INCREASING THE ENTROPY OF RANDOM NUMBERS

By combining random processes with bijective or compressive S-boxes and XOR gates, the entropy of random numbers is enhanced, addressing the need for high-entropy random numbers in cryptography.

DE102023112670A1Pending Publication Date: 2024-11-14INFINEON TECHNOLOGIES AG
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
DE102023112670
Authority / Receiving Office
DE · DE
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-12
Publication Date
2024-11-14

AI Technical Summary

Technical Problem

Existing cryptographic systems require random numbers with high entropy, typically approaching a per-bit entropy of 1, which current noise generators and S-boxes struggle to achieve efficiently.

Method used

A method and device that combine multiple random processes, using permutations, bijective or compressive S-boxes, and XOR gates to increase the entropy of random numbers by linking and mapping random numbers, optimizing the selection of S-boxes for maximum entropy gain.

Benefits of technology

The combined random numbers achieve significantly higher entropy values, meeting cryptographic security requirements with optimized hardware efficiency and reduced implementation costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

The approaches proposed here concern increasing the entropy of a random number, whereby several random processes each produce a random number, which are combined into a single random number, and at least one of the random numbers is previously mapped to another random number.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Perfect random numbers have a per-bit entropy of 1 (100%). Many applications, especially cryptographic ones, require the best possible random numbers—that is, random numbers that come as close as possible to an entropy of 1. For example, noise sources (also called noise generators) are used to generate random numbers. The state of the art knows a variety of possible implementations for noise sources, e.g., analog or digital noise generators.

[0002] So-called S-boxes are known as basic components of cryptography: An S-box is a substitution operation for converting one binary number into another. Such a mapping can, for example, be invertible (bijective) or compressive.

[0003] One task here is to increase the entropy of random numbers in order to meet high cryptographic security requirements.

[0004] This object is achieved according to the features of the independent claims. Preferred embodiments can be found in particular in the dependent claims.

[0005] The examples proposed herein may be based on at least one of the following solutions. In particular, combinations of the following features may be used to achieve a desired result. The features of the method may be combined with any feature(s) of the device, or vice versa.

[0006] To solve this problem, a method is proposed to increase the entropy of a random number, - in which several random processes each produce a random number, - in which the plurality of random numbers are combined to form a combined random number, with at least one of the random numbers being mapped to another random number beforehand.

[0007] Mapping one random number to another random number may involve permuting or reordering.

[0008] It is a further development that the combined random number is used in a cryptographic procedure.

[0009] It is a further development that every random process includes at least one physical random process.

[0010] In principle, any mechanism that produces a reasonably arbitrary (random) number can be used as a random process. The random number can be a number with a higher or lower entropy. Ultimately, such a random number can also be generated deterministically or by a (pseudo-)random generator. Since the increase in entropy is achieved by combining and mapping multiple random numbers, lower requirements for the entropy of the individual random numbers can be achieved.

[0011] It is a further development that the random number is mapped into the other random number using at least one S-box.

[0012] It is a further development that the S-box is a bijective S-box or a compressive S-box.

[0013] It is a further development that the random process and a digitizer are part of a noise source.

[0014] It is a further development that the random number is mapped into the other random number using the digitizer.

[0015] It is a further development that the plurality of random numbers are combined to form the combined random number by means of at least one XOR operation, in particular by means of at least one XOR gate.

[0016] The XOR operation can be implemented using XOR gates or other logic gates. In particular, an XOR operation of more than 2 bits can be split into multiple 2-bit XOR operations.

[0017] For example, an implementation can be chosen that combines noise sources using XOR gates. One advantage of using XOR gates is that it allows for an exact qualitative statement and thus documentation of the entropy increase.

[0018] It is a further development that the mapping of at least one random number to the other random number is determined which leads to an increase in the entropy of the combined random number.

[0019] It is a further development that the mapping of the at least one random number to the other random number is determined which leads to an increase in the entropy of the combined random number by testing a large number of mappings.

[0020] Such testing can be performed, for example, using a computer search: All (or a subset of all) existing S-boxes can be tried, and for each selected S-box or set of S-boxes, the entropy of the combined random number is determined. From the best S-boxes, at least one S-box can be selected that can be implemented most easily or inexpensively (with respect to at least one objective variable, e.g., space, cost, time, effort, etc.).

[0021] To achieve this, a device for increasing the entropy of a random number is also proposed, wherein the device has a processing unit which is configured to carry out the steps of the method according to one of the preceding claims.

[0022] Furthermore, a device for increasing the entropy of a random number is proposed - with several random generators, each of which provides a random number, - with at least one mapping unit by means of which at least one of the random numbers can be mapped to at least one other random number, - with a combining unit which is coupled to the random number generators and the at least one mapping unit such that a combined random number can be determined based on the random numbers and the at least one other random number.

[0023] Each random generator can correspond to a random process or contain at least one random process.

[0024] In particular, it is an option for the random number generator coupled to the mapping unit to process the random number of the mapping unit instead of the random number determined by the random number generator. If a mapping unit is provided for each random number generator, only the mapped random numbers can be processed in the combining unit.

[0025] It is a further development that the mapping unit comprises at least one bijective and / or at least one compressing S-box.

[0026] It is an option for the mapping unit to be multi-stage or for multi-stage mapping to be implemented, possibly with intermediate XOR gates.

[0027] It is a further development that the combining unit includes at least one XOR operation.

[0028] It is a further development that the random generator and the imaging unit are part of a noise source.

[0029] In this case, the imaging unit can also be called a digitizer.

[0030] The above-described properties, features, and advantages of this invention, as well as the manner in which they are achieved, are described below in conjunction with a schematic description of exemplary embodiments, which are explained in more detail in conjunction with the drawings. For clarity, identical or equivalent elements may be provided with the same reference numerals. Fig. Figure 1 shows a combination of a large number of noise sources that are connected together via an XOR gate. Fig. Figure 2 shows a combination of a large number of noise sources with S-boxes that are connected via an XOR gate. Fig. 3 shows an exemplary arrangement for determining a combined random word using two noise sources connected via an XOR gate. Fig. 4 shows an exemplary arrangement for determining a combined random word by means of two paths connected via an XOR gate, each path comprising a noise source and an S-box. Fig. Figure 5 shows an exemplary arrangement for increasing entropy using two noise sources and a bijective S-box. Fig. Figure 6 shows an exemplary arrangement for increasing entropy using two noise sources and two compressing S-boxes. Fig. 7 shows a section of the arrangement according to Fig. 6, where the S-boxes are implemented using XOR gates as an example. Fig. Figure 8 shows a table with example probability distributions for several noise sources. Fig. Figure 9 shows an exemplary arrangement for increasing entropy using four noise sources and several bijective and compressing S-boxes. Fig. 10 shows an exemplary arrangement for increasing entropy by modifying a digitizer arranged in the noise source.

[0031] Fig. 1 shows an arrangement with b ≥ 2 noise sources RQ1, RQ2,..., RQ b , which are connected to each other via an XOR gate 101.

[0032] The noise sources RQ1, RQ2,..., RQ b independently generate n-bit wide random words (also called random vectors or random numbers), with each noise source having a certain entropy.

[0033] Per unit of time, for example per clock cycle of a processor unit (CPU), b-random words are generated, each of which represents one of the individual noise sources RQ1, RQ2,..., RQ bcorresponding entropy. The combination of the noise sources via the XOR gate 101 causes an increase in entropy (ie, an entropy compression in the random word). The entropy can be further increased by S-boxes S1, S2, ..., S b after the noise sources.

[0034] Fig. 2 shows an exemplary arrangement based on Fig. 1 each S-Box S1, S2,..., S b in the path between noise source RQ1, RQ2,..., RQ b and the XOR gate 201.

[0035] The examples described here determine, in a set of noise sources, those S-boxes that increase the entropy of the combined random word. This increase in entropy can be an optimization, in particular a maximization of entropy. For example, from a 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, e.g., the two S-boxes that lead to the greatest possible increase in entropy (e.g., with regard to a given (maximum) implementation effort).

[0036] The following cases are considered as examples: Case 1 bijective S-boxes: m = n, the number of bits at the input of the S-box is identical to the number of bits at its output. Case 2 compressing S-boxes: m < n, the number of bits at the input of the S-box is greater than the number of bits at its output, a reduction or compression of the bits takes place.

[0037] The b noise sources in the network of noise sources operate independently of one another. A single noise source is either memoryless or has a short-term memory. For a noise source with no memory, the random word generated at time t is (statistically) independent of all previously generated random words, i.e., in particular, statistically independent of the random word generated at time t - 1. For a noise source with a short memory, the random word generated at time t depends on a few previous words. Noise sources with a short-term memory can be modeled using Markov chains. In the Markov chain model, the stationary probability distribution is first calculated and then used as the basis for further analysis. In this way, the case of a noise source with a short-term memory can be traced back to the case of a noise source with no memory. Bijective S-boxes

[0038] Bijective functions are entropy-preserving, meaning that the random word before and after a bijective S-box has the same entropy. However, the presence of bijective S-boxes in the noise source cluster influences the entropy value of the combined random number after the XOR gate.

[0039] This is illustrated by a noise source array comprising two noise sources, each producing a 2-bit random word. These are two memoryless noise sources, RQ1 and RQ2, that generate 2-bit random words X and Y per CPU cycle. The generated 2-bit random words can take four different values: 0:=(00),1:=(10),2:=(01),3:=(11).

[0040] The four values ​​are assumed with certain, possibly constant (and generally different) probabilities. From these probabilities, the probabilities for the XOR sum Z = X ⊕ Y can be calculated, which then yields the entropy H(Z) of the combined random word.

[0041] Fig. Figure 3 shows an exemplary arrangement (without S-boxes) with the two noise sources RQ1 and RQ2, each of which supplies 2-bit random words X and Y to the inputs of an XOR gate 301. At the output of the XOR gate 301, a 2-bit random word Z is generated per clock cycle, which can assume values ​​from 0 to 3. The corresponding probabilities W0 to W3 are: W0=Pr(Z=0) =Pr(X=0)⋅Pr(Y=0)+Pr(X=1)⋅Pr(Y=1) +Pr(X=2)⋅Pr(Y=2)+Pr(X=3)⋅Pr(Y=3). W1=Pr(Z=1) =Pr(X=0)⋅Pr(Y=1)+Pr(X=1)⋅Pr(Y=0) +Pr(X=2)⋅Pr(Y=3)+Pr(X=3)⋅Pr(Y=2). W2=Pr(Z=2) =Pr(X=0)⋅Pr(Y=2)+Pr(X=1)⋅Pr(Y=3) +Pr(X=2)⋅Pr(Y=0)+Pr(X=3)⋅Pr(Y=1). W3=Pr(Z=3) =Pr(X=0)⋅Pr(Y=3)+Pr(X=1)⋅Pr(Y=2) +Pr(X=2)⋅Pr(Y=1)+Pr(X=3)⋅Pr(Y=0).

[0042] Here, Pr(K = l) denotes the probability that the random word K takes the value l.

[0043] From the probabilities W0, W1, W2 and W3 the entropy of the random word Z is H(Z)=−∑i=03Wilog2(Wi).

[0044] Fig. 4 shows based on Fig. 3 shows an exemplary arrangement with bijective S-boxes S and T in the path between the noise source and the XOR gate 401. The noise source RQ1 supplies a 2-bit random word X to the input of the S-box S, via whose output a modified 2-bit random word X' is supplied to the first input of an XOR gate 401. The noise source RQ2 supplies a 2-bit random word Y to the input of the S-box T, via whose output a modified 2-bit random word Y' is provided to the second input of the XOR gate 401. At the output of the XOR gate 401, a 2-bit random word Z' is generated per clock cycle, which can assume values ​​from 0 to 3. The corresponding probabilities V0 to V3 are: V0=Pr(Z'=0) =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)) V1=Pr(Z'=1) =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)) V2=Pr(Z'=2) =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)) V3=Pr(Z'=3) =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))

[0045] Although the modified random words X' and Y' contain the same entropy as the original random words X and Y, the random word Z' = X' ⊕ Y' takes on the four possible values ​​0 to 3 with different probabilities than the random word Z = X ⊕ Y. This results in a new entropy value for Z' (compared to Z).

[0046] The entropy of Z' = X'⊕Y' is determined by H(Z')=−∑i=03Vilog2(Vi).

[0047] The choice of S-boxes S and T used influences the entropy H(Z'). For example, one goal is to determine those S-boxes S and T for which H(Z') is maximum. This can be achieved by determining the corresponding entropy H(Z') for all possible pairs (S, T) of S-boxes. Now, for example, the pair (S, T) for which the entropy H(Z') assumes the maximum possible value can be selected. In addition, at least one objective can be taken into account within the framework of a multi-objective optimization: For example, only those S-boxes can be considered for which the implementation costs do not exceed a specified threshold, or it can be determined that the implementation costs should also be as low as possible, so that a weighted optimization can be carried out between the objective of increasing entropy and the objective of reducing costs.In general, different optimization strategies are possible that lead to the selection of suitable S-boxes. For example, a suboptimal entropy value may be sufficient if it already meets certain requirements and / or if the costs for the associated S-boxes are above a specified threshold. Example: Two noise sources, one bijective S-box

[0048] Given are two memoryless noise sources, each generating 3-bit wide random words (i.e. 8 different symbols 0 to 7) with the following probabilities (p i is the probability for the symbol i): p0=0.001,p1=0.02p2=0.05,p3=0.08,p4=0.16,p5=0.17,p6=0.20,p7=0.319.

[0049] The two noise sources produce random words with 2.4783 bits of entropy and an entropy value of 82.62%, respectively.

[0050] One goal is to increase the entropy of the combined random number. This is achieved, for example, with a single S-box.

[0051] Fig. Figure 5 shows an exemplary arrangement with two noise sources RQ1 and RQ2, wherein an S-box 502 is arranged in the path between the noise source RQ1 and an XOR gate 501. The S-box 502 is denoted by S in this example. It is, for example, a non-linear S-box, which is determined by S=(0123456707145263).

[0052] The noise source RQ2 is directly connected to the XOR gate 501. The combined random number is provided at the output of the XOR gate 501.

[0053] The two noise sources RQ1 and RQ2, as explained above, provide 3-bit wide random numbers with an entropy value of 82.61% per noise source. Fig. 5, it can be achieved that the combined random number at the output of the XOR gate 501 has an entropy of 2.9956 bits or an entropy value of 99.85% (in contrast to 2.8133 bits or 93.78% if the S-box 502 is omitted).

[0054] The algebraic description of the S-Box 502 given here as an example is determined by S:(x,y,z)→(x+y+z,x+xy+yz,x+z), where x, y and z are given by the three bits at the input of the S-Box 502. Compressing S-boxes

[0055] A compressing S-box is a function that maps binary words of length n to binary words of length m, where m < n. By using compressing S-boxes in the noise source cluster, an additional compression of the entropy in the final (combined) random word can be achieved. In many cases, the final random word already has a sufficiently high per-bit entropy, so that post-processing to further increase entropy is unnecessary.

[0056] By eliminating post-processing, the costs of the hardware implementation of the random number generator are reduced. Another advantage is that random numbers can be generated at a higher data rate (i.e., more random numbers per unit of time or in a shorter time).

[0057] In particular, it is proposed to use improved or optimized compressing S-boxes in the noise source network so that the entropy of the combined random words is already sufficiently high and, for this reason, post-processing can be dispensed with.

[0058] "Sufficiently high" can specifically mean that the entropy value of the combined random word meets a requirement, e.g., of an authority, a customer, or a standard. In many cryptographic applications, it is common practice to specify the quality of the random number based on a minimum entropy value that must be achieved. Example: Two noise sources, two compressing S-boxes

[0059] Fig. 6 shows the two noise sources RQ1 and RQ2 from Fig. 5 with identical probability distributions. In contrast to Fig. 5, an S-box S 602 is arranged in the path between the noise source RQ 1 and an XOR gate 601, and another S-box T 603 is arranged in the path between the RQ 2 and the XOR gate 601. Each of the S-boxes S and T compresses 3-bit words into 2-bit words.

[0060] Each of the two noise sources RQ 1 and RQ 2 generates 3-bit long random words with an entropy of 2.4783 bits and an entropy value of 82.61%, respectively.

[0061] The S-boxes are defined as follows: S=(0123456701322310) and T=(0123456701233210).

[0062] Through the Fig. 6, the entropy of the combined random word provided by the XOR gate 601 is increased to 1.99984 bits or 99.992%.

[0063] The S-boxes S and T are linear in this example and can be represented as 2x3 matrices as follows: S=(110011) and T=(101011).

[0064] Fig. Figure 7 shows an exemplary arrangement of a possible implementation of the S-boxes 602 and 603 using XOR gates, which follows directly from the 2x3 matrices shown above.

[0065] A combination of bijective and compressive S-boxes can also be used in a noise source network. Furthermore, the individual noise sources can exhibit different probability distributions, for example, if identical noise sources are operated with different system parameters (e.g., different oscillators with different frequencies). Noise sources of different types or designs can also be used. Example: Four different noise sources

[0066] In this example, four different noise sources are to be combined to form a noise source network.

[0067] Fig. Figure 8 shows a table with an example probability distribution for each of the four noise sources: Each of the noise sources generates eight symbols per time unit, labeled 0 to 7. The noise sources RQ1 to RQ4 occupy rows 2 to 5 of the table. For example, the second entry 0.02 in the second row means that the noise source RQ1 generates the symbol 0 with a probability of p = 0.02. The noise source RQ b generates only six different symbols 1 to 6, the symbols 0 and 7 are not generated by RQ3.

[0068] If the noise source generates the symbols 0 to 7 with the Fig. 8, the entropy of the 3-bit random word generated by the respective noise source is h=−∑i=07pilog2(pi).

[0069] Here, p denotes i the probability for the symbol i with i = 0, ... , 7. The last two columns in Fig. 8 indicate the bit entropy (i.e. the entropy of an output symbol of the noise source, which is a 3-bit random number or a random vector of length 3 bits) or the entropy value (also called percentage entropy value) for the respective noise source.

[0070] In this example, this noise source system is combined with S-boxes. Fig. Figure 9 shows an exemplary arrangement comprising the four noise sources RQ1 to RQ4 mentioned here, in Fig. 9 also designated 901 to 904.

[0071] The noise source RQ1 901 supplies a 3-bit random word to a bijective S-box S 905, whose output supplies a modified 3-bit random word to the first input of an XOR gate 907.

[0072] The noise source RQ2 902 supplies a 3-bit random word to the second input of the XOR gate 907.

[0073] The noise source RQ b903 supplies a 3-bit random word to a bijective S-box T 906, whose output supplies a modified 3-bit random word to the first input of an XOR gate 908.

[0074] The noise source RQ4 904 supplies a 3-bit random word to the second input of the XOR gate 908.

[0075] The output of the XOR gate 907 is connected to the input of a compressing S-box J 909, which modifies the incoming 3-bit random word into a compressed 2-bit random word and forwards it to the first input of an XOR gate 911.

[0076] The output of the XOR gate 908 is connected to the input of a compressing S-box K 910, which modifies the incoming 3-bit random word into a compressed 2-bit random word and forwards it to the second input of the XOR gate 911.

[0077] The combined 2-bit random word is available at the output of the XOR gate 911.

[0078] The noise sources RQ1 to RQ4 have the Fig. 8 given entropies. The arrangement according to Fig. 9 causes the following gradual increases in entropy values: - at the output of the XOR gate 907: 94.47%, - at the output of the XOR gate 908: 98.65%, - at the output of the S-Box J 909: 99.78%, - at the output of the S-Box K 910: 99.94% and - at the output of the XOR gate 911: 99.99995%.

[0079] The S-boxes used here are defined as follows: S=(0123456703746815) and S=(0123456707214356), J=(0123456713130002) and K=(0123456701201323)

[0080] The algebraic description of the S-boxes is: S:(x,y,z)→(x+y+xz,x+y+z+xz,y+z+xz);T:(x,y,z)→(x+yz,x+y+yz,x+z+ xy);J:(x,y,z)→(1+z,x+xz+xyz);K:(x,y,z)→(x+z+xy+xz+yz,y+xy+xz).

[0081] It should also be noted that the approach presented here is highly robust. For example, if 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 lower entropy limit of 99.7% required by the German Federal Office for Information Technology.

[0082] A further advantage is that, in accordance with the standard, a total failure is detected as early as possible: Since the combination logic proposed here is memoryless, in the event of a total failure of all noise sources, the same random words would always be provided at the output of the XOR gate 911, thus detecting the total failure without delay. Noise sources with customized digitizers

[0083] The above-mentioned Fig. 4 shows the entropy compression using two S-boxes S and T into a combined random word Z'=X'⊕Y'=S(X)⊕T(Y) at the output of the XOR gate 401. This entropy compression is achieved, for example, by the use of bijective S-boxes S and T (e.g., determined by computer search).

[0084] Another approach involves appropriate modification of the noise sources involved, especially their digitizers. The same applies to noise source bundles with more than two noise sources. This approach is advantageous from an implementation perspective (gate area, logic depth of the circuit, power consumption).

[0085] The noise source consists of two parts, the actual physical noise source and the digitizer. The physical noise source

[0086] The physical noise source implements a physical random process. A random event (also called a random experiment) occurs per unit of time (determined by an external clock, e.g., the clock of a processor unit). The result of the random experiment is represented by a measured quantity ξ, which is an element of the so-called event space Ω: ξ ∈ Ω.

[0087] The event space Ω is the disjoint union of k = 2 n , n ≥ 1, subareas A0, A1, ..., A k-1 : Ω=A0∪A1⋯∪Ak−1 with Ai∩A j = θ for 0≤i <j≤k-1.

[0088] The outcome of the random experiment is unpredictable. But for each i∈{0,1,…,k−1} is the probability P(A i ) that the measured quantity ξ in the subset A i will land. The k-tuple (P(A0),P(A1),…,P(Ak−1)) with P(A i ) ≥ 0 for 0 ≤ i ≤ k - 1 and P(A0)+P(A1)+⋯+P(Ak−1)=1 is called the probability distribution of the noise source. The probability distribution of a noise source (used in practice) is known.

[0089] If the noise source (as assumed here) is memoryless, then the (Shannon) entropy h of the noise source results from the probability distribution according to the formula h=−∑i=0k−1P(Ai)log2P(Ai).

[0090] Entropy is therefore a pure property of the physical noise source and independent of the design of the digitizer. The digitizer

[0091] Let ξ ∈ Ω be the result of a random experiment. Since Ω has k non-overlapping regions A i is divided, there are k different events ξ ∈ A i , each with probability P(A i ) occur, i = 0, 1, ... , k - 1.

[0092] The digitizer converts each of the k possible events ξ ∈ A i into a unique symbol. Therefore, k different symbols are required. The k symbols can be represented by integers 0, 1, ... , k - 1.

[0093] Initially, it remains unclear which event should be represented by which symbol. One possibility is the following assignment: The event ξ ∈ A0 is represented by the symbol 0: When the event ξ ∈ A0 occurs in the physical noise source, the digitizer generates the symbol 0, and this symbol 0 is output. When the event ξ ∈ A1 occurs, the symbol 1 is generated and output. When ξ ∈ A2 occurs, the symbol 2 is output, and so on.

[0094] The digitizer implemented according to this mapping rule is called the canonical digitizer for the noise source.

[0095] In principle, the k events ξ ∈ A i, i = 0,1, ... , k - 1, the k symbols 0,1, ... , k - 1 are assigned in any order. Therefore, there are k!=k(k−1)(k−2)…3⋅2⋅1 Ways to specify the digitizer.

[0096] For example, if k = 8, then there are 8!=40320 ways to implement the digitizer. All of these variations result in a noise source with the same entropy.

[0097] Furthermore, π=π(0),π(1),…,π(k−1) an arbitrary permutation (arrangement) of the k symbols 0,1, ... ,k- 1. The digitizer for the permutation π can be determined as follows: If the event ξ ∈ A i occurs, then the symbol π(i) is generated and output, i = 0,1, ... , k - 1.

[0098] The following S-box defined by the permutation π is considered: S=(01…k−1π(0)π(1)…π(k−1)).

[0099] We have π(i) = S(i) for all i ∈ {0, 1, ... , k - 1}. The permutation π thus defines a bijective S-box. Conversely, a given bijective S-box defines S:{0,1,…,k−1}→{0,1,…,k−1} the corresponding permutation π :=S(0), S(1), ... , S(k - 1), of the k symbols 0,1, ... , k-1.

[0100] In particular, there are the following two variants: Variant (A) is a physical noise source with a canonical digitizer and a downstream S-box S. Variant (B) comprises a noise source without an S-box, where the digitizer of the noise source is arranged according to the permutation π=S(0),S(1),…,S(k−1) is implemented.

[0101] Both variants (A) and (B) produce for each event ξ ∈ A i always the same output symbol. In variant (A), this output symbol appears at the output of the S-Box S, and in variant (B), the output symbol is provided directly by the noise source's digitizer.

[0102] It is therefore also proposed that in a noise source network (as shown for example in Fig. 1) Noise sources with specifically defined digitizers can be used. Such digitizers implement permutations for which the entropy of the combined random word is increased or, in particular, optimized. Suitable permutations can be determined in advance using computer search.

[0103] It should be noted that in the first approach, the noise sources are provided with external, real S-boxes, whereas in the second approach, the same S-boxes are included internally and virtually in the digitizer.

[0104] To explain the term “digitizer”: k = 2 n . For example, if we consider the case k = 16, then n = 4. The physical noise source can have 16 different internal states ξ ∈ A i, 0 ≤ i ≤ 15. The corresponding 16 symbols in the digitizer are determined to be 0, 1,..., 15. In the hardware implementation of the noise source, these symbols are represented by four-bit binary vectors. For example, the symbol 15 is represented by the vector (1, 1, 1, 1). Thus, the event ξ ∈ A 15 in the physical noise source is digitized into the vector (1, 1, 1, 1). Example: Noise sources with digitizers

[0105] In the Fig. In the example shown in Figure 5, there are two noise sources RQ1 and RQ2, which use the same physical random process. Each noise source can contain eight different events ξ ∈ A i occur, 0 ≤ i ≤ 7. The probabilities p i = P(A i ) = Pr(ξ ∈ A i ) for the occurrence of the individual events are given by: p0=0.001p1=0.02p2=0.05p3=0.08,p4=0.16,p5=0.17p6=0.20,p7=0.319.

[0106] Each noise source generates (in each step) h=−∑i=07pilog2pi=2.4783

[0107] Bit entropy, which is 82.61% of the maximum possible entropy (=3) for a three-bit wide random vector.

[0108] In the example to Fig. 5, it was assumed that the two noise sources contain identical (canonical) digitizers. Then, the XOR gate 501 extracts the maximum entropy when one of the two noise sources is matched with the bijective S-box S=(0123456707145263) is provided.

[0109] The maximum entropy of 99.85% is also achieved when the digitizer of a noise source is permuted according to the π=0,7,1,4,5,2,6,3 is implemented. Using identical digitizers in both noise sources, the combined random word at the output of XOR gate 501 would only achieve an entropy value of 93.78%.

[0110] Fig.10 shows an arrangement for increasing the entropy by modifying a digitizer arranged in the noise source.

[0111] A noise source RQ1 1001 comprises a physical random process 1002 and a digitizer 1003 and a noise source RQ2 1004 comprises a physical random process 1005 and a digitizer 1006.

[0112] The digitizer 1003 provides a 3-bit wide random word to the first input of an XOR gate 1007 and the digitizer 1006 provides a 3-bit wide random word to the second input of the XOR gate 1007. At the output of the XOR gate 1007, 3-bit wide combined random words with the entropy value 99.85% are provided.

[0113] The digitizer 1003 is according to the permutation π=0,1,2,3,4,5,6,7 realized and in the digitizer 1006 the permutation π=0,7,1,4,5,2,6,3 implemented.

[0114] The XOR gates with 3-bit wide inputs shown in the examples can be implemented using three XOR gates with only two inputs by performing a bitwise XOR operation, e.g.: 4

Claims

[1] Method for increasing the entropy of a random number, - in which several random processes each produce a random number, - in which the plurality of random numbers are combined to form a combined random number, with at least one of the random numbers being mapped to another random number beforehand. [2] Method according to claim 1, wherein the combined random number is used in a cryptographic method. [3] Method according to one of the preceding claims, wherein each random process comprises at least one physical random process. [4] Method according to claim 3, wherein the random number is mapped into the other random number by means of at least one S-box. [5] The method of claim 4, wherein the S-box is a bijective S-box or a compressive S-box. [6] A method according to any preceding claim, wherein the random process and a digitizer are part of a noise source. [7] Method according to claim 6, wherein the random number is mapped into the other random number by means of the digitizer. [8] Method according to one of the preceding claims, in which the plurality of random numbers are combined to form the combined random number by means of at least one XOR operation, in particular by means of at least one XOR gate. [9] Method according to one of the preceding claims, in which the mapping of the at least one random number to the other random number is determined which leads to an increase in the entropy of the combined random number. [10] A method according to claim 9, wherein the mapping of the at least one random number to the other random number is determined which leads to an increase in the entropy of the combined random number by testing a plurality of mappings. [11] Device for increasing the entropy of a random number with a processing unit which is arranged to carry out the steps of the method according to one of the preceding claims. [12] Device for increasing the entropy of a random number - with several random generators, each of which provides a random number, - with at least one mapping unit by means of which at least one of the random numbers can be mapped to at least one other random number, - with a combining unit which is coupled to the random number generators and the at least one mapping unit such that a combined random number can be determined based on the random numbers and the at least one other random number. [13] Apparatus according to claim 12, wherein the mapping unit comprises at least one bijective and / or at least one compressive S-box. [14] Device according to one of claims 12 or 13, wherein the combining unit comprises at least one XOR operation. [15] Apparatus according to any one of claims 12 to 14, wherein the random generator and the mapping unit are part of a noise source.

Citation Information

Patent Citations

  • Method for generating an output from a random source of a random number generator

    DE102014200163A1

  • Generation of random numbers

    DE102015100760A1

  • pseudorandom number generator for a stream cipher

    DE10347455A1