Method and apparatus for improving entropy of random numbers
Patent Information
- Application Number
- CN202410573381.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-05-12
- Filing Date
- 2024-05-10
- Publication Date
- 2024-11-12
AI Technical Summary
The prior art is difficult to generate a random number with sufficiently high entropy in cryptographic applications and cannot meet high security requirements.
By combining multiple random processes, the random numbers are mapped and linked using S-box and XOR logic operations to form combined random numbers and increase the entropy value. Specific methods include optimizing the entropy of combined random numbers using noise sources, dual-surface S-boxes, compressed S-boxes, and digitized converters.
The entropy value of random numbers is significantly improved, and cryptographic applications that meet high security requirements are used to achieve precise quantity improvement and cost optimization of entropy.
Smart Images

Figure CN118944862A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method and a device for improving the entropy of random numbers. Background Art
[0002] The entropy per bit of a perfect random number is 1 (100%). In many applications, especially cryptography, it is necessary to have as good a random number as possible, i.e. a random number whose entropy is as close to 1 as possible. For example, a noise source (also called a noise generator) is used to generate random numbers. The prior art knows many possible implementations for a noise source, such as an analog or digital noise generator.
[0003] So-called S-boxes are known as basic building blocks of cryptography: An S-box is a substitution operation for converting one binary number into another binary number. Such a mapping can be, for example, reversible (bijective) or compressive. Summary of the invention
[0004] The aim here is to increase the entropy of the random numbers in order to meet high cryptographic security requirements.
[0005] The object is achieved according to the features of the invention. Preferred embodiments can be found in particular in this context.
[0006] The examples presented herein may be based on at least one of the following solutions. In particular, the following combinations of features may be used in order to achieve the desired result. Features of the method may be combined with any feature(s) of the device or vice versa.
[0007] In order to solve this problem, a method for improving the entropy of random numbers is proposed.
[0008] - where random numbers are provided for multiple random processes respectively,
[0009] - wherein a plurality of random numbers are linked to form a combined random number, wherein at least one of the random numbers was previously mapped onto another random number.
[0010] The mapping of one random number to another random number can include a permutation or rearrangement.
[0011] An improvement is to use combined random numbers in cryptographic methods.
[0012] A further development is that each stochastic process includes at least one physical stochastic process.
[0013] Basically, any mechanism that provides seemingly quite arbitrary (random) numbers can be used as a random process. Random numbers can be numbers with a more or less high entropy. Ultimately, such random numbers can also be generated in a deterministic manner or by a (pseudo)random generator. Since the entropy increase is achieved by combining and mapping multiple random numbers, lower requirements can optionally be placed on the entropy of the individual random numbers.
[0014] One development consists in that the random number is mapped into another random number by means of at least one S-box.
[0015] An improvement is that the S-box is a bijective S-box or a compressed S-box.
[0016] A refinement is that the random process and the digitizer are part of the noise source.
[0017] One development consists in mapping the random number into another random number by means of a digitizer.
[0018] One development consists in that a plurality of random numbers are linked to form a combined random number by means of at least one XOR logic operation, in particular by means of at least one XOR gate.
[0019] The XOR logic operation can be implemented by means of an XOR gate or other logic gates. In particular, an XOR logic operation of more than 2 bits can be divided into a plurality of 2-bit XOR logic operations.
[0020] For example, an implementation can be chosen which links the noise sources by means of XOR gates. The advantage of using an XOR logic operation is that an exact quantitative conclusion and documentation of the entropy increase is possible.
[0021] One development consists in determining a mapping of at least one random number to further random numbers which leads to an increase in the entropy of the combined random number.
[0022] One development consists in determining a mapping of at least one random number to further random numbers which leads to an increase in the entropy of the combined random number by testing a plurality of mappings.
[0023] Such a test can be performed, for example, by means of a computer search: thus, all existing S-boxes (or all subsets) can be tried once and the entropy of the combined random number can be determined for each selected S-box or set of S-boxes. Now, at least one S-box can be selected from the best S-boxes, which can be realized most simply or most advantageously (with respect to at least one target variable, such as space, cost, time, expenditure, etc.).
[0024] As a solution, a device for improving the entropy of random numbers is also proposed, wherein the device has a processing unit configured to execute the steps of the method according to any one of the above embodiments.
[0025] Furthermore, a device for increasing the entropy of random numbers is proposed.
[0026] - having a plurality of random generators, each of said random generators providing a random number,
[0027] at least one mapping unit, by means of which at least one of the random numbers is mapped to at least one further random number,
[0028] - a combining unit, said combining unit being coupled to the random generator and to at least one mapping unit,
[0029] Such that a combined random number can be determined based on the random number and at least one further random number.
[0030] Each random generator may correspond to a random process or include at least one random process.
[0031] In particular, one option is that for a random generator coupled to a mapping unit, the random numbers of the mapping unit are processed by the combination unit instead of the random numbers determined by the random generator. If a mapping unit is provided for each random generator, only the mapped random numbers can be processed in the combination unit.
[0032] A further development is that the mapping unit comprises at least one bijection and / or at least one compressed S-box.
[0033] One option is to design the mapping unit in multi-stage or to implement a multi-stage mapping, optionally with XOR gates connected in between.
[0034] In one development, the combination unit comprises at least one XOR logic operation.
[0035] An improvement is that the random generator and the mapping unit are part of the noise source.
[0036] The mapping unit can in this case be referred to as a digitizer. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] The characteristics, features and advantages of the present invention described above and the manner and method of achieving the characteristics, features and advantages are described below in conjunction with the schematic description of an embodiment, which is explained in detail in conjunction with the drawings. For the sake of overview, the same or identically functioning elements may be provided with the same reference numerals.
[0038] Figure 1Shown is a combination of multiple noise sources connected together via XOR gates.
[0039] Figure 2 Shown is a combination of multiple noise sources connected together via XOR gates along with S-boxes.
[0040] Figure 3 An exemplary arrangement is shown for determining a combined random word by means of two noise sources connected together via an XOR gate.
[0041] Figure 4 An exemplary arrangement is shown for determining a combined random word by means of two paths connected together via an XOR gate, wherein each path comprises a noise source and an S-box.
[0042] Figure 5 An exemplary arrangement for increasing entropy by means of two noise sources and a bijective S-box is shown.
[0043] Figure 6 An exemplary arrangement for increasing entropy by means of two noise sources and two compression S-boxes is shown.
[0044] Figure 7 Show according to Figure 6 Part of an arrangement of, in which the S-boxes are implemented, for example, by means of XOR gates.
[0045] Figure 8 A table with exemplary probability distributions for a number of noise sources is shown.
[0046] Fig. 9 An exemplary arrangement for increasing entropy with the aid of four noise sources and a plurality of bijection and compression S-boxes is shown.
[0047] Fig.10 An exemplary arrangement for increasing the entropy by means of a variant of a digitizer provided in the noise source is shown. DETAILED DESCRIPTION
[0048] Figure 1 It is shown that there are b≥2 noise sources RQ1, RQ2, ..., RQ b In the arrangement, the noise sources are connected to each other via an XOR gate 101.
[0049] Noise sources RQ1, RQ2, ..., RQ b Random words (also called random vectors or random numbers) with a width of n bits are generated independently of one another, each noise source having a specific entropy.
[0050] The processor unit (CPU) generates b random words per time unit, for example, per clock, wherein each random word has a value corresponding to each noise source RQ1, RQ2, . . . , RQb The combination of the noise sources via the XOR gate 101 causes an entropy increase (i.e., entropy compression in the random word). By using S-boxes S1, S2, ..., S b , which can further increase the entropy.
[0051] Figure 2 An exemplary arrangement is shown, which is based on Figure 1 In the noise sources RQ1, RQ2, ..., RQ b The path between the XOR gate 201 and the XOR gate 202 has an S-box RQ1, RQ2, ..., RQ b .
[0052] The example described in this article determines the following S-box in the combination of noise sources, which increases the entropy of the combined random words. The entropy increase can be an optimization of entropy, especially maximization. For example, from the set of S-boxes, at least one S-box can be determined (and used) that contributes most to the entropy increase. Multiple S-boxes can also be used in this sense, such as two S-boxes that cause the maximum increase in entropy (e.g., with respect to a preset (maximum) implementation cost).
[0053] Consider the following situation as an example:
[0054] Case 1 Bijective S-box: m=n applies, the number of bits at the input of the S-box is equal to the number of bits at its output.
[0055] Case 2 Compressed S-box: m<n applies, the number of bits at the input of the S-box is greater than the number of bits at its output, and a reduction or compression of bits occurs.
[0056] The b noise sources of the combination of noise sources work in particular independently of each other. A single noise source is memoryless or has a temporary memory. In a noise source without memory, the random words generated at time t are (statistically) independent of all previously generated random words, i.e., especially statistically independent of the random words generated at time t-1. In a noise source with a short memory, the random words generated at time t are related to a small number of process words. A noise source with a temporary memory can be modeled by a Markov chain. In the Markov chain model, a static probability distribution is first calculated and then used as the basis for further analysis. Therefore, the situation of a noise source with a temporary memory can be attributed to the situation of a noise source without memory.
[0057] Bijective S-box
[0058] The bijective function is entropy conserving, i.e. the random words before and after the bijective S-box have the same entropy. However, the presence of the bijective S-box in the noise source combination affects the entropy value of the combined random number after the XOR gate.
[0059] This is explained exemplarily based on a noise source combination, which includes two noise sources, each of which generates a 2-bit wide random word. These are therefore two memoryless noise sources RQ1 and RQ2, which generate 2-bit wide random words X and Y per CPU clock. The generated 2-bit random words can take four different values:
[0060]
[0061] These four values are adopted with specific, optionally constant (and generally different) probabilities. From these probabilities, the probability for XOR and , from which the entropy H(Z) of the combined random word is then derived.
[0062] Figure 3 An exemplary arrangement is shown with two noise sources RQ1 and RQ2 (without S-boxes), each of which supplies 2-bit wide random words X and Y to the input of XOR gate 301. At the output of XOR gate 301, a 2-bit wide random word is generated per clock cycle, which can assume values from 0 to 3. The associated probabilities W0 to W3 result as follows:
[0063] W0=Pr(Z=0)
[0064] =Pr(X=0)·Pr(Y=0)+Pr(X=1)·Pr(Y=1)+Pr(X=2)·Pr(Y=2)+Pr(X=3)·Pr(Y=3);
[0065] W1=Pr(Z=1)
[0066] =Pr(X=0)·Pr(Y=1)+Pr(X=1)·Pr(Y=0)+Pr(X=2)·Pr(Y=3)+Pr(X=3)·Pr(Y=2);
[0067] W2=Pr(Z=2)
[0068] =Pr(X=0)·Pr(Y=2)+Pr(X=1)·Pr(Y=3)+Pr(X=2)·Pr(Y=0)+Pr(X=3)·Pr(Y=1);
[0069] W3=Pr(Z=3)
[0070] =Pr(X=0)·Pr(Y=3)+Pr(X=1)·Pr(Y=2)+Pr(X=2)·Pr(Y=1)+Pr(X=3)·Pr(Y=0).
[0071] In this case, Pr(K=l) represents the probability that a random word K takes the value l.
[0072] The entropy of a random word Z derived from probabilities W0, W1, W2, and W3 is:
[0073]
[0074] Figure 4 based on Figure 3 An exemplary arrangement with bijective S-boxes S and T in the path between the noise source and the XOR gate 401 is shown. The noise source RQ1 supplies a 2-bit wide random word X to the input of the S-box S, via the output of which a modified 2-bit wide random word X' is supplied to the first input of the XOR gate 401. The noise source RQ2 supplies a 2-bit wide random word Y to the input of the S-box T, via the output of which a modified 2-bit wide random word Y' is supplied to the second input of the XOR gate 401. At the output of the XOR gate 401, a 2-bit wide random word Z' is generated per clock, which can take values from 0 to 3. The associated probabilities V0 to V3 result in:
[0075] V0=Pr(Z′=0)
[0076] =Pr(X=S -1 (0))·Pr(Y=T -1 (0))+Pr(X=S -1 (1))·Pr(Y=T -1 (1))+Pr(X=S -1 (2))·Pr(Y=T -1 (2))+Pr(X=S -1 (3))·Pr(Y=T -1 (3));
[0077] V1=Pr(Z′=1)
[0078] =Pr(X=S -1 (0))·Pr(Y=T -1 (1))+Pr(X=S -1 (1))·Pr(Y=T -1 (0))+Pr(X=S -1 (2))·Pr(Y=T -1 (3))+Pr(X=S -1 (3))·Pr(Y=T -1 (2));
[0079] V2=Pr(Z′=2)
[0080] =Pr(X=S -1(0))·Pr(Y=T -1 (2))+Pr(X=S -1 (1))·Pr(Y=T -1 (3))+Pr(X=S -1 (2))·Pr(Y=T -1 (0))+Pr(X=S -1 (3))·Pr(Y=T -1 (1));
[0081] V3=Pr(Z′=3)
[0082] =Pr(X=S -1 (0))·Pr(Y=T -1 (3))+Pr(X=S -1 (1))·Pr(Y=T -1 (2))+Pr(X=S -1 (2))·Pr(Y=T -1 (1))+Pr(X=S -1 (3))·Pr(Y=T -1 (0)).
[0083] Although the modified random words X' and Y' contain the same entropy as the original random words X and Y, the random words With random words The different probabilities assume four possible values 0 to 3. This results in a new entropy value for Z' (relative to Z).
[0084] The entropy of is determined as follows:
[0085]
[0086] The selection of the S-boxes S and T used has an influence on the entropy H(Z'). For example, the goal is to determine the following S-boxes S and T for which H(Z') is the largest. This can be achieved in the following way, that is, for all possible pairs of S-boxes (S, T), the corresponding entropy H(Z') is determined. Now, for example, it is possible to select the pair (S, T) for which the entropy H(Z') takes the maximum possible value. In addition, in the process of multi-objective optimization, at least one objective variable can be considered together: for example, only the following S-boxes can be considered, for which the implementation cost does not exceed a preset threshold, or it can be determined that, in addition, the implementation cost should be as small as possible, so that a weighted optimization between the goal "entropy improvement" and the goal "cost reduction" can be performed. In general, different optimization strategies are possible, which cause the selection of suitable S-boxes. For example, a suboptimal entropy value will be sufficient when it has met a specific preset and / or when the cost for the corresponding S-box exceeds a preset threshold.
[0087] Example: Two Noise Sources, a Bijective S-Box
[0088] Given two memoryless noise sources, the noise sources generate 3-bit wide random words (i.e., 8 different symbols 0 to 7) with the following probabilities (pi is the probability of symbol i):
[0089] p0=0,001, p1=0,02, p2=0,05, p3=0,08,
[0090] p4=0,16, p5=0,17, p6=0,20, p7=0,319.
[0091] These two noise sources produce random words with 24783 bits of entropy or an entropy value of 82.62%.
[0092] The goal is now to increase the entropy of the combined random numbers. This is achieved, for example, with the aid of a unique S-box.
[0093] Figure 5 An exemplary arrangement with two noise sources RQ1 and RQ2 is shown, wherein an S-box 502 is arranged in the path between the noise source RQ1 and the XOR gate 501. The S-box 502 is denoted by S in this example. This is exemplarily a nonlinear S-box, which is determined as follows:
[0094]
[0095] The noise source RQ2 is directly connected to the XOR gate 501. At the output of the XOR gate 501 a combined random number is provided.
[0096] As explained above by way of example, the two noise sources RQ1 and RQ2 provide a 3-bit wide random number with an entropy value of 82.61% per noise source. Figure 5 The linking shown in with the S-box 502 enables the combined random number at the output of the XOR gate 501 to have an entropy of 29956 bits or an entropy value of 99.85% (compared to 28133 bits or 93.78% if the S-box 502 is omitted).
[0097] The algebraic description of the S-box 502 exemplarily described here is determined as follows:
[0098]
[0099] where x, y and z are given by three bits at the input of S-box 502 .
[0100] Compressed S-box
[0101] A compressed S-box is a function that maps a binary word of length n to a binary word of length m, where m<n applies. By using a compressed S-box in a noise source combination, an additional compression of the entropy in the final (combined) random word can be achieved. The final random word then has a sufficiently high entropy per bit in many cases, so that the compressed S-box can dispense with post-processing for further entropy increase.
[0102] By saving on post-processing, the costs in the hardware implementation of the random number generator are reduced. It is also advantageous that the random numbers can be generated at a high data rate (ie more random numbers per unit time or in a shorter time).
[0103] In particular, it is proposed hereby to use improved or optimized compressed S-boxes in the noise source combination, so that the entropy of the combined random words is already sufficiently high and post-processing can also be dispensed with for the reasons described.
[0104] Sufficiently high can in particular mean that the entropy value of the combined random word meets, for example, official, customer or standard specifications. In many cryptographic applications, it is common to describe the quality of a random number based on an entropy value that must be at least achieved.
[0105] Example: Two noise sources, two compressed S-boxes
[0106] Figure 6 Show that the probability distribution of Figure 5 There are two noise sources RQ1 and RQ2 in the Figure 5 In contrast, an S-box S602 is provided in a path between the noise source RQ1 and the XOR gate 601, and another S-box T603 is provided in a path between the noise source RQ2 and the XOR gate 601. Each of the S-boxes S and T compresses a 3-bit word into a 2-bit word.
[0107] Each of the two noise sources RQ1 and RQ2 generates a 3-bit long random word with an entropy of 24783 bits or an entropy value of 82.61%.
[0108] The S-box is illustratively defined as follows:
[0109] and
[0110] By Figure 6 With the arrangement shown in FIG. 6 , the entropy of the combined random words provided by XOR gate 601 is increased to 199,984 bits or 99.992%.
[0111] The S-boxes S and T are linear in this example and can be shown as a 2×3 matrix as follows:
[0112] and
[0113] Figure 7 An exemplary arrangement of S-boxes 602 and 603 by means of XOR gates is shown, which results directly from the 2×3 matrix shown above.
[0114] In the noise source combination, a combination of bijection and compressed S-boxes can also be used. In addition, the individual noise sources can have different probability distributions if, for example, noise sources of the same structure are operated with different system parameters (e.g. different oscillators with different frequencies). It is also possible to use noise sources of different types or in different structural ways.
[0115] Example: Four different noise sources
[0116] In this example, four different noise sources should be combined with each other to form a noise source combination.
[0117] Figure 8 A table showing an exemplary probability distribution for each of the four noise sources: Each of the noise sources generates eight symbols per unit time, the symbols being denoted 0 to 7. Noise sources RQ1 to RQ4 occupy rows 2 to 5 of the table. For example, the second entry 0.02 in the second row indicates that noise source RQ1 generates symbol 0 with probability p=0.02. Noise source RQ3 generates only six different symbols 1 to 6, symbols 0 and 7 are not generated by RQ3.
[0118] If these noise sources are Figure 8 The probabilities described in generate symbols 0 to 7, then the entropy of the 3-bit random words generated by the corresponding noise source is obtained as:
[0119]
[0120] Here, pi represents the probability of symbol i, where i=0,...,7. Figure 8 The last two rows in give the bit entropy (ie the entropy of the output symbol of the noise source, which is a 3-bit random number or a random vector of length 3 bits) or entropy value (also called percentage entropy value) for the corresponding noise source.
[0121] In the present example, the noise source assembly is combined with an S-box. Fig. 9 An exemplary arrangement comprising the four noise sources RQ1 to RQ4 mentioned here is shown, in Fig. 9 It is also represented by 901 to 904.
[0122] The noise source RQ1 901 provides a 3-bit random word to a bijective S-box S 905 , the output of which conducts a modified 3-bit random word to a first input of an XOR gate 907 .
[0123] Noise source RQ2 902 provides a 3-bit random word to a second input of XOR gate 907 .
[0124] The noise source RQ3 903 provides a 3-bit random word to a bijective S-box T 906 , the output of which conducts a modified 3-bit random word to a first input of an XOR gate 908 .
[0125] Noise source RQ4 904 provides a 3-bit random word to a second input of XOR gate 908 .
[0126] The output of the XOR gate 907 is connected to the input of a compression S-box J 909 which modifies the applied 3-bit random word into a compressed 2-bit random word and forwards it to the first input of an XOR gate 911 .
[0127] The output of the XOR gate 908 is connected to the input of a compression S-box K 910 , which modifies the applied 3-bit random word into a compressed 2-bit random word and forwards it to the second input of an XOR gate 911 .
[0128] The combined 2-bit random word is provided at the output of the XOR gate 911.
[0129] Noise sources RQ1 to RQ4 have Figure 8 The entropy described in . According to Fig. 9 The arrangement of leads to the following stepwise increase in entropy:
[0130] - at the output of XOR gate 907: 94.47%,
[0131] - at the output of XOR gate 908: 98.65%,
[0132] - at the output of S-box J 909: 99.78%,
[0133] - at the output of the S-box K 910: 99.94%, and
[0134] - At the output of the XOR gate 911: 99.99995%.
[0135] The S-box used here is determined exemplarily as follows:
[0136] and
[0137] and
[0138] The algebraic description of the S-box is:
[0139]
[0140]
[0141]
[0142]
[0143] In addition, it should be noted that the method proposed here is largely robust. If, for example, one of the four noise sources fails, the entropy value at the output of XOR gate 911 is still at least 99.9%. This value exceeds the entropy lower limit of 99.7% required by the German Federal Agency for Information Technology.
[0144] A further advantage is that all failures are detected as early as possible in compliance with the standard: Since the combinatorial logic proposed here is memoryless, in the event of a complete failure of all noise sources the same random word is always available at the output of XOR gate 911 , so that all failures are noticed without delay.
[0145] Noise Source with Custom Digitizer
[0146] In the above Figure 4 The entropy is shown compressed into random words at the output of the XOR gate 401 by means of two S-boxes S and T.
[0147]
[0148] The entropy compression is achieved, for example, by using bijective S-boxes S and T (determined, for example, by means of a computer search).
[0149] Another approach consists in appropriately modifying the participating noise sources, in particular their digitizers. The same applies to noise source bundles with more than two noise sources. The solution described is advantageous from an implementation point of view (gate area, circuit logic depth, power consumption).
[0150] The noise source consists of two parts, the actual physical noise source and the digitizer.
[0151] Physical noise sources
[0152] Physical noise sources implement physical random processes. Random events (also called random experiments) occur per unit time (determined by an external clock generator, such as the clock of a processor unit). The results of random experiments are shown by measurement variables ξ, which are elements of the so-called event space Ω: ξ∈Ω.
[0153] The event space Ω is k=2 n , n≥1 sub-areas A0, A1, ..., A k-1The disjoint union of :
[0154] Ω=A0∪A1∪…∪A k-1
[0155] For 0≤i<j≤k-1,
[0156] The output of a random experiment is unpredictable. But for each
[0157] i∈{0, 1, ..., k-1}
[0158] Given the probability P(A i ): The measured variable ξ is in subset A i In. k-tuple
[0159] (P(A0), P(A1), ..., P(A k-1 ))
[0160] For 0≤i≤k-1, P(A i )≥0 and
[0161] P(A0)+P(A1)+…+P(A k-1 )=1
[0162] It is called the probability distribution of the noise source. The probability distribution of the noise source (used in practice) is known.
[0163] If the noise source (as assumed here) is memoryless, then the (Shannon) entropy h is derived from the noise distribution as follows:
[0164]
[0165] The entropy is therefore a pure property of the physical noise source and is independent of the design of the digitizer.
[0166] Digitizer
[0167] ξ∈Ω is the result of a random experiment. Because Ω is divided into k non-overlapping regions A i , there are k different outcomes ξ∈A i The results are expressed as probabilities P(A i ) appears, i=0, 1,…, k-1.
[0168] The digitizer converts k possible outcomes ξ∈A i Each result in is converted into a symbol determined one by one. Therefore, k different symbols are required. The k symbols can also be represented by integers 0, 1, ..., k-1.
[0169] First of all, it is clear which event should be represented by which symbol. One possibility is the following association: the event ξ∈A0 is represented by the symbol 0: if the event ξ∈A0 occurs in the physical noise source, the digitizer generates the symbol 0 and outputs it. If the event ξ∈A1 occurs, the symbol 1 is generated and output. If ξ∈A2 occurs, the symbol 2 is output, and so on.
[0170] The digitizer implemented according to the association rule is called the canonical digitizer for the noise source.
[0171] In principle, k events ξ∈A i , i = 0, 1, ..., k-1, can be associated with k symbols 0, 1, ..., k-1 in any order. Therefore, there exists
[0172] k! = k(k-1)(k-2)...3·2·1
[0173] Possibility to specify a digitizer.
[0174] If, for example, k=8, then there are 8!=40320 possibilities for realizing the digitizer. All these variants result in a noise source with the same entropy.
[0175] Furthermore, π=π(0), π(1), ..., π(k-1) is an arbitrary arrangement (disposition) of k symbols 0, 1, ..., k-1. The digitizer for the arrangement π can be determined as follows: If the event ξ∈A i If π(i) occurs, then the symbol π(i) is generated and output, i=0, 1, ..., k-1.
[0176] Consider the following S-box defined by permutation π:
[0177]
[0178] For all i∈{0, 1, ..., k-1}, π(i)=S(i) holds. The permutation π thus defines a bijective S-box. Conversely, a given bijective S-box
[0179] S: {0, 1, ..., k-1} → {0, 1, ..., k-1}
[0180] Define the permutation π of k symbols 0, 1, ..., k-1: = S(0), S(1), ..., S(k-1).
[0181] There are therefore two variants in particular: Variant (A) is a physical noise source with a standard digitizer and an S-box S connected downstream. Variant (B) comprises a noise source without an S-box, wherein the digitizer of the noise source is arranged according to
[0182] π=S(0), S(1),...,S(k-1)
[0183] accomplish.
[0184] Two variants (A) and (B) for each event ξ∈A i The same output symbol is always produced. In variant (A) the output symbol appears at the output of the S-box S, whereas in variant (B) the output symbol is provided directly by the digitizer of the noise source.
[0185] It is therefore also proposed that in a noise source assembly (such as for example in Figure 1 In the example shown in FIG. 1 , a noise source with a specially defined digitizer can be used. Such a digitizer realizes an arrangement for which the entropy of the combined random words is increased or in particular optimized. A suitable arrangement can be determined in advance by means of a computer search.
[0186] It should be noted here that in the first approach the noise source is provided with an external, real S-box, whereas in the second approach the same S-box is contained internally and virtually in the digitizer.
[0187] For the purpose of explaining the term "digitizer": k = 2 applies n If, for example, we consider the case where k = 16, then n = 4. The physical noise source can adopt 16 different internal states ξ∈A i , 0≤i≤15. The corresponding 16 symbols in the digitizer are determined as 0, 1, ..., 15. In the hardware implementation of the noise source, the symbols are represented by a four-bit long binary vector. For example, the symbol 15 is represented by the vector (1, 1, 1, 1). Therefore, the event ξ∈A 15 Digitized into a vector (1, 1, 1, 1) in a physical noise source.
[0188] Example: Noise Sources with Digitizers
[0189] exist Figure 5 In the example shown in , there are two noise sources RQ1 and RQ2, which use the same physical random process. For each noise source, eight different events ξ∈A can occur. i , 0≤i≤7. The probability of each event occurring is p i =P(A i )=Pr(ξ∈A i) is obtained as follows:
[0190] p0=0,001, p1=0,02, p2=0,05, p3=0,08,
[0191] p4=0,16, p5=0,17, p6=0,20, p7=0,319.
[0192] Each noise source (in each step) generates
[0193]
[0194] bits, which is 82.61% of the maximum possible entropy (=3) for a random vector three bits wide.
[0195] exist Figure 5 In the example of , it is assumed that the two noise sources contain the same (standard) digitizer. If one of the two noise sources is equipped with a bijective S-box
[0196]
[0197] , then XOR gate 501 extracts the maximum entropy.
[0198] If the noise source digitizer is arranged according to
[0199] π=0,7,1,4,5,2,6,3
[0200] When the same digitizer is used in both noise sources, the combined random words at the output of XOR gate 501 will only achieve an entropy value of 93.78%.
[0201] Fig.10 An arrangement for increasing the entropy by means of a modification of a digitizer provided in the noise source is shown.
[0202] Noise source RQ1 1001 includes a physical random process 1002 and a digitizer 1003 , and noise source RQ2 1004 includes a physical random process 1005 and a digitizer 1006 .
[0203] Digitizer 1003 provides a 3-bit wide random word to a first input of XOR gate 1007 and digitizer 1006 provides a 3-bit wide random word to a second input of XOR gate 1007. At the output of XOR gate 1007 a 3-bit wide combined random word with an entropy value of 99.85% is provided.
[0204] The digitizer 1003 is arranged according to
[0205] π=0,1,2,3,4,5,6,7
[0206] Implementation, and conversion of the arrangement in the digitizer 1006
[0207] π=0,7,1,4,5,2,6,3.
[0208] The XOR gate shown in the example with a 3-bit wide input can be implemented with the aid of three XOR gates with only two inputs by performing the XOR logic operation bit by bit, for example:
[0209]
Claims
1. A method for increasing the entropy of random numbers, - Multiple random processes provide random numbers separately, - wherein a plurality of said random numbers are linked to form a combined random number, wherein one of said random numbers is mapped to another random number beforehand.
2. The method according to claim 1, The combined random number is used in a cryptographic manner.
3. The method according to any one of the preceding claims, Each of the random processes includes at least one physical random process.
4. The method according to claim 3, In this case, the random number is mapped into another random number by means of at least one S-box.
5. The method according to claim 4, The S-box is a bijective S-box or a compression S-box.
6. The method according to any one of the preceding claims, The random processes and digitizers are part of the noise source.
7. The method according to claim 6, In this case, the random number is mapped into a further random number by means of the digitizer.
8. The method according to any one of the preceding claims, In this case, a plurality of the random numbers are linked to form a combined random number by means of at least one XOR logic operation, in particular by means of at least one XOR gate.
9. The method according to any one of the preceding claims, In this case, a mapping of at least one random number to another random number is determined which leads to an increase in the entropy of the combined random number.
10. The method according to claim 9, In this case, a mapping of at least one random number to another random number is determined which leads to an increase in the entropy of the combined random number by testing a plurality of mappings. 11 . A device for increasing the entropy of random numbers, the device comprising a processing unit configured to carry out the steps of the method according to claim 1 .
12. A device for increasing the entropy of random numbers - having a plurality of random generators, each of the plurality of random generators providing a random number, - having at least one mapping unit, by means of which at least one random number can be mapped into at least one further random number, A combining unit is provided, which is coupled to the random generator and the at least one mapping unit in such a way that a combined random number can be determined based on the random number and the at least one further random number.
13. The device according to claim 12, The mapping unit comprises at least one bijective S-box and / or at least one compressed S-box.
14. The device according to any one of claims 12 or 13, The combination unit includes at least one XOR logic operation.
15. The device according to any one of claims 12 to 14, wherein the random generator and the mapping unit are part of a noise source.
