Combined threshold mask protection method of Ascon authentication encryption algorithm
By combining the second-order threshold implementation and exchange bit selection algorithm in the Ascon algorithm, redundant masks are constructed and XOR operations are performed, the problem of insufficient security of the Ascon algorithm in the face of combined attacks is solved, and the second-order power consumption is achieved Effective resistance to attacks and first-order failure attacks, and optimizes hardware implementation efficiency.
Patent Information
- Application Number
- CN202411240884.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-05
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-09-05
AI Technical Summary
When the existing Ascon algorithm faces power consumption attacks and fault attacks, a single protection strategy is vulnerable to combined attacks, and the existing protection methods cannot effectively resist second-order power consumption attacks and fault attacks.
Using a combined threshold mask protection method, the second-order threshold implementation of Ascon algorithm is combined with exchange-based infection technology, partially shared exchange is performed through the exchange bit selection algorithm, the redundant mask of Ascon algorithm is constructed, and the integration is generated through XOR operation. The ciphertext output forms a combined threshold mask protection method for Ascon authentication encryption algorithm.
Effectively resisting second-order power consumption attacks and first-order differential fault attacks, reducing algorithm delay periods, using fewer additional random numbers and smaller circuit implementation area, significantly reducing hardware implementation overhead.
Smart Images

Figure CN118944856A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of information security, and in particular to a combined threshold mask protection method for an Ascon authentication encryption algorithm. Background Art
[0002] Power Attack (PA) is a side channel attack method that infers keys or other sensitive information by analyzing the changes in power consumption of encryption devices when performing encryption operations. This attack method takes advantage of the leakage of encryption devices at the physical level. The idea is that when encryption devices process data, their power consumption will change depending on the type of operation and the specific value of the data processed. Attackers can collect these power consumption traces through the device and extract characteristic information related to the encryption operation from them. By further analyzing these characteristic information, it is possible to infer the keys or other sensitive information used by the device, thereby undermining the security of the encryption system. Masking technology is an effective means of protecting against power consumption attacks. By randomizing the intermediate values processed by the cryptographic device during the encryption process, the energy consumption of the device has no dependency on the intermediate values of the executed cryptographic algorithm.
[0003] Fault Attack (FA) introduces artificial faults (such as voltage fluctuations, clock signal interference, temperature changes, laser interference, etc.) to cause the encryption device to generate erroneous output when performing encryption operations, and analyzes the output of the encryption algorithm in abnormal conditions and compares it with the normal output to infer key information inside the device, such as keys or other sensitive information. This type of attack uses the incorrect behavior of the encryption device when processing faulty data to obtain additional information to help attackers crack the encryption system. Infection countermeasures are currently one of the most popular countermeasures against fault attacks. In addition to the original calculations in the cryptographic algorithm, redundant calculations are also required. Infection countermeasures are implemented by disrupting the data paths of the original calculations and redundant calculations, so that the final erroneous ciphertext is independent of the key.
[0004] Power consumption attacks and fault attacks are two main physical attacks against cryptographic implementations. The above two countermeasures only consider a single power consumption attack or fault attack. Currently, countermeasures against these two attacks often only consider a single attack type. One of the main disadvantages is that they are easily vulnerable to another attack or a combination of attacks. For example, masking technology can effectively resist power consumption attacks by randomizing power consumption characteristics to cover up sensitive information, but this technology usually cannot defend against fault attacks. Due to the limitations of a single protection strategy, existing cryptographic implementations often adopt a method that combines anti-power consumption attack technology with anti-fault attack technology to defend against both attacks at the same time.
[0005] The Ascon algorithm is a lightweight authenticated encryption algorithm developed by an international research team, specifically for resource-constrained environments such as Internet of Things (IoT) devices and embedded systems. The Ascon algorithm was first proposed in 2014 and was widely recognized in the subsequent CAESAR (Competition for Authenticated Encryption: Security, Applicability, and Robustness) competition. In February 2023, the Ascon algorithm was rated as a lightweight cryptographic standard by the NIST (National Institute of Standards and Technology). Due to its efficient, lightweight and secure design, its excellent performance and security, it stood out in this selection and became one of the lightweight encryption standards recommended by NIST. It is widely used in various information systems around the world, including industrial control, wireless sensor networks, smart homes and other fields. With the rapid development of the Internet of Things and embedded devices, the protection capabilities and performance optimization of the Ascon algorithm have received more and more attention.
[0006] In recent years, researchers have focused their work on the mask protection of the Ascon algorithm. For example, Prasad et al. [1] Based on the guard replacement technique, a second-order DOM masking method for the Ascon algorithm with low latency and no additional random numbers is proposed. Gigerl et al. [2] The previously proposed Keccak algorithm S-box first-order mask based on Toffoli gate is used to extend it to the second order without any online random number, and the second-order mask of Ascon algorithm is realized in software. However, these protection methods can only resist the second-order power consumption attack and cannot effectively resist the fault attack. Therefore, Kandi et al. [3] A combined protection implementation method based on the Ascon algorithm of first-order threshold mask and triple / majority fault attack protection is proposed, which effectively resists first-order power consumption attack and first-order differential fault attack, but this work does not increase the security order of threshold implementation to second order. In summary, when designing and implementing cryptographic algorithms, the characteristics of power consumption attack and fault attack must be comprehensively considered, and effective combined protection methods must be proposed for these two attacks to ensure the overall security of the cryptographic algorithm.
[0007] References:
[0008] [1]SHPrasad,F.Mendel,M.Schlaeffer,et al.Efficient Low-LatencyMasking of Ascon without Fresh Randomness.Cryptology ePrint Archive,Paper2023 / 1914;
[0009] [2] B. Gigerl, F. Mendel, M. et al.Efficient Second-Order MaskedSoftware Implementations of Ascon in Theory and Practice.Cryptology ePrintArchive,Paper 2024 / 755;
[0010] [3]A.Kandi,A.Baksi,P.Gan,et al.Side-Channel and Fault Resistant ASCONImplementation:A Detailed Hardware Evaluation(Extended Version).CryptologyePrint Archive,Paper 2024 / 984. Summary of the invention
[0011] The present invention proposes a combined threshold mask protection method for the Ascon authentication encryption algorithm, which combines the threshold implementation with the exchange-based infection technology, constructs a second-order threshold implementation of the Ascon algorithm, and completely copies it as a redundant mask, optimizes the exchange-based infection technology, and performs partially shared exchanges through the exchange bit selection algorithm proposed in the present invention, thereby ensuring that the algorithm can effectively resist second-order power consumption attacks and first-order differential fault attacks.
[0012] The technical method for achieving the purpose of the present invention is:
[0013] A combined threshold mask protection method for an Ascon authentication encryption algorithm comprises the following steps:
[0014] (1) Construct a second-order threshold implementation of the Ascon algorithm;
[0015] The Ascon algorithm structure is analyzed and its second-order threshold is constructed as the initial mask to ensure correctness, incompleteness and uniformity. The second-order detection security is achieved through three shared thresholds.
[0016] (2) Construct the redundancy mask of the Ascon algorithm;
[0017] The constructed second-order threshold implementation of the Ascon algorithm is fully copied as its redundant mask;
[0018] (3) Construct the Ascon algorithm combined threshold mask based on the replicated exchange technique;
[0019] The original mask and redundant mask of the previously constructed Ascon algorithm are assigned to the same plaintext information and encrypted at the same time. The partial sharing of the initial mask and the redundant mask is selected through the exchange bit selection algorithm, and the partial sharing is exchanged. Finally, all the sharing of the initial mask encryption result is integrated using the XOR operation to generate the ciphertext output, forming the final Ascon authentication encryption algorithm combined threshold mask protection method.
[0020] Furthermore, the second-order threshold of the Ascon algorithm described in step (1) is implemented as follows:
[0021] (1.1) Rearrange the Ascon algorithm round function;
[0022] The S-box is the only nonlinear component in the Ascon algorithm. The design of the S-box of the Ascon algorithm is realized by introducing linear transformation and affine transformation on the input and output of the χ mapping function of the Keccak algorithm.
[0023] Move the linear transformation in the S-box behind the linear diffusion layer in the round function, and add the round constant of the next round to the linear transformation of the S-box after calculation. The permutation starts from the χ mapping of the S-box and rearranges the round function of the Ascon algorithm.
[0024] (1.2) Nonlinear function mask;
[0025] The χ mapping function is to perform an XOR operation on each bit with the nonlinear terms of the other two input bits, given by It consists of three bit functions;
[0026] The algebraic normal form of the decomposed χ mapping function is as follows:
[0027] x(a,b,c,d,e)=de+a+d
[0028] y(a,b,c,d,e)=ae+b+e
[0029] z(a,b,c,d,e)=ab+a+c
[0030] t(a,b,c,d,e)=bc+b+d
[0031] w(a,b,c,d,e)=cd+c+e
[0032] Where a, b, c, d, e represent five input bits, and x, y, z, t, w represent five output bits;
[0033] (1.2.1) According to the d+1 sharing threshold implementation proposed by Bilgin et al. in the prior art, a block cipher can achieve d-order security through d+1 input sharing. Therefore, three input sharings are required to construct a second-order security sharing, and at least 9 component functions are required to achieve the second-order incompleteness of the quadratic function.
[0034] (1.2.2) Use ring refresh mask technology;
[0035] The component functions are masked and refreshed using a ring masker, and only the cross terms need to be refreshed. Additional random numbers are used in the ring refresh configuration to ensure second-order incompleteness;
[0036] (1.2.3) Use paired S-box masking technique;
[0037] The Ascon algorithm uses 64 S-boxes in parallel to perform pairings in one round of encryption and decryption, without adding any additional masking costs.
[0038] On the basis of (1.2.2), the two inputs of each input in the other S-box in the paired S-box are shared and added to the method to form the χ mapping function mask method of the Ascon algorithm S-box;
[0039] (1.3) Linear function mask;
[0040] Masking a linear function is achieved by replicating the same XOR operation for each shared component. According to (1.1), the linear transformation and affine transformation in the S-box are linear functions, and the same XOR operation needs to be performed on each shared component to achieve masking;
[0041] (1.4) Complete the construction of the second-order threshold masking method of the Ascon algorithm;
[0042] Use manual layout and routing methods to achieve circuit layout and form a mask method.
[0043] Furthermore, the specific implementation process of constructing the redundancy mask of the Ascon algorithm described in step (2) is as follows:
[0044] The initial mask M is completely copied to M′, which is called a redundant mask. The same plaintext P and key K are input into M and M′ for encryption respectively, and M and M′ have the same number of input shares and output shares. The sharing method and random number used in the masking method are exactly the same.
[0045] Furthermore, the Ascon algorithm combined threshold mask based on the replication exchange technology described in step (3) is specifically implemented as follows:
[0046] (3.1) Determine the range of the number of exchange bits;
[0047] Let the number of exchange bits be w = HW(R 0 ,R 1 ,R 2 ), satisfies w≤d, where d is the detection security order, then the number of exchanged bits ranges from 0, 1, to 2; HW represents the Hamming weight; R 0 ,R 1 ,R 2 Indicates the selected swap bit position;
[0048] (3.2) R is obtained by the exchange bit selection algorithm 0 ,R 1 ,R 2 ;
[0049] (3.2.1) Select input share a 0 ,a 1 ,a 2 , and concatenate them into a decimal number and assign it to A;
[0050] (3.2.2) Calculate the value of A mod5. If A mod5 = 0, go to (3.2.3); if A mod5 = 1, go to (3.2.4); if A mod5 = 2, go to (3.2.5); if A mod5 = 3, go to (3.2.6); if A mod5 = 4, go to (3.2.7);
[0051] (3.2.3) If A mod5 = 0, the input will share a 0 ,a 1 ,a 2 Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0052] (3.2.4) If A mod5 = 1, the input will be shared b 0 ,b 1 ,b 2 Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0053] (3.2.5) If A mod5 = 2, the input will be shared c 0 ,c1 ,c 2 Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0054] (3.2.6) If A mod5 = 3, the input shared d 0 ,d 1 ,d 2 Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0055] (3.2.7) If A mod5 = 4, the input will be shared by e 0 ,e 1 ,e 2 Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0056] (3.2.8) The binary number sel 0 ,sel 1 ,sel 2 Concatenate and convert into a decimal number and assign it to R;
[0057] (3.2.9) Calculate Rmod7, convert the result into a three-digit binary number, and assign it to R 0 ,R 1 ,R 2 ;
[0058] (3.3) Select exchange sharing through G function;
[0059] Let y 0 ,y 1 ,y 2 is the output share of the initial mask, y′ 0 ,y 1 ′,y′ 2 For the output sharing of the redundancy mask, the 3-bit swap bit R is obtained according to the algorithm in (3.2) 0 ,R 1 ,R 2 (R i ∈0,1), and y i ,y i ′ and R 0 ,R 1 ,R 2 (R i ∈0,1) is input into the G function, where
[0060]
[0061] (3.3.1) Determine R i Is it 0? If R i =0, go to (3.3.2); if R i =1, go to (3.3.3);
[0062] (3.3.2) If R i =0, then y i and i ′’s sharing remains unchanged;
[0063] (3.3.3) If R i =1, the corresponding y i With y i 'Perform partial sharing exchange;
[0064] (3.4) After each round of nonlinear operation, the partial shares are exchanged according to the steps of (3.3);
[0065] (3.5) After the last round, all shares of the initial mask encryption results are integrated using an XOR operation to generate the ciphertext output, forming the final Ascon authenticated encryption algorithm combined threshold mask method.
[0066] The beneficial effects of the method of the present invention are:
[0067] (1) The method of the present invention can effectively resist second-order power consumption attacks and first-order differential fault attacks.
[0068] (2) The method of the present invention reduces a delay cycle by rearranging the round function of the algorithm.
[0069] (3) The method of the present invention uses fewer additional random numbers and a smaller circuit implementation area. BRIEF DESCRIPTION OF THE DRAWINGS
[0070] Figure 1 It is a schematic diagram of the S-box structure of the Ascon authentication encryption algorithm of the present invention;
[0071] Figure 2 It is a schematic diagram of the Ascon authentication encryption algorithm implemented based on the round of the present invention;
[0072] Figure 3 Schematic diagram of the paired S-box structure used in the present invention. DETAILED DESCRIPTION
[0073] The content of the present invention is further described below in conjunction with the drawings and embodiments, but the present invention is not limited thereto.
[0074] Example:
[0075] A comprehensive threshold mask protection implementation method for the Ascon authentication encryption algorithm includes three parts:
[0076] Step (1) constructs the second-order threshold implementation of the Ascon algorithm. The specific implementation process is as follows:
[0077] (1.1) Rearrange the Ascon algorithm round function;
[0078] The S-box is the only nonlinear component in the Ascon algorithm. Figure 1 As shown, the design of the S-box of the Ascon algorithm is realized by introducing linear transformation and affine transformation on the input and output of the χ mapping function of the Keccak algorithm;
[0079] The linear transformation in the S-box is moved behind the linear diffusion layer in the round function, and the round constant of the next round is calculated and added to the linear transformation of the S-box. The permutation starts from the χ mapping of the S-box and rearranges the round function of the Ascon algorithm. The round function after arrangement is as follows Figure 2 As shown;
[0080] (1.2) Nonlinear function mask;
[0081] The χ mapping function is to perform an XOR operation on each bit with the nonlinear terms of the other two input bits, given by It consists of three bit functions;
[0082] The algebraic normal form of the decomposed χ mapping function is as follows:
[0083] x(a,b,c,d,e)=de+a+d
[0084] y(a,b,c,d,e)=ae+b+e
[0085] z(a,b,c,d,e)=ab+a+c
[0086] t(a,b,c,d,e)=bc+b+d
[0087] w(a,b,c,d,e)=cd+c+e
[0088] Where a, b, c, d, e represent five input bits, and x, y, z, t, w represent five output bits;
[0089] (1.2.1) According to the d+1 sharing threshold implementation proposed by Bilgin et al., a block cipher can achieve d-order security through d+1 input sharing, so three input sharings are required to construct a second-order security sharing, such as: Then at least 9 component functions are needed to realize the second-order incompleteness of quadratic functions, such as:
[0090] (1.2.2) Use ring refresh mask technology;
[0091] The component functions are masked and refreshed using a ring masker, and only the cross terms need to be refreshed. Extra random numbers are used in the ring refresh configuration to ensure second-order incompleteness.
[0092] Take x(a,b,c,d,e)=de+a+d as an example:
[0093]
[0094] (1.2.3) Use paired S-box masking technique;
[0095] The Ascon algorithm uses 64 S-boxes for parallel calculation in one round of encryption and decryption, and pairs the S-boxes in pairs. The structure of the paired S-boxes is as follows: Figure 3 As shown, no additional mask cost is required;
[0096] The two paired S-boxes are denoted as S 0 (a 0 ,a 1 ,a 2 ,b 0 ,b 1 ,b 2 ,c 0 ,c 1 ,c 2 ,d 0 ,d 1 ,d 2 ,e 0 ,e 1 ,e 2 ) and S 1 (k 0 ,k 1 ,k 2 ,l 0 ,l 1 ,l 2 ,m 0 ,m 1 ,m 2 ,n 0 ,n 1 ,n 2 ,o 0 ,o 1 ,o 2 ), where a i ,b i ,c i ,d i ,e i,i∈0,1,2 is the first S-box S 0 Input sharing, k i ,l i ,m i ,n i ,o i ,i∈0,1,2 is another S-box S 1 Input sharing,
[0097] Based on (1.2.2), S 1 The two inputs of each input are added to the method in pairs, forming the χ mapping function mask method of the Ascon algorithm S-box;
[0098] The specific masking method of the χ mapping function of the S-box in the Ascon algorithm is as follows:
[0099]
[0100]
[0101] (1.3) Linear function mask;
[0102] Masking a linear function is achieved by replicating the same XOR operation for each shared component. According to (1.1), the linear transformation and affine transformation in the S-box are linear functions, and the same XOR operation needs to be performed on each shared component to achieve masking.
[0103] (1.4) Complete the construction of the second-order threshold masking method of the Ascon algorithm;
[0104] Manual placement and routing methods are used to implement circuit layout and form a threshold mask method.
[0105] Step (2) constructs the redundancy mask of the Ascon algorithm. The specific implementation process is as follows:
[0106] The initial mask M is completely copied to M', which is called a redundant mask. The same plaintext P and key K are input to M and M' for encryption, and M and M' have the same input sharing and output sharing numbers. The sharing method and random number used in the masking method are exactly the same.
[0107] Step (3) constructs the Ascon algorithm combined threshold mask based on the replicated exchange technology. The specific implementation process is as follows:
[0108] (3.1) Determine the range of the number of exchange bits;
[0109] Let the number of exchange bits be w = HW(R 0 ,R 1 ,R 2), satisfies w≤d, where d is the detection security order, then the number of exchanged bits ranges from 0, 1, to 2; HW represents the Hamming weight; R 0 ,R 1 ,R 2 Indicates the selected swap bit position;
[0110] (3.2) R is obtained by the exchange bit selection algorithm 0 ,R 1 ,R 2 ;
[0111] (3.2.1) Select input share a 0 ,a 1 ,a 2 , and concatenate them into a decimal number and assign it to A;
[0112] (3.2.2) Calculate the value of A mod5. If A mod5 = 0, go to (3.2.3); if A mod5 = 1, go to (3.2.4); if A mod5 = 2, go to (3.2.5); if A mod5 = 3, go to (3.2.6); if A mod5 = 4, go to (3.2.7);
[0113] (3.2.3) If A mod5 = 0, the input will share a 0 ,a 1 ,a 2 Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0114] (3.2.4) If A mod5 = 1, the input will be shared b 0 ,b 1 ,b 2 Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0115] (3.2.5) If A mod5 = 2, the input will be shared c 0 ,c 1 ,c 2 Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0116] (3.2.6) If A mod5 = 3, the input shared d 0 ,d 1 ,d 2Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0117] (3.2.7) If A mod5 = 4, the input will be shared by e 0 ,e 1 ,e 2 Assign values to sel respectively 0 ,sel 1 ,sel 2 , transfer (3.2.8);
[0118] (3.2.8) The binary number sel 0 ,sel 1 ,sel 2 Concatenate and convert into a decimal number and assign it to R;
[0119] (3.2.9) Calculate Rmod7, convert the result into a three-digit binary number, and assign it to R 0 ,R 1 ,R 2 ;
[0120] The pseudo code of the exchange bit selection algorithm is as follows:
[0121]
[0122] (3.3) Select exchange sharing through G function;
[0123] Let y 0 ,y 1 ,y 2 is the output share of the initial mask, y′ 0 ,y 1 ′,y′ 2 is the output share of the redundancy mask. According to the algorithm in (3.2), we get the 3-bit swap bit R 0 ,R 1 ,R 2 (R i ∈0,1), and y i ,y i ′ and R 0 ,R 1 ,R 2 (R i ∈0,1) is input into the G function, where
[0124]
[0125] (3.3.1) Determine R i Is it 0? If R i=0, go to (3.3.2); if R i =1, go to (3.3.3);
[0126] (3.3.2) If R i =0, then y i and i ′’s sharing remains unchanged;
[0127] (3.3.3) If R i =1, the corresponding y i With y i 'Perform partial sharing exchange;
[0128] (3.4) After each round of nonlinear operation, the partial shares are exchanged according to the steps of (3.3);
[0129] (3.5) After the last round, all shares of the initial mask encryption results are integrated using an XOR operation to generate the ciphertext output, forming the final Ascon authenticated encryption algorithm combined threshold mask method.
[0130] The advantage of the method of the present invention is that it can enable the Ascon algorithm to effectively resist second-order power consumption attacks and first-order differential fault attacks, and only requires the use of fewer additional random numbers and a smaller circuit implementation area, which can significantly reduce its hardware implementation overhead cost.
Claims
1. A combined threshold mask protection method for the Ascon authentication encryption algorithm, characterized in that: The steps include: (1) Construct a second-order threshold implementation of the Ascon algorithm; The Ascon algorithm structure is analyzed and its second-order threshold is constructed as the initial mask to ensure correctness, incompleteness and uniformity. The second-order detection security is achieved through three shared thresholds. (2) Construct the redundancy mask of the Ascon algorithm; The constructed second-order threshold implementation of the Ascon algorithm is fully copied as its redundant mask; (3) Construct the Ascon algorithm combined threshold mask based on the replicated exchange technique; The initial mask and redundant mask of the previously constructed Ascon algorithm are assigned to the same plaintext information and encrypted at the same time. The partial sharing of the initial mask and redundant mask is selected through the exchange bit selection algorithm, and the partial sharing is exchanged. Finally, all the sharing of the initial mask encryption result is integrated using the XOR operation to generate the ciphertext output, forming the final Ascon authentication encryption algorithm combined threshold mask protection method.
2. The combined threshold mask protection method of the Ascon authentication encryption algorithm according to claim 1 is characterized in that: The second-order threshold of the Ascon algorithm described in step (1) is implemented as follows: (1.1) Rearrange the Ascon algorithm round function; The S-box is the only nonlinear component in the Ascon algorithm. The design of the S-box of the Ascon algorithm is realized by introducing linear transformation and affine transformation on the input and output of the χ mapping function of the Keccak algorithm. Move the linear transformation in the S-box behind the linear diffusion layer in the round function, and add the round constant of the next round to the linear transformation of the S-box after calculation. The permutation starts from the χ mapping of the S-box and rearranges the round function of the Ascon algorithm. (1.2) Nonlinear function mask; The χ mapping function is to perform an XOR operation on each bit with the nonlinear terms of the other two input bits, given by It consists of three bit functions; (1.2.1) According to the d+1 sharing threshold implementation proposed in the prior art, the block cipher achieves d-order security through d+1 input sharing, so three input sharings are required to construct a second-order security sharing, and at least 9 component functions are required to achieve the second-order incompleteness of the quadratic function; (1.2.2) Use ring refresh mask technology; The component functions are masked and refreshed using a ring masker, and only the cross terms need to be refreshed. Extra random numbers are used in the ring refresh configuration to ensure second-order incompleteness. (1.2.3) Use paired S-box masking technique; The Ascon algorithm uses 64 S-boxes in parallel to perform pairings in one round of encryption and decryption, without adding any additional masking costs. On the basis of (1.2.2), the two inputs of each input in the other S-box in the paired S-box are shared and added to the method to form the χ mapping function mask method of the Ascon algorithm S-box; (1.3) Linear function mask; Masking a linear function is achieved by replicating the same XOR operation for each shared component. According to (1.1), the linear transformation and affine transformation in the S-box are linear functions, and the same XOR operation needs to be performed on each shared component to achieve masking. (1.4) Complete the construction of the second-order threshold masking method of the Ascon algorithm; Use manual layout and routing methods to achieve circuit layout and form a mask method.
3. The combined threshold mask protection method of the Ascon authentication encryption algorithm according to claim 1 is characterized in that: The specific implementation process of constructing the redundant mask of the Ascon algorithm described in step (2) is as follows: the initial mask M is completely copied to M′, which is called the redundant mask. The same plaintext P and key K are input into M and M′ for encryption respectively, and M and M′ have the same input sharing and output sharing numbers. The sharing method and random number used in the masking method are exactly the same.
4. The combined threshold mask protection method of the Ascon authentication encryption algorithm according to claim 1 is characterized in that: The construction of the Ascon authentication encryption algorithm combined threshold mask based on the exchange technology of replication described in step (3) is specifically implemented as follows: (3.1) Determine the range of the number of exchange bits; Let the number of exchange bits be w = HW (R0, R1, R2), satisfying w ≤ d, where d is the detection security order, then the number of exchange bits ranges from 0, 1, 2; HW represents the Hamming weight; R0, R1, R2 represent the selected exchange bit positions; (3.2) Get R0, R1, and R2 through the exchange bit selection algorithm; (3.2.1) Select the shared inputs a0, a1, a2, and concatenate them into decimal numbers and assign them to A; (3.2.2) Calculate the value of Amod5. If Amod5 = 0, go to (3.2.3); if Amod5 = 1, go to (3.2.4); if Amod5 = 2, go to (3.2.5); if Amod5 = 3, go to (3.2.6); if Amod5 = 4, go to (3.2.7); (3.2.3) If Amod5 = 0, assign the shared inputs a0, a1, a2 to sel0, sel1, sel2 respectively, and go to (3.2.8); (3.2.4) If Amod5 = 1, assign the shared inputs b0, b1, b2 to sel0, sel1, sel2, go to (3.2.8); (3.2.5) If Amod5 = 2, assign the shared inputs c0, c1, c2 to sel0, sel1, sel2 respectively, go to (3.2.8); (3.2.6) If Amod5 = 3, assign the shared inputs d0, d1, d2 to sel0, sel1, sel2 respectively, go to (3.2.8); (3.2.7) If Amod5 = 4, assign the shared inputs e0, e1, e2 to sel0, sel1, sel2 respectively, go to (3.2.8); (3.2.8) Concatenate the binary numbers sel0, sel1, sel2 and convert them into a decimal number and assign it to R; (3.2.9) Calculate Rmod7, convert the result into a three-digit binary number, and assign it to R0, R1, and R2 respectively; (3.3) Select exchange sharing through G function; Let y0, y1, y2 be the output shares of the initial mask, y′0, y1′, y′2 be the output shares of the redundant mask, and according to the algorithm in (3.2), we can get the 3-bit swap bits R0, R1, R2 (R i ∈0,1), and y i ,y i ′ and R0, R1, R2 (R i ∈0,1) is input into the G function, where (3.3.1) Determine R i Is it 0? If R i =0, go to (3.3.2); if R i =1, go to (3.3.3); (3.3.2) If R i =0, then y i and i ′’s sharing remains unchanged; (3.3.3) If R i =1, the corresponding y i With y i 'Partial sharing exchange; (3.4) After each round of nonlinear operation, the partial shares are exchanged according to the steps of (3.3); (3.5) After the last round, all shares of the initial mask encryption results are integrated using an XOR operation to generate the ciphertext output, forming the final Ascon authenticated encryption algorithm combined threshold mask method.
Citation Information
Patent Citations
Automatic threshold masking method for 4-bit password S box
CN113794549A
Rapid search method and system for three-component second-order threshold implementation scheme
CN117271569A
Threshold secret sharing apparatus, threshold secret sharing method, and program
JP2008145917A
