compared with the table set corresponding to the above original information, and by detecting the ↑ h information that matches each other, the above original information is obtained. An information protection management method that decrypts .