KR20030059500A - Pseudo random number generator formed by spn structure using a block code and method thereof - Google Patents

Abstract

The present invention relates to a pseudo random number generator using a block cipher having an SPN structure. That is, in the pseudorandom number generator proposed by the present invention, an update algorithm for updating a key value used for generating a pseudorandom number instead of writing a timestamp on one key like a random number generator using a block cipher proposed in ANSI X9.17. By changing the key value every time a random number is generated there is an advantage to further enhance the random number. In the present invention, there is an advantage in that the efficiency can be improved by updating a previously generated key through a simple logical operation, instead of regenerating the key in changing the key value, IC card in a limited environment as well as a PC This has the advantage of being applicable to all the same various platforms.

Description

Pseudo-random number generator and method using block cipher having S.P.N structure (PSEUDO RANDOM NUMBER GENERATOR FORMED BY SPN STRUCTURE USING A BLOCK CODE AND METHOD THEREOF)

The present invention relates to a pseudo random number generator, and more particularly, to a pseudo random number generator using a block cipher having an SPN structure.

In general, it is impossible to generate a true random number, so a pseudo-random number that is difficult to distinguish from a random number is generated and used together with the random number. Pseudorandom numbers are numbers that do not have an efficient algorithm to distinguish them from random numbers.

The pseudo random number is generated by using a hardware chip or by using a cryptographic function.

The random number generation method using the hardware chip includes a method of using a high-entropy noise source obtained by hardware such as heat and noise generated from the hardware chip, a quantum effect of a semiconductor, a radiation source, and a turbulence in a disk drive after simple filtering, and a key stroke. For example, there is a method that uses the output value as a random function by inputting a noise source having a lower entropy than the above noise, such as CPU time and a hash function or block cipher algorithm.

As a representative example of generating pseudorandom numbers using the hardware, there is a pseudorandom number generating method that uses heat generated inside the hardware of the Intel 82802 Fireware Hub (FWH) chip to generate pseudorandom numbers, as described above. In the case of generating pseudorandom using, the generated pseudorandom is very high randomness like true random number, while it is difficult to obtain high entropy noise, difficult to apply to various platforms, and expensive There was a lifting issue.

Next, in the case of generating pseudorandom numbers using cryptographic functions, there is a disadvantage of low randomness. However, it is sufficient to assume that the hash function or block cipher algorithm is difficult to analyze without revealing the key or seed used in generating pseudorandom numbers. It can guarantee stability, can be applied to multiple platforms, and has the advantage of generating pseudorandom sequence of desired length.

As a cryptographic algorithm, ANSI X9.17 is mainly used as a conventional block cipher. ANSI X9.17 uses a time stamp on a key to generate random numbers or seeds. When performing about 2 to ³² iterations of the time output, there is a weak point that the output of the X9.17 pseudorandom number generator can be distinguished from the true random bit sequence.

Therefore, the object of the present invention is applicable to a variety of platforms, especially in a limited hardware environment such as an IC with a limited noise source, and also compared with the random number generator using the block cipher proposed in the existing ANSI X9.17. The present invention provides a pseudo random number generator and method using a block cipher having a new type of SPN structure.

In the present invention for achieving the above object, in the pseudo random number generator and method using a block cipher having an SPN structure, the noise to be collected for each platform to be used as input of a random function when the pseudo random number is generated from the noise information ( keying generation module for generating a value; Two random function value converters, and generate a first random function value (Rand1) using the key value and a state value as input values of a first random function value converter, and generate the first random function value and A pseudorandom number generation module for generating a second random function value (Rand2) by using a key value as an input value of a second random function value conversion unit, and outputting the second random function value as a pseudorandom value; Generates a third random function value (Rand3) by using the second random function value generated from the pseudo random number generation module and the updated new key value as an input value of a random function value conversion unit, and generates a random number similar to the receiving generation module An update module configured to provide a module to update and change state information values of the key value and the random function value conversion unit; and implement a pseudo random number generator using a block cipher having an SPN structure including

(a) generating a seed value from the noise information; (b) generating a key value using the seed value; (c) calculating a first random function value (Rand1) from a random function having the key value and the last updated state value from the update module as an input data value; (d) calculating a second random function value (Rand2) output from the random function using the first random function value (Rand1) and the key value as an input data value; (e) outputting a value generated as the second random function value (Rand2) as a pseudo-random value; (f) generating an update random function value (Rand3) from a random function having the second random function value (Rand2) and the updated new key value as an input data value to update the seed value and the state value of the random function; It is characterized in that it implements a method of generating a pseudo-random number using a block cipher having a SPN structure to proceed.

1 is a functional block diagram of a pseudorandom number generator according to an embodiment of the present invention.

2 is a functional block diagram of a key generation unit according to an embodiment of the present invention.

3 is a functional block diagram of a KF function unit in a key generation unit according to an embodiment of the present invention.

4 is a functional block diagram of a random function unit in a pseudorandom number generation module according to an embodiment of the present invention.

5 is a functional block diagram of the Feistel function of the second round SPN structure in the random function according to an embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings will be described in detail the operation of the preferred embodiment according to the present invention.

1 illustrates a functional block configuration of a pseudorandom number generator according to an embodiment of the present invention. Referring to FIG. 1, the pseudo random number generator 100 is composed of a receiving module 102, a pseudo random number generating module 104, and an update module 106. As shown in FIG. 1, the receiving module 102 includes a noise generator 108 that provides noise information, a seed generator 110 that generates a seed from the noise information, and a seed. It consists of a key generation unit 112 for generating a key using.

The seed generator 110 generates a seed for generating a pseudo random number using noise information that is difficult to predict or control, and determines a function for collecting noise to be used according to each platform on which a program is implemented. In this case, the collected noise should be at least 128 bits.

At this time, the seed generator 110 generates a seed using the 128-bit output data value of the randomization function and the 128-bit data value generated from the noise generator 108, as shown in FIG. The seed generator 110 uses 64-bit O1 to O6 using 128-bit noise data and 128-bit randomization function (RandomFnt) output data applied from the noise generator 108 as shown in Equation 1 below. Generates seed values up to.

(Noise || state) = (I1 || I2 || I3 || I4)

→ (O1 || O2) = RandomFnt (I1 || I2, key)

→ (O3 || O4) = RandomFnt (I2 O2 || I3, key)

→ (O5 || O6) = RandomFnt (I3 O4 || I4, key)

Then, the seed generator 110 selects only three 64-bit O1, O3, and O5, and provides O1 as two 32-bit constant values used for the key generator 112 below, and the O3 and O5 seed values. The bit string operation value of is output as new 128-bit seed data (Seed) as shown in [Equation 2] below.

Seed = (O3 || O5)

The key generation unit 112 is a software module for generating key data consisting of four 32-bit words to be used in the pseudorandom number generation module 104. As shown in FIG. 2, the KF function unit 200 and 202 are 64-bit inputs. ) And a linear transformation unit 204.

3 illustrates an internal block configuration of the KF function unit 200. Hereinafter, the operation of the KF function will be described with reference to FIG. 3. First, when the keywords A ^r and B ^r of the 32-bit word generated and input from the key generator 112 are input to the KF function unit 200, the A ^r word is generated by the first adder 302. After addition to the 32-bit constant C1 generated in the unit 112, the modulo operation is performed to 2 ³² and output.

Then, the A word ^r is input to claim 1 Sbox (Substitution box) using Sbox transformed B ^r words in 300, after performing a DDR (Date Dependant Rotation) operation, the 2 Sbox (304). The DDR operation refers to an operation that consists of four bits of four MSBs (Most Significant Bits) of the four-byte result data obtained by Sbox converting the B ^r word, and then rotates the left four bits by the configured four bits. Following the DDR performs the operation A ^r word is output to the second is converted back Sbox in Sbox (304) KF function to convert the A ^r, word values.

On the other hand, the word B ^r is non-linearly transformed in the first Sbox 300, and then applied to the multiplier 306 to be multiplied by an A ^r word value Sbox-converted in the second Sbox 304. In the adder 308, the constant C2 is added, modulated by 2 ³² , and output as a KF function-converted B ^r, word value. As described above, the KF function-converted keyword (A ^r ,, B ^r, ) is linearly transformed through the linear transform unit 204 and generated as a final keyword (A ^{r + 1} , B ^{r + 1} ) to generate a pseudorandom number. Is applied to the generation module 104.

The pseudo-random number generating module 104 is composed of two random function value converting units 114 and 116 as shown in FIG. 1, and is generated from the key generating unit 112 as shown in Equation 3 below. A pseudo-random number is generated from a random function by using an applied key value and a state value as input values of the random function value converters 114 and 116.

Rand1 = RandomFnt (state, key)

Rand2 = RandomFnt (Rand1, key)

The second random function value Rand2 is a pseudorandom number output from the pseudorandom number generation module 104, and the first random function value Rand1 is used as an update algorithm input. The state value corresponds to the plaintext of the block cipher algorithm, and is used as an input of a random function (RandomFnt) together with a key and has a size of 128 bits. The size of the keyword value is 128 x round bits. Since the randomization algorithm is basically eight rounds, the size of the key applied to the pseudo-random number generation module 104 is 128 x 8 = 1024 bits. However, since the keyword is generated from a 128-bit seed, the safety of the full search of the randomization function depends on the size of the seed.

4 illustrates an internal functional block configuration of the random function value converter 114. Hereinafter, the operation of the random function value converter 114 in the pseudorandom number generation module 104 will be described in detail with reference to FIG. 4.

That is, the random function value converting unit 114 is formed in a 3-Feistel structure as shown in FIG. 4 and uses a keyword value and a state value generated from the receiving module 102 as input values. It outputs a value (Rand1), the random function value (Rand1) is applied to the second random function value converter 116 as a state value to affect the generation of pseudo-random value that is the second random function value (Rand2) do.

That is, the random function value converter 114 divides the 128-bit state value into 32-bit words A ₀ , B ₀ , C ₀ , and D ₀ as shown in FIG. 4, and inputs the keyword. Value {{K} `_ {r-1} ^ {A}, {K}` _ {r-1} ^ {B}, {K} `_ {r-1} ^ {C}, {K} `_ {r-1} ^ {D}) with using Feistel conversion r case of displaying the randomized intermediate value outputted round after treatment with a _r, B _r, C _r, D _r the randomization function median a _r , B _r , C _r , and D _r can be expressed as in Equation 4 below.

5 illustrates an internal functional block configuration of the Feistel function unit 400. Hereinafter, the operation of the Feistel function will be described with reference to FIG. 5. As shown in FIG. 5, the Feistel function includes two key exclusive OR operations through two eXclusive OR gates 500 and 502, and a first and second four Sboxes. It is formed of a two-round SPN structure consisting of two Sbox transformations through the Sbox units 504 and 506 and one linear transformation through the linear transformation unit 508. The F functions used for odd and even words are called F1 and F2, respectively, and use different Sboxes and linear transformations. However, the keyword value and Sbox inputted to the exclusive ora 500 and 502 are repeatedly used.

In this case, the Sbox is a non-linear transformation as described above. In the present invention, the Sbox is designed to be safer according to the following three design principles in order to solve the vulnerability of the Sbox.

(A) Design GF (2 ⁸ ) using m (x) = x ⁸ + x ⁶ + x ⁵ + x ⁴ +1 (corresponding to 171 _x )

(B) transformation x ^-1 = x ²⁵⁴ on galoache GF (2 ⁸ )

GF (2 ⁸ ) → GF (2 ⁸ ): x x ^-1

(C) again apply affine transformation on {0,1} ⁸

In this case, the affine transformation is used to remove the fixed point, and the form is y = ax when a is an 8 × 8 regular matrix. has type b The matrix a and the constant b are as shown in Equation 5 below.

= ------- S1

= ------- S2

The generated Sbox is as shown in [Table 1] and [Table 2] below.

114, 143, 38, 213, 172, 138, 129, 211, 69, 10, 121, 91, 95, 76, 88, 203, 224, 92, 32, 154, 84, 98, 161, 176, 104, 174, 214, 83, 39, 190, 192, 128, 3, 141, 238, 246, 117, 160, 65, 79, 78, 180, 74, 53, 186, 146, 20, 247, 250, 162, 85, 202, 7, 177, 1, 175, 58, 227, 109, 50, 230, 235, 201, 8,22, 57, 54, 185, 122, 52, 226, 72, 61, 134, 44, 100, 41, 17, 80, 152, 198, 4, 221, 243, 15, 168, 18, 234, 164, 193, 225, 82, 97, 119, 116, 23,139, 206, 60, 191, 216, 0, 86, 200, 223, 46, 130, 25, 6, 71, 195, 153,253, 244, 133, 99, 165, 68, 93, 157, 218, 56, 37, 9, 208, 101, 48, 11, 113, 167, 123, 125, 148, 207, 34, 124, 210, 64, 132, 87, 19, 205, 31, 75,178, 35, 16, 106, 28, 252, 147, 212, 67, 197, 62, 209, 135, 245, 81, 156,63, 228, 89, 66, 33, 187, 189, 24, 127, 194, 140, 131, 184, 233, 179, 21,229, 47, 112, 142, 149, 43, 151, 2, 204, 40, 45, 173, 171, 183, 231, 51,239, 137, 159, 163, 36, 59, 251, 111, 126, 90, 144, 1 05, 94, 169, 70, 145, 49, 219, 12, 248, 217, 249, 158, 155, 73, 232, 240, 13, 96, 136, 199, 215,196, 27, 242, 55, 150, 241, 220, 108, 115, 254, 118, 26, 120, 102, 14, 103, 110, 236, 237, 255, 42, 30, 166, 170, 222, 188, 5, 107, 77, 29, 182, 181

163, 102, 4, 185, 62, 241, 30, 214, 167, 136, 76, 184, 131, 28, 64, 166,233, 123, 107, 150, 108, 196, 190, 117, 135, 106, 65, 175, 168, 13, 73, 178,51, 50, 145, 225, 96, 18, 156, 228, 72, 78, 115, 155, 54, 129, 151, 77,205, 70, 192, 10, 8, 217, 103, 198, 79, 17, 16, 88, 134, 6, 242, 220, 195, 183, 99, 206, 180, 55, 189, 39, 140, 221, 71, 171, 208, 0, 87, 194, 38, 240, 174, 118, 31, 5, 248, 170, 41, 229, 69, 3, 60, 52, 204, 111, 93, 49, 32, 161, 57, 203, 56, 94, 250, 19, 230, 23, 164, 243, 177, 110, 14, 181, 37, 104, 133, 11, 215, 85, 109, 27, 252, 112, 46, 7, 12, 36, 91, 209, 24, 119, 238, 157, 63, 219, 122, 48, 137, 148, 84, 201, 120, 223, 33, 147, 172, 211, 128, 162, 45, 21, 200, 222, 116, 130, 113, 25, 251, 249,216, 210, 236, 100, 199, 154, 245, 187, 35, 29, 158, 74, 98, 82, 141, 59,126, 191, 247, 202, 66, 132, 22, 159, 101, 124, 235, 146, 253, 182, 42, 244,61, 9, 1, 234, 80, 227, 97, 68, 143, 20, 213, 43, 4 7, 169, 95, 121,160, 193, 231, 153, 149, 53, 173, 58, 139, 254, 142, 75, 144, 165, 138, 237,114, 67, 218, 207, 186, 34, 2, 188, 15, 246, 152, 239, 224, 255, 179, 83,232, 197, 105, 90, 40, 212, 125, 81, 86, 89, 92, 127, 176, 44, 26, 226

That is, the data converted into the values shown on the table in the Sbox through the primary Sbox as described above is input to the 4 × 4 linear transformation unit 508 and linearly transformed as shown in Equation 6 below. Outputted as function values L1 and L2.

Finally, the update module 106 inputs the second random function value Rand2, which is an output value from the second random function value conversion unit 116 of the pseudorandom number generation module 116, as a state value, and receives the read. The first random function value Rand1, which is an output value from the first random function value conversion unit 114 in the pseudo random number generation module 104, is used to generate the keyword value generated from the key generation unit 112 in the generation module 102. The third random function value converting unit 118 is configured as a third random function value converting unit 118 using another keyword value obtained by performing an exclusive ora operation. Rand3) is provided to the seed generator 110 in the receiving module 102, and is also provided to the first random function value converter 114 in the pseudorandom number generation module 104. Accordingly, as shown in FIG. 1, the seed generator 110 generates new 64-bit bits O1 through 64-bit using the Rand3 value and the keyword value generated from the key generator 112 in the receiving generation module 102. By generating the seed value up to O6 it is possible to maintain the robustness from the attack from the outside, and the first random function value conversion unit 114 in the pseudo random number generation module 104 is the state of the newly updated Rand3 value (state) By inputting the feedback as the data value to be input as), a new random function value (Rand3) having a high randomness that is hardly related to the first random function value (Rand1) can be obtained.

That is, since the randomized function value (Rand2) generated from the pseudorandom number generation module can be easily exposed to the attacker, if the keyword value is exposed to the attacker without newly updating the keyword value, the pseudorandom number generation module 104 is then performed. And all output values of the randomization function of the update module 106 can be calculated.

At this time, the exclusive OR 120 operation performs an exclusive OR operation on the keyword 128 bits used in each round and the output value Rand1 of the first random function value conversion unit 114 in Equation 7 below. To update the keyword,

Rand1

state ← RandomFnt (Rand2, key)

The pseudorandom number generation module 104 of the pseudorandom number generation module 104 is used as the output value Rand3 of the third random function value conversion unit 118 that receives the updated keyword and the output value Rand2 of the second random function value conversion unit 116. The state value of one random function value converting unit 114 is updated.

Meanwhile, in the above description of the present invention, specific embodiments have been described, but various modifications may be made without departing from the scope of the present invention. Therefore, the scope of the invention should be determined by the claims rather than by the described embodiments.

As described above, the pseudorandom number generator proposed by the present invention uses a key value used for generating pseudorandom numbers instead of writing a timestamp on one key like a random number generator using a block cipher proposed in ANSI X9.17. An update algorithm for updating has an advantage of improving stability by changing a key value whenever a random number is generated. In the present invention, there is an advantage in that the efficiency can be improved by updating a previously generated key through a simple logical operation, instead of regenerating the key in changing the key value, IC card in a limited environment as well as a PC This has the advantage of being applicable to all the same various platforms.

Claims (30)

In pseudo-random number generator using block cipher with SPN structure, A receiving generation module for collecting a noise suitable for each platform and generating a key value to be used as an input of a random function when a similar random number is generated from the noise information; Two random function value converters, and generate a first random function value (Rand1) using the key value and a state value as input values of a first random function value converter, and generate the first random function value and And generating a second random function value (Rand2) by using a key value as an input value of the second random function value conversion unit, and outputting the second random function value as a pseudo random value. Pseudo-random number generator. In pseudo-random number generator using block cipher with SPN structure, A receiving generation module for collecting a noise suitable for each platform and generating a key value to be used as an input of a random function when a similar random number is generated from the noise information; Two random function value converters, and generate a first random function value (Rand1) using the key value and a state value as input values of a first random function value converter, and generate the first random function value and A pseudorandom number generation module for generating a second random function value (Rand2) by using a key value as an input value of a second random function value conversion unit, and outputting the second random function value as a pseudorandom value; Generates a third random function value (Rand3) using the second random function value generated from the pseudorandom number generation module and the updated new key value as an input value of a random function value conversion unit, and generates a random number similar to the receiving generation module And an update module provided to the module to update and change state information values of the key value and the random function value conversion unit. The method of claim 2, The receiving generation module includes: a noise generator for collecting and providing noise to be used according to each platform; A seed generator for generating a seed value from the noise information; And a key generator for generating a key value to be used as an input value in a random function value converter for generating pseudorandom numbers from the seed value generated by the seed generator. The method of claim 3, The seed generator may separate the noise data and the state data last updated from the update module into two word data having the same bit length, and then, the key value and the four values. A pseudo-random number generator, characterized by generating seed values (O1 to O6) through the operation of a random function value that takes word data as input. [Equation] (Noise || state) = (I1 || I2 || I3 || I4) → (O1 || O2) = RandomFnt (I1 || I2, key) → (O3 || O4) = RandomFnt (I2 O2 || I3, key) → (O5 || O6) = RandomFnt (I3 O4 || I4, key) Noise: Noise Data state: state data last updated from the update generation module I1, I2: Word data divided by equal length of noise data I3, I4: Word data divided by equal length of status data RandomFnt: Random Function The method of claim 4, wherein The seed generator generates a pseudo random number generator, characterized in that the 128-bit noise data and the state data are divided into four word data in 64-bit units and applied as an input value of the random function for generating a seed value. . The method of claim 4, wherein The seed generator is a pseudo-random number generator, characterized in that the O3 and O5 seed value of the generated seed value is bit-linked as shown in the following equation to use the new 128-bit seed data (Seed). [Equation] Seed = (O3 || O5) The method of claim 4, wherein The seed generator divides the seed value O1 into equal bit lengths and provides the seed generator O1 as a constant value to a key generation algorithm of the key generator. The method of claim 7, wherein The seed value O1 is 64-bit data and is divided into two word data in 32-bit units for providing to the key generator. The method of claim 3, The key generation unit includes: a KF function conversion unit for dividing the new seed value generated from the seed generation unit into four word data having the same bit length, and converting each word data into a KF function value; And a linear converter for linearly converting the KF-converted word data and outputting the input key value of a random function for generating a pseudo random number. The method of claim 9, The KF function conversion unit adds the seed constant value C1 to one word data input to one input terminal among two word data input to the KF function conversion unit, and modulates the result to 2 ³² . A first adder for outputting a value; A first sbox unit for sbox converting the other word data and outputting the same; A DDR unit for extracting an MSB 4 byte of 4-byte data configured as an output from the first Sbox unit and outputting the output of the first adder by 4 bits in a direction preset by a user; A second Sbox unit for converting the word data shifted through the DDR unit into one KF function-converted data value; A multiplication unit for multiplying Sbox-converted word data from the first and second Sbox units; A second adder configured to add another seed constant value C2 to word data output from the multiplier, and then output a result value obtained by performing a modulo operation to another ³² as another KF function converted data value; Pseudo random number generator comprising a. The method of claim 2, The pseudo-random number generating module generates a first random function value (Rand1) using the four key data generated from the receiving generation module and the state data last updated from the key update module as input values. Wealth; And a second random function unit configured to generate a second random function value (Rand2), which is a pseudo random value, using the first random function value and the key value as input values. The method of claim 2, The key value is pseudo-random number generator, characterized in that for updating the new key value through an XOR operation with the first random function value (Rand1). The method of claim 2, The random function value conversion unit, pseudo-random number generator, characterized in that consisting of a 3-Feistel structure. The method of claim 11, Each word data A _r , B _r , C _r , and D _r of the random function value generated through the random function operation is divided into four word data of 32 bits as shown in the following equation. A pseudo-random number generator, characterized in that output as a value of the Feistel function with the input value and key value. [Equation] F: Feistel function The method of claim 14, The Feistel structure, pseudo-random number generator, characterized in that implemented in a two-round SPN structure. In the pseudo random number generator using a block cipher having an SPN structure including a receiving generation module, a pseudo random number generation module, and an update module, (a) generating a seed value from the noise information; (b) generating a key value using the seed value; (c) calculating a first random function value (Rand1) from a random function having the key value and the last updated state value from the update module as input data values; (d) calculating a second random function value (Rand2) output from the random function using the first random function value (Rand1) and the key value as an input data value; and (e) outputting the value generated as the second random function value (Rand2) as a pseudo random value. In the pseudo random number generator using a block cipher having an SPN structure including a receiving generation module, a pseudo random number generation module, and an update module, (a) generating a seed value from the noise information; (b) generating a key value using the seed value; (c) calculating a first random function value (Rand1) from a random function having the key value and the last updated state value from the update module as an input data value; (d) calculating a second random function value (Rand2) output from the random function using the first random function value (Rand1) and the key value as an input data value; (e) outputting a value generated as the second random function value (Rand2) as a pseudo-random value; (f) generating an update random function value (Rand3) from a random function having the second random function value (Rand2) and the updated new key value as an input data value to update the seed value and the state value of the random function; A pseudorandom number generating method comprising the step; The method of claim 17, In the step (f), the key value is similar random number generating method, characterized in that for updating the new key value through an Exclusive OR (XOR) operation with the first random function value (Rand1). The method of claim 17, Step (a) may include: (a1) dividing the noise data and the last updated state value from the update module into two word data having the same bit length; (a2) generating a seed value through a calculation of a random function using the key value and the separated four word data as input values. The method of claim 19, In step (a2), the generated seed values (O1 to O6) are generated by a random function operation using the key value and the four word data as input values as in the following equation. How it occurs. [Equation] (Noise || state) = (I1 || I2 || I3 || I4) → (O1 || O2) = RandomFnt (I1 || I2, key) → (O3 || O4) = RandomFnt (I2 O2 || I3, key) → (O5 || O6) = RandomFnt (I3 O4 || I4, key) Noise: Noise Data state: state data last updated from the update generation module I1, I2: Word data divided by equal length of noise data I3, I4: Word data divided by equal length of status data RandomFnt: Random Function The method of claim 20, The noise data and the state data are generated in 128 bits, each of which is divided into two word data in units of 64 bits, and is input as an input value of the random function. The method of claim 20, The O3 and O5 seed values of the generated seed values are used as new 128-bit seed data by bit connection as in the following equation. [Equation] Seed = (O3 || O5) The method of claim 20, The O1 seed value is divided into two word data in 32-bit units and is provided as a constant value when a key generation algorithm is performed. The method of claim 17, Step (b) comprises: (b1) dividing the seed value into four word data having the same bit length; (b2) converting each word data into a KF function; (b3) generating a key value of a random function for generating a similar random number by linearly converting each word data converted by the KF function, and proceeding to the pseudo random number generation method. The method of claim 24, Step (b2) may include: (b21) receiving two word data of the four data as an input value; (b22) adding the seed constant value C1 to one word data of the two word data, and then performing modulo operation with 2 ³² ; (b23) converting the other word data into a first Sbox; (b24) extracting four bits of MSB from the modulated word data and the first Sbox-converted four-byte word data and performing a four-bit raft shift; (b25) converting the raft-shifted word data into second K-box transformed data as one KF function-transformed data value; (b26) multiplying the first Sbox-converted word data by the second Sbox-converted word data, adding the other constant value C2, and performing a modulo operation with 2 ³² to convert another KF function Outputting a data value; and generating a pseudo random number. The method of claim 17, The first random function value (Rand1) is a pseudo-random number generation method, characterized in that it is generated by a random function operation of the 3-Feistel structure with the key value and the last updated state value from the update module as an input value. The method of claim 17, The second random function value (Rand2) is a pseudo-random number generation method, characterized in that it is generated by a random function operation of the 3-Feistel structure with the key value and the first random function value (Rand1) as an input value. The method of claim 17, The third random function value Rand3 is generated through a random function operation of a 3-Feistel structure using the updated key value and the second random function value Rand2 as input values. Way. The method according to any one of claims 26 to 28, The Feistel function is a pseudo random number generation method, characterized in that implemented in a two-round SPN structure. The method according to any one of claims 26 to 28, Each word data A _r , B _r , C _r , D _r of a random function value generated through the random function operation is divided into four word data in 32-bit units as shown in the following equation. A pseudorandom number generating method characterized by outputting a Feistel function value having a value and a key value as input values. [Equation] F: Feistel function