JP4003723B2 - Information processing equipment, tamper resistant processing equipment - Google Patents

Description

  The present invention relates to a tamper resistant device such as a highly confidential IC card.

  An IC card is a device mainly used for holding information that cannot be rewritten without permission, encrypting data using an encryption key that is secret information, and decrypting ciphertext. Since the IC card does not have a power source, when it is inserted into the reader / writer, it is supplied with power and becomes operable. When it becomes operable, it receives a command from the reader / writer and transfers data according to the command. General explanations of IC cards can be found in “IC Cards” written by Junichi Mizusawa, edited by the Institute of Electronics, Information and Communication Engineers, Ohm.

  As shown in FIG. 1, the IC card has a structure in which an IC card chip 102 is mounted on a card 101. In general, an IC card has a contact, through which power is supplied from a reader / writer and data is communicated with the reader / writer.

  The configuration of the IC card chip is basically the same as that of a microcomputer. As shown in FIG. 2, the configuration includes a central processing unit 201, a storage device 204, an input / output port 207, and a co-processor 202. The central processing unit 201 is a device that performs logical operations and arithmetic operations, and the storage device 204 is a device that stores programs and data. The input / output port is a device that communicates with the reader / writer. The co-processor is a special arithmetic device for performing a remainder operation, and is a device used for arithmetic operations of RSA that is an asymmetric cipher. Many IC card processors do not have a co-processor. The data bus 203 is a bus that connects each device.

  The storage device 204 includes a ROM (Read Only Memory), a RAM (Random Access Memory), an EEPROM (Electrical Erasable Programmable Read Only Memory), and the like. ROM is a memory that cannot be changed, and is a memory that mainly stores programs. The RAM is a rewritable memory, but when the power supply is interrupted, the stored contents are not erased. When the IC card is removed from the reader / writer, the power supply is interrupted, so that the RAM contents cannot be retained. The EEPROM is a memory that can retain the contents even when the supply of power is interrupted. Used to store data that needs to be rewritten, even if the IC card is removed from the reader / writer. For example, the prepaid frequency in a prepaid card is rewritten every time it is used, and since it is necessary to retain data even if it is removed from the reader / writer, it is retained in the EEPROM.

  The IC card is used to store important information and perform cryptographic processing in the card because programs and important information are sealed in an IC card chip. The difficulty of decrypting cryptographic processing with an IC card was thought to be the same as the difficulty of decrypting cryptographic algorithms. However, it has been suggested that by observing and analyzing the current consumption when the IC card is performing encryption processing, it is possible that the content of the encryption processing and the encryption key can be estimated more easily than the decryption of the encryption algorithm. The current consumption can be observed by measuring the current supplied from the reader / writer. Such hazards are described in 8.5.1.1 Passive protective mechanisms (page 263) of "Smart Card Handbook" by John Wiley & sons W. Rankl & W. Effing.

  The CMOS constituting the IC card chip consumes current when the output state changes from 1 to 0 or from 0 to 1. In particular, since the bus of the data bus 203 has a large electric capacity, when the bus value changes from 1 to 0 or from 0 to 1, a large current is consumed. Therefore, the observation of current consumption suggests the possibility of knowing what is working in the IC card chip.

FIG. 3 shows a waveform of current consumption in one cycle of the IC card chip. Depending on the data being processed, the current waveforms are different, such as 301 and 302. Such a difference occurs depending on data flowing through the bus 203 and data processed by the central processing unit 201.

  Consider a case where data is transferred to a 16-bit precharge bus. The precharge bus is a bus that aligns all bus values to “0” before data transfer. When data having the same number of “1” bits, for example, “88” and “11” are transferred to this bus as hexadecimal numbers having “1” bits of 2, the current waveform Have almost the same waveform. This is because, since the number of bits changed from “0” to “1” is the same, current is consumed and consumed in the same manner, and the same current waveform is obtained. If data having a different number of “1” bits, for example, “89” or “19” in which the number of “1” bits is 3, the number of “1” bits is 2. Current consumption is different from data. This is because the bus value for three bits has changed from “0” to “1”, and that much current is consumed. Therefore, the current consumption is increased by 1 bit compared to the data in which the previous 2 bits have changed. In general, there is a regularity that the current waveform becomes higher as the number of “1” bits is larger. From this regularity, the data being transferred can be estimated.

  The current waveform in FIG. 3 is a sum of currents consumed not only by the bus but also by components constituting the IC card chip. However, a microcomputer such as an IC card chip is divided into a phase in which data is mainly transferred to the bus, a phase in which the CPU is mainly performing calculations, and a phase in which data is written in registers. Therefore, paying attention to the phase, it is possible to know mainly the difference in current consumption in which part and estimate the data operation in that part.

  The difference between specific instructions will be described by taking the following left shift instruction as an example.

shiftl R1 (Formula 1)
This instruction is an instruction that shifts the contents of the register 1 to the left, shifts the bit string of the register to the left, and enters the value of the most significant bit into the condition code register as a carrier. Since the most significant bit of the register R1 is transferred to the condition code register via the data bus, it is possible to identify whether the most significant bit is “0” or “1” by comparing the current waveform sizes. There is sex. If important data is contained in R1, it is one bit of the data, but there is a possibility that “0” or “1” can be recognized. In particular, in encryption processing such as DES, operations for shifting the encryption key frequently occur. When this shift operation is performed, a current waveform capable of estimating the encryption key data is generated, and there is a risk that the encryption key is estimated.

  The same applies to the operation of the co-processor 202. If the calculation contents are biased depending on the encryption key, the bias may be obtained from the current consumption, and the encryption key may be estimated.

  An object of the present invention is to reduce the relationship between data processing and current consumption in an IC card chip. If the relationship between current consumption and chip processing decreases, it will be difficult to estimate the processing within the IC card chip and the encryption key from the observed current consumption waveform. The point of the present invention is to make it difficult to estimate the processing and the encryption key from the waveform of the current consumption by modifying and processing the data to be processed by the IC card chip.

  A tamper resistant device typified by a chip for an IC card includes a program storage unit for storing a program, a storage device having a data storage unit for storing data, and a central processing unit that performs predetermined processing according to the program and performs data processing. The computer can be considered as an information processing device that includes one or more data processing means including processing instructions that give execution instructions to the central processing unit. The present invention, as a method for reducing the relationship between the data being processed and the consumption current of the IC card chip, transforms the data to be processed with disturbance data, processes the data with the modified data, Later, the data for disturbance is inversely transformed to obtain a correct processing result. As the disturbance data used after the data processing, if necessary, processed disturbance data that has been subjected to the same processing as the data processing is used. Further, the disturbance data is set at random for each data processing. In this way, in data processing, different original data can be used each time without using original data. This makes it difficult to estimate data from the current waveform.

  Specifically, first, the disturbance data Xi is created, the data D1 is transformed, and the transformed data H1 is created. Examples of transformation methods include exclusive OR, addition, and multiplication. In the data processing, the deformation data H1 is used to perform data processing to generate processed deformation data H2. Originally, the data D1 is used. However, since the deformation data H1 is used here, it is difficult to estimate the data D1 from the current waveform for processing the deformation data H1. Since the deformation data H1 is deformed by different disturbance data Xi for each process, the deformation data H1 is different every time. Therefore, the current waveform for processing the deformation data H1 also becomes a different waveform each time, and it is meaningless to estimate the deformation data H1 from the waveform.

  When the disturbance data Xi needs to be processed in the same manner as the data D1, the disturbance data Xi is input, data processing is performed, and processed disturbance data Xo is generated. Then, the processed deformation data H2 is processed using the processed disturbance data Xo, and processed data D2 is obtained as a result of processing the input data D1 by the input data processing means.

  When it is necessary to use a different data transformation method, it may be necessary to concatenate and execute the data transformation. In that case, data transformation, transformation data processing, disturbance data processing, and data reverse transformation processing are nested or combined so that data transformation is continued so that original data is not processed.

  INDUSTRIAL APPLICABILITY The present invention can be used for information concealment such as transposition for replacing data, substitution, or table access in an encryption algorithm. In the case of data replacement, exclusive OR is used for data transformation, and in the data processing, the deformed data and the disturbance data are processed in the same way, and the transformed data and the disturbance data are always subjected to exclusive OR. One effective method is to perform deformation processing so that original data can be obtained when executed.

  According to the present invention, it is difficult to estimate the processing and the encryption key from the waveform of the current consumption by modifying the processing data in the IC card chip.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 shows the appearance of an IC card. The size, the position of the IC card chip 102, the number of contacts, and the assignment of the IC card 101 are defined by the ISO7816 standard.

  FIG. 2 shows an internal configuration of the IC card chip 102. The configuration is as already described in the description of the prior art. The present invention makes it difficult to estimate real data from the current waveform consumed by the IC card chip hardware generated during processing by adding disturbance to the data processed by the program 205.

  The basic idea will be explained using the following simple instruction sequence as an example.

logica_shiftl R1 (Formula 2)
xor R1 R2 (Formula 3)
Expression 2 is an instruction that logically loops the value of the register R1 to the left. The most significant bit moves to the least significant bit. The result is stored in the register R1. The result is the exclusive OR of the contents of the register R2 and the result is an instruction sequence stored in the register R2. Such an instruction appears frequently in a cryptographic algorithm such as DES. In Expressions 2 and 3, since the data to be processed is handled as it is, the magnitude of the current waveform is changed depending on the contents of the data. Therefore, there is a possibility that data can be estimated by observing the current waveform.

  The instruction sequence is changed to the following so as not to directly handle the data processed by the expressions 2 and 3.

xor X1 R1 (Formula 4)
xor X2 R2 (Formula 5)
logica_shiftl R1 (Formula 6)
xor R1 R2 (Formula 7)
logica_shiftl X1 (Formula 8)
xor X1 X2 (Formula 9)
xor X2 R2 (Formula 10)
Here, X1 and X2 are random numbers selected arbitrarily, and are disturbance data. Expressions 4 and 5 are deformation processes in which the random numbers X1 and X2 are exclusive-ORed with the registers R1 and R2, and the real data is deformed using the disturbance data. Expressions 6 and 7 are the same data processing as Expressions 2 and 3, but since the deformation process is performed, the values of R1 and R2 are different from the values of the real data. Expressions 8 and 9 perform data processing on the disturbance data itself. This is an inverse transformation process that returns the original data by exclusive ORing the processing result of the disturbance data and the processing result of Expression 7 in Expression 10 .

  This can be expressed by the following numerical values. Let R1 and R2 be the following values.

R1: 11001010 (Formula 11)
R2: 01010111 (Formula 12)
The value of R1 as a result of Expression 2 is as follows.

R1: 10010101 (Formula 13)
And the processing result of Formula 3 is as follows.

R2: 11000010 (Formula 14)
Next, a case where the present invention is modified will be described. First, the disturbance data is as follows. The values of the registers R1 and R2 are the same.

X1: 10010111 (Formula 15)
X2: 00111010 (Formula 16)
As a result of processing Equations 4 and 5, the values of the registers R1 and R2 are as follows.

R1: 01011101 (Formula 17)
R2: 01101101 (Formula 18)
The values of the registers R1 and R2 as a result of processing Expression 6 and Expression 7 are as follows.

R1: 10111010 (Formula 19)
R2: 11010111 (Formula 20)
The result of data processing on the disturbance data X1 and X2 in the same way using Equation 8 and Equation 9 is as follows.

X1: 00101111 (Formula 21)
X2: 00010101 (Formula 22)
As a result of the inverse transformation processing of Expression 10, the same result as the processing result Expression 14 in which real data is handled as it is can be obtained.

R2: 11000010 (Formula 23)
As shown in this example, the original value can be obtained by performing deformation processing, performing the same data processing on the deformed data and the disturbance data, and performing reverse deformation using the result. In the data processing, since the real data is not used, the deformation data being processed can be estimated even by looking at the current waveform, but it is difficult to estimate the real data.

  The example shown above with specific numerical values is shown in a more general form. The actual processing is as follows.

Output (j) = f (Input (i)) (Equation 24)
This is a process of inputting i inputs Input, performing a process f, and outputting j outputs Output. In the examples of Expressions 2 and 3, there are two inputs R1 and R2, one output, and the example stored in R2. In order to make it difficult to estimate the data to be processed from the current waveform during the processing of Equation 24, the following processing is performed.

InputX (i) = h (Input (i), X (i)) (Equation 25)
OutputX (j) = f (inputX (i)) (Formula 26)
Xoutput (j) = f (X (i)) (Formula 27)
Output (j) = g (OutputX (i), Xoutput (i)) (Formula 28)
Expression 25 is a transformation process for transforming the input data Input (i) using the disturbance data X (i) and creating the transformed input data InputX (i). h is a deformation operation. Expression 26 is data processing for performing data processing using the deformed input data. Expression 27 is a disturbance data process for processing the disturbance data in the same manner as the input data. Expression 28 represents the modified input data data processing result OutputX (j) and the disturbance data processing result Xoutput (j). This is an inverse deformation process for performing inverse transformation of deformation. g is a reverse deformation operation.

In the previous example, Equations 4 and 5 correspond to the processing of Equation 25, and the transformation operation h is an exclusive OR. Equations 6 and 7 correspond to the data processing of the modified input data of Equation 26. Expressions 8 and 9 are data processing of the disturbance data of Expression 27. Expression 10 is the inverse deformation process of Expression 28. The transformation operation g is exclusive OR.
The selection of the deformation operation h and the reverse deformation operation g is determined by the characteristics of the data processing f. In the processing of Equations 2 and 3, the exclusive OR is the transformation operation h and the inverse transformation operation g. This is because the shift operation, the XOR operation, etc. can select the exclusive OR as the reverse deformation operation g by selecting the exclusive OR as the deformation operation h. This is because the exclusive OR is the logical zero if the same data is exclusive ORed, and the operation that takes the exclusive OR is lost.

If the data processing f is addition / subtraction, addition or subtraction can be selected for the deformation operation h, and subtraction or addition can be selected for the reverse deformation operation. For example,
Output = Input (1) + Input (2)-Input (3) (Equation 29)
For operations like, perform the following transformation process,
InputX (1) = Input (1) + X (1) (Equation 30)
InputX (2) = Input (2) + X (2) (Equation 31)
InputX (3) = Input (3) + X (3) (Formula 32)
By processing the deformed input data, the processing result of the deformed input data can be obtained.

OutputX = InputX (1) + InputX (2)-InputX (3) (Formula 33)
Data processing is performed in the same manner as the disturbance data.

Xoutput = X (1) + X (2)-X (3) (Formula 34)
Then, reverse deformation processing is performed.

Output = g (Xoutput) = OutputX-Xoutput
      = InputX (1) + InputX (2)-InputX (3)-(X (1) + X (2)-X (3))
      = Input (1) + Input (2)-Input (3) (Formula 35)
Correct output data can be obtained. This is because, in addition and subtraction operations, a certain result is added and the operation is performed, and a correct result can be obtained by subtracting the value added to the final result.

If the data processing f is multiplication / division, the transformation processing and the inverse transformation processing can be realized by selecting multiplication or division for the transformation operation h and selecting division or multiplication for the inverse transformation operation. The reason for this is the same as addition and subtraction. In multiplication / division operations, a correct result can be obtained by multiplying (dividing) a certain value and dividing the final result by division (multiplication). From that .

  If the data processing f is addition / subtraction of remainder calculation, addition / subtraction of an integral multiple of modulus N can be selected for the transformation operation h. For example, consider the following addition / subtraction remainder calculation.

Output = Input (1) + Input (2)-Input (3) mod N (Equation 36)
The input Input (i) is transformed as follows.

InputX (i) = Input (i) + k (i) * N (Formula 37)
Then, an addition / subtraction remainder calculation is performed using the transformed input data.

OutputX = InputX (1) + InputX (2)-InputX (3) mod N (Formula 38)
This equation is obtained by using equation 37
OutputX = (Input (1) + k (1) * N) + (Input (2) + k (2) * N)-(Input (3) + k (3) * N
mod N
= (Input (1) + Input (2)-Input (3)) + (k (1) * N + k (2) * N-k (3) * N)
mod N (Formula 39)
And can be transformed. And it is a feature of the remainder operation
0 = k * N mod N (Formula 40)
, The value in the second parenthesis of Equation 39 is 0, and Equation 39 is
OutputX = (Input (1) + Input (2)-Input (3)) mod N (Formula 41)
It becomes. That is, the operation result of the modified input data is the same as the original operation result. This is an example in which the data processing of the disturbance data and the inverse transformation processing are unnecessary by using the feature of the remainder calculation. This is because the result of data processing of disturbance data
Xoutput (i) = k (1) * N + k (2) * N-k (3) * N mod N (Formula 42)
However, since it becomes 0 in the first place, the data processing of the disturbance data and the reverse deformation processing are not necessary .

  If the data processing f is multiplication of a remainder operation, a transformation operation h obtained by adding 1 to an integral multiple of the modulus N can be used. For example, consider the following modular multiplication operation.

Output = Input (1) * Input (2) * Input (3) mod N (Formula 43)
The input Input (i) is transformed as follows.

InputX (i) = Input (i) * (k (i) * N + 1) (Formula 44)
Then, a modular multiplication operation is performed using the transformed input data.

OutputX = InputX (1) * InputX (2) * InputX (3) mod N (Formula 45)
This equation is obtained by using Equation 44.
OutputX = (Input (1) * (k (1) * N + 1)) * (Input (2) + (k (2) * N + 1) * (Input (
3 ) m od N
= (Input (1) * Input (2) * Input (3)) *
((k (1) * N + 1) * (k (2) * N + 1) * (k (3) * N + 1)) mod N (Equation 46)
And can be transformed. And it is a feature of the remainder operation
0 = k * N mod N (Formula 47)
By using, Equation 46 becomes
OutputX = (Input (1) * Input (2) * Input (3)) * (1 * 1 * 1) mod N
= Input (1) * Input (2) * Input (3) mod N (Formula 48)
It becomes. That is, the operation result of the deformed input data is the same as the real operation result. This is also an example in which the data processing of the disturbance data and the inverse transformation processing are unnecessary by using the feature of the remainder calculation .

  Further, if the data processing f is multiplication of the remainder operation, the number X in the transformation operation h and Y that is the reciprocal number with respect to the modulus N can be used.

1 = X * Y mod N (Formula 49)
One such simple number is one where X is 2 and Y is (N + 1) / 2. This is to obtain an actual processing result by multiplying X as a deformation operation and multiplying Y by the number of times X is multiplied by reverse deformation. For example, consider the following addition / subtraction remainder calculation.

Output = Input (1) * Input (2) * Input (3) mod N (Formula 50)
The input Input (i) is transformed as follows.

InputX (i) = Input (i) * X (Formula 51)
Then, an addition / subtraction remainder calculation is performed using the transformed input data.

OutputX = InputX (1) * InputX (2) * InputX (3) mod N (Formula 52)
When the inverse deformation process g is selected by multiplying Y by the number of times multiplied by X, the correct result can be obtained by operating the inverse deformation process g on OutputX.

Output = OutputX * Y * Y * Y mod N
= InputX (1) * InputX (2) * InputX (3) * Y * Y * Y mod N
= Input (1) * X * Input (2) * X * Input (3) * X * Y * Y * Y mod N
= Input (1) * Input (2) * Input (3) * X * X * X * Y * Y * Y mod N
= Input (1) * Input (2) * Input (3) * X * Y * X * Y * X * Y mod N
= Input (1) * Input (2) * Input (3) mod N (Formula 53)
In the equation 53, the property of the equation 49 is used. In this example, the data processing of the disturbance data is not necessary, but a correct result can be obtained by multiplying the reciprocal number by the inverse transformation process by the modulus N as many times as the number of times the disturbance data is multiplied .

  So far, the data processing f has been an example of data operation, but disturbing the data extraction operation from the table is also necessary in order not to estimate the data from the current waveform. An embodiment relating to disturbance of table data and disturbance of table address will be described using an example of extracting data from the table of FIG.

  33 is exclusive ORed with the disturbance data X1. Here, “9” is selected as the disturbance data, and the exclusive OR of the table value and 9 is executed. The result is the table of FIG. Next, in order to disturb the address of the table, exclusive OR of the row number and “3” selected as the disturbance data X2 is executed, and the column number and “2” selected as the disturbance data X3 are obtained. Perform exclusive OR and sort the table. The result is shown in FIG. 33, “0”, which is the data 3301 in the first row and the second column of the table of FIG. 33, is “9” (3401) in the table of FIG. 34 as a result of performing exclusive OR with the disturbance data X1. It becomes. As a result of executing exclusive OR of the row number and the column number with the disturbance data X2 and the disturbance data X3, respectively, 3401 moves to a location 3501. Such a table is created in order to disturb the data extraction operation from the table.

  It is assumed that the row number variable Gyou and the column number variable Retsu have already been transformed by exclusive OR with the disturbance data Y1 and Y2 until the address calculation. In other words, the correct row number Gyou and column number Retsu cannot be obtained unless Y1 and Y2 are exclusive ORed with Gyou and Retsu. The relationship is as follows.

Gyou = GyouY1 xor Y1 (Formula 54)
Retsu = RetsuY2 xor Y2 (Formula 55)
However, performing this reverse deformation process and using the table of FIG. 33 uses the true address data, so there is a possibility that the address data is estimated from the current waveform. Therefore, first, the disturbance data X2 and the disturbance data X3 used to disturb the row number and the column number when the table of FIG. 35 is created are used.

GyouY1X2 = GyouY1 xor X2 (Formula 56)
RetsuY1X3 = RetsuY2 xor X3 (Formula 57)
Then, reverse deformation processing of the disturbance data that has been used so far is performed.

GyouX2 = GyouY1X2 xor Y1 (Formula 58)
RetsuX3 = RetsuY2X3 xor Y2 (Formula 59)
By doing so, since true row numbers and column numbers are not used, it is difficult to estimate the values from the current waveform. Using GyouX2 and RetsuX3, refer to table TableX1X2X3 in FIG.

DataX1 = TableX1X2X3 (GyouX2, RetsuX3) (Formula 60)
Since the table of FIG. 35 has already been transformed with the disturbance data X1, processing is performed with the disturbance data as X1 thereafter. True data is not used in the processing from Expression 56 to Expression 60 .

  For each fixed process, the table data disturbance data X1, the row number disturbance data X2, and the column number disturbance data X3 are randomly generated to deform the table. By doing so, since the table is deformed for each process, it is difficult to estimate data from the current waveform.

  The type of disturbance data and the method of data transformation have been described above. Next, the processing procedure is shown. FIG. 4 shows an embodiment of an information concealment procedure using basic disturbance data.

  FIG. 4 shows the basic procedure. Disturbance data Xi is generated by the disturbance data generating means (401). As a general method, there is a method of generating a random number having a required length using a random number generator or a pseudo-random number. Next, the data transformation processing means (406) transforms the input data D1 with the disturbance data Xi (405) to generate the transformed data H1 (407). As described above, the transformation methods include exclusive OR, addition / subtraction, multiplication / division, and the like. Then, the modified data processing means (408) performs data processing using the modified data H1 to generate processed modified data H2. On the other hand, the disturbance data Xi is subjected to the same data processing as the input data by the disturbance data processing means (403) to create processed disturbance data Xo (404). Then, using the disturbance data Xo and the processed deformation data H2, the data reverse deformation processing means (410) obtains the true processed data D2 (411). The methods of the data transformation processing means (406) and the data inverse transformation processing means (410) include the already described exclusive OR, addition / subtraction, multiplication / division, and remainder calculation.

  The embodiment of FIG. 5 is an example using two pieces of disturbance data, and the information hiding process using the second disturbance data is included in the information hiding process using the first disturbance data. This is the case.
The main flow is the same as in the embodiment of FIG. The transformation data H1 (507) transformed by the first disturbance data is processed, and the second transformation data X2i is used as the processed transformation data H2 (509), which is the processing result. Deformation is performed by means (510), and processed deformation data H3 (511) is created. The data is processed by the second modified data processing means (512), the processed modified data H4 is generated, and the inverse transformation of the second disturbance data is performed by the second data inverse modified processing means (520). Processed deformation data H5 (521) is generated. Then, inverse transformation of the first disturbance data is performed by the first data inverse transformation processing means (514) to obtain the true processed data D2 (515). An example of using exclusive OR for transformation is as follows.

H1 = D1 xor X1i
H2 = f1 (H1)
X1o = f1 (X1i)
H31 = H2 xor X2i
H32 = D2 xor X2i (Formula 61)
H4 = f2 (H31, H32)
X2o = f2 (X2i, X2i)
H5 = H4 xor X2o
D2 = H5 xor X1o
Here, f1 and f2 are data processing operations. As in this example, when another data D2 is used in the second data processing f2, and the data D2 is used after being transformed with the second disturbance data, the processing procedure of this embodiment is used. And effective.

  The embodiment of FIG. 6 is also an example using two pieces of disturbance data. A major difference from the embodiment of FIG. 5 is the case where the information concealment process using the first disturbance data and the information concealment process using the second disturbance data are connected. This is a procedure for transforming with the second disturbing data and concealing the process using the true data before the reverse deformation process for returning the conversion by the first disturbing data. The modified data H1 (609) transformed by the first disturbance data is subjected to data processing by the first modified data processing means (610), and the processed transformed data H2 (611) is converted into second modified data creating means ( In step 612), deformation is performed using the second disturbance data, and processed deformation data H3 (613) is generated. Then, inverse transformation of the first disturbance data is performed by the first data inverse transformation processing means (605). Using the processed deformation data H4 (606) that is the processing result, the second deformation data processing means (614) is performed to generate processed deformation data H5 (615). Then, the second data inverse transformation processing means (616) performs inverse transformation to generate true processed data D2 (617). An example of using exclusive OR for transformation is as follows.

H1 = D1 xor X1o
H2 = f1 (H1)
X1o = f1 (X1i)
H3 = H2 xor X2i (Formula 62)
H4 = H3 xor X1o
H5 = f2 (H4)
X2o = f2 (X2i)
D2 = H5 xor X2i
This is effective when there are a plurality of processing operations and it is necessary to use a plurality of disturbance data.

  In the embodiment of FIG. 7, from the viewpoint of the disturbance data, data processing is calculated in advance to improve the processing efficiency. The processed disturbance data Xo is generated in advance by the disturbance data processing means (703), and recorded by the processed disturbance data recording means (706). In the processing, the data inverse transformation processing means (713) reads the recorded processed disturbance data (714) and uses it. This is effective in terms of efficiency when the same data processing is executed many times. However, since the disturbance data is used many times, it is more effective to change the disturbance data every time as in the embodiment of FIG. This is determined by a trade-off between processing speed and information concealment.

In the embodiment of FIG. 8, the inverse transformation of the first and second disturbance data is performed in an integrated manner, and thereafter, the result is used to inversely transform the data. As the first and second disturbance data, processed disturbance data X1o and X2o are generated by the first disturbance data processing means (803) and the second disturbance data processing means (807), respectively. These data are integrated by the data reverse transformation integration means, and the integrated disturbance data Xo is generated. Using this data, the inverse transformation processing means (820) of the processed transformation data H4 (819), which is the result of processing the first transformation data processing (814) and the second transformation data processing (818), is performed, and the true processing is performed. Generated data D2. This is a method in which the processed disturbance data is integrated, and then the deformation processing is performed collectively, rather than the case of individually performing reverse deformation. This is effective when the reverse deformation process takes a long time .

  Next, an embodiment using symmetric encryption DES (data encryption standard) as an example will be described. The present invention can also be used for other ciphers. Cryptography is described in “Introduction to Cryptography” by Eiji Okamoto, Kyoritsu Publishing Co., Ltd.

  DES encrypts and decrypts 64-bit data (plaintext or ciphertext) with a 56-bit encryption key. Since the same encryption key is used for encryption and decryption, it is called symmetric encryption. DES performs encryption by randomly exchanging bits of plaintext (text to be encrypted) so as to make it random by playing cards. Data exchange is performed according to the encryption key. When decrypting, the data is replaced in the reverse order according to the encryption key, and the data is restored. There are two replacement methods in the DES processing: a bit unit and a unit of a plurality of bits. The former is called transpose and the latter is called substitution.

  The DES encryption method will be described with reference to FIG. The deformation process a (901), the deformation process b (904), and the reverse deformation process (916) relate to the present invention and are not related to the original encryption process of DES. The ciphertext is first transposed in an initial transposition (IP) 902. This replaces the 64-bit data of the ciphertext in bit units according to the initial transposition table. Thereafter, a set of operations is performed 16 steps up to the initial transposition (IP-1) 916.

  In each one-stage process, the first half or the latter half of the result of the previous stage and the 32-bit data and the encryption key are input and a process called an f function 903 is performed, and the output is exclusive logic using the remaining half of the previous stage. An operation of taking the sum 909 is performed. The encryption key is also exchanged. First, selective transposition PC-1 (905) using a table called PC-1 is performed on the encryption key. Thereafter, selective transposition PC-2 (908) using a table called PC-2 is performed, and the encryption key is replaced. In the next stage, 28 bits are cycled and used according to the LS table.

  In the present embodiment, before the IP processing, a transformation process a (901) for transforming plaintext, a transformation process b (904) for transforming the encryption key, and a process (916) for reverse transformation at the end are performed. to add. In the transformation process a (901), the plaintext is transformed and the current of the process is processed by processing the transformed plaintext instead of processing the plaintext itself by the IP processing (902) or the f function (903). The plaintext data cannot be estimated from the waveform. In the transformation process b (904), the PC-1 process (905), the LS process (907), the PC-2 process (908), and the f function (903) do not process the encryption key itself, but the modified encryption key is processed. By processing, it is difficult to estimate the encryption key from the current waveform of the processing.

  The processing of the f function 303 is shown in FIG. First, the input sentence to the f function is selectively transposed based on the selective transposition matrix E (1002). Next, an exclusive OR is performed between the input data subjected to the selective transposition and the encryption key (1003), the S box processing is performed (1004), and the P transposition processing is performed (1005). The S box process is a process of taking out 6 bits from 48 bits, which are the processing result of the exclusive OR of 1003, and obtaining the row numbers and column numbers of the 8 S box tables to generate 4-bit data. Each S box table differs depending on the position of the data shown every 6 bits. The P transposition process is an operation of exchanging 32-bit bit positions according to the P transposition table.

  The deformation process a (901) and the deformation process b (902) are basically the same process. With reference to FIG. 11, the plain text deformation data creation process of the deformation process a will be described. Disturbance data X1 is randomly generated. This is generated using a random number generator or a pseudo-random number for each DES encryption (or decryption) process (1102). Use different disturbance data each time. Next, the disturbance data X1 and the plaintext P are subjected to XOR (exclusive OR) to generate a modified plaintext (transformed plaintext) PX1 (1103). In the case of DES, the plain text is 64 bits, but the generated random number may be 64 bits or 8 bits. However, if it is 64 bits or less, it is necessary to perform expansion and generate 64-bit disturbance data X1. If the generated random number is 8 bits, it may be repeated 8 times to generate 64-bit disturbance data X1. Here, since the transformation is performed using exclusive OR (XOR), the plaintext P is generated when the disturbance data X1 and the modified plaintext PX1 are XORed.

  FIG. 36 shows a procedure for creating modified data of the encryption key in the deformation process b (904). 11 is different from the embodiment of FIG. 11 in that the encryption key K and the disturbance data X2 are used instead of the plain text and the disturbance data X1. In DES, the encryption key is the same 64 bits as plain text. A cryptographic key KX2 transformed by the processing is generated.

Next, IP processing (902) will be described. In the IP processing, the plaintext 64-bit sequence is switched according to the table shown in FIG. According to the table, the first bit of the output is replaced with the 58th bit of the input, the second bit of the output is replaced with the 50th bit of the input, and the 64th bit of the output is replaced with the seventh bit of the input. The IP processing in this embodiment will be described with reference to FIG. First, the modified plaintext PX1 is IP-processed to generate an IP-processed transformed plaintext PX1IP (1202). Bit replacement follows the table in FIG. Next, the disturbance data X1 is similarly IP-processed to generate IP-processed disturbance data X1IP (1203). If the IP-processed modified plaintext PX1IP and the IP-processed disturbance data X1IP are exclusive-ORed, a result of IP processing of the plaintext can be generated. This is because the IP processing is bit movement, and the disturbance data is also moved in the same way as the modified plaintext PX1, so that the relationship that true data is obtained when exclusive OR is performed for each bit is maintained. It is. The lower 32 bits of the IP processing are used for the first stage f function (903) and the second stage exclusive OR, and the upper 32 bits are the input of the exclusive OR (909) .

  In the IP processing, the value of the bit of the modified plaintext PX1 is different from the value of the original plaintext bit, and it is difficult to estimate the plaintext data even when looking at the current waveform of the IP processing. Although the current consumption increases depending on the number of “1” bits, the number of “1” bits in the modified plaintext has nothing to do with the number of “1” bits in the plaintext. Therefore, it is difficult to estimate plaintext data from the size of the current waveform. In this way, by transforming plaintext with the disturbance data, it is possible to make it difficult to estimate the original data even if the current waveform being processed is observed.

  PC-1 processing is almost the same as IP processing. Using the conversion table for PC-1 in FIG. 38, the 64-bit encryption key is made 56 bits by removing 8 bits of parity bits, and the order of bits is changed. The way of viewing the table of FIG. 38 is the same as the way of viewing the table of FIG. When the modified encryption key KX2PC1 processed with PC-1 and the disturbance data X2PC1 processed with PC-1 are subjected to exclusive OR, a correct encryption key processed with PC-1 is obtained.

In the LS process, the 58-bit key generated in the PC-1 process is divided into 28 bits to the right and left, and each is shifted 1 bit or 2 bits to the left according to the LS table. An embodiment will be described with reference to FIG. First, at 1502, the result of the PC-1 processed modified encryption key KX2PC1 is LS processed to generate PC-1 and LS processed modified encryption key KX2PC1LS. At 1503, the PC-1 processed processed encryption data X2PC1 The result is subjected to LS processing to generate PC-1 and LS processed disturbance data X2PC1LS. Since the LS process is also a bit position exchange, if the exclusive OR of the PC-1 and LS processed modified encryption key KX2PC1LS and the PC-1 and LS processed disturbance data X2PC1LS is executed, the encryption that processed the true LS The key result is obtained. In the LS processing, since the disturbance data is used, the actually operated data is different from the true encryption key, and therefore it is difficult to estimate the encryption key even if the current waveform is observed.

  In the PC-2 processing, 56 bits of the LS processing result are reduced and transposed to 48 bits according to the PC-2 table. In 1402, the PC-1 and LS processed modified encryption key KX2PC1LS are subjected to PC-2 processing to generate PC-1 and LS, PC-2 processed modified encryption key KX2PC1LSPC2. In 1403, the PC-1 and LS processed disturbance data X2PC1LS are subjected to PC-2 processing to generate PC-1 and LS, PC-2 processed disturbance data X2PC1LSPC2. Since it is basically transposition using a table, it is basically the same as PC-1 processing.

  Next, processing of the f function 903 will be described. As shown in FIG. 10, the f function includes a selective transposition E process (1002), an exclusive OR of the encryption key and the selective transposition execution result (1003), an S box process (1004), and a P transposition process (1005). ).

  The selective transposition E process will be described with reference to FIG. Selective transposition E is to change the order of bits using the transposition table of FIG. At 1602, the IP processed modified plaintext PXIP is selectively transposed E to generate IP processed and E transposed modified plaintext PXIPE. In step 1603, the IP-processed disturbance data XIP is selectively transposed E to generate IP-processed and E-transposed-processed disturbance data XIPE. As in the case of IP processing and PC-1 processing, if the modified plaintext PXIPE after IP processing and E transposition processing and the disturbance data XIPE after IP processing and E transposition processing are XORed, correct IP processing and E transposition processing are performed. Finished plaintext. In addition, when performing transposition E table, the value of the bit is the value of the bit that has been transformed, so it is not possible to estimate the true data even by looking at the current waveform of the process. Have difficulty.

  Next, an exclusive OR process of the encryption key and the result of the selective transposition, which is the second process in the f function, is performed. This process will be described with reference to FIG. In 1702, XOR of the plain text PXIPE after IP processing and E transposition processing generated from the plain text and the PC-1 and LS, PC-2 processed modified cipher key KX2PC1LSPC2 generated from the encryption key, and the input of the S box processing 48-bit S box input data SinputX is created. Next, in 1703, the IP processing and E transposition processed disturbance data XIPE generated from the plaintext disturbance data, and the PC-1 and LS generated from the encryption key disturbance data and the PC-2 processed disturbance XOR the data X2PC1LSPC2 to generate S box input data disturbance data XSinput, which is disturbance data for the S box input data SinputX. Due to the nature of exclusive OR, S box input data disturbance data XSinput is composed of two disturbance data (PC-1 and LS, PC-2 processed disturbance data X2PC1LSPC2, PC-1 and LS, PC-2 It can be generated by XORing the processed disturbance data X2PC1LSPC2). This is illustrated by a simple example. Here, the plaintext is P and the key is K, the modified plane is PX1, and the modified key is KX2. Their relationship is Equation 63 and Equation 64. X1 and X2 are plaintext and key disturbance data.

PX1 = P xor X1 (Formula 63)
KX2 = K xor X2 (Formula 64)
Then, if the result of executing the exclusive OR of P and K is Z, the relationship between the results of executing the exclusive OR of PX1 and PX2 Z1 and Z is as follows.

Z = P xor K (Formula 65)
Z1 = PX1 xor KX2
   = (P xor X1) xor (K xor X2)
   = P xor X1 xor K xro X2
   = (P xor K) xor (X1 xor X2)
   = Z xor (X1 xor X2) (Formula 66)
That is, it can be understood that it is sufficient to use exclusive OR of P and K disturbance data as disturbance data for returning Z1 to true data. In the exclusive OR processing of the encryption key and the selective transposition execution result, as the disturbance data of the S box input data SinputX, the IP data generated from the plaintext disturbance data and the E transposed disturbance data XIPE It can be seen that the S box input data disturbance data XSinput generated by XORing the PC-1 and LS generated from the encryption key disturbance data and the PC-2 processed disturbance data X2PC1LSPC2 may be used.

  Next, the processing of the S box will be described with reference to FIG. Six bits are extracted from the S box input data SinputX, and processing is performed for eight S boxes. FIG. 25 shows the first S box used in DES. The eight S boxes have the same format, but the field data is different. In each S box process, first, the upper i-th 6-bit sub-data SubSinputX (i) of the S box input data Sinput is extracted (1805). XOR is performed on the address disturbance data Xsa (i) and SubSinputX (i) in the modified S box table obtained by modifying the previously created S box to generate SubSinputXXsa (i) (1806). The SubSinputXsa (i) is XORed with the upper i-th 6 bits of the S box input data disturbance data XSinput to generate SubSinputXsa (i) (1807). SubSinputXsa (i) is a value obtained by exclusive-ORing address disturbance data Xsa (i) with real address data for extracting the i-th S box. When XORing SubSinputX (i) and XSinput (i), it returns to the true data, so before XORing with XSinput (i), SubSinputX (i) and Xsa ( Perform exclusive OR of i), and then execute exclusive OR (XOR) on XSinput (i). By doing so, it is not necessary to process real data, and it is difficult to estimate data from the current waveform. Next, the address of the modified S box table is calculated using SubSinputXsa (i) (1808). Since the address for referring to the original S box table is modified, the table needs to be modified in advance. Based on the calculated address, S box output data SoutputX3 (i) is extracted from the modified S box table S (i) (1809). At the same time, the disturbance data X3 (i) for the output data SoutputX3 (i) of the S box is extracted (1810). Process 8 S-boxes and generate SoutputX3 and X3 by concatenating SoutputX3 (i) and X3 (i) with i from 1 to 8, respectively. After this, the processing data is SoutputX3 and the disturbance data is X3.

  Next, a method for generating the modified S box table will be described with reference to FIGS. In this S box, S (i) box address disturbance data Xsa (i) and data disturbance data X3 (i) are created (2306). Xsa (i) is 6 bits and X3 (i) is 4 bits. The disturbance data X3 is 32-bit data obtained by collecting 8 pieces of X3 (i) created every 4 bits. Next, a modified S (i) box table creation routine is called (2307). The processing of the i-th modified S box table creation routine will be described with reference to FIG. k designates a row number, and l designates a column number. The processing of k rows and l columns is 2408 to 2413. The table of the first S box is as shown in FIG. First, the data d of the row number k and column number 1 of the i-th original S box is taken out (2408). Then, the data d and the disturbance data X3 (i) are exclusive-ORed to obtain d2 (2409). If the disturbance data is “7”, the result is 2604 with respect to the original S box data 2504. When this is performed for all fields, the result is as shown in FIG. The table in FIG. 26 is obtained by exclusive-ORing the disturbance data as “7” with respect to the data in the first S box in FIG.

Next, the address is disturbed. First, let Xsa1 be 2 bits made from the upper 1 bit and the lower 1 bit of Xsa (i), and Xsa2 be 4 bits of data made from the upper 2 bits of Xsa (i). This is derived from the method for calculating the address of the S box. Then, assuming that the row number and the column number in FIG. 26 are k and l, respectively, Xsa1 and Xsa2 are exclusive-ORed (2412). If the newly created row numbers and column numbers are k2 and l2, the data d2 is stored in the k2 row l2 column of the i-th modified S box table S (i) (2413). An example of this processing is shown in FIG. FIG. 27 is created by selecting “2” and “9” as disturbance data for the rows and columns of FIG. To make it easier to understand, only the row and column numbers are changed while the row and column data positions remain unchanged. In FIG. 25, the data 12 (2504) in 3 rows and 1 column is moved to 1 row and 8 columns in FIG. In this example, the data disturbance data is “7”, the address disturbance data has a row of “2”, and a column of “9”. In this way, the eight S boxes are deformed. In this embodiment, this processing is performed at the beginning of DES. The modified S box table is used in 16 stages of DES .

  When the processing of the S box is completed, the processing data becomes 32-bit SoutputX3, and the disturbance data becomes 32-bit X3. This is an input of the transposition P process (1005) which is the last process of the f function. The transposition P process will be described with reference to FIG. SinputX3 that is the output of the S box is transposed P to generate SinputX3P (1902). The perturbation data X3 of SinputX3 is transposed P to generate X3P (1903). The table used for the turnover P is as shown in FIG. The usage of this table is the same as the IP processing table.

  When the processing of the f function is completed, transposition P processing and XOR of the result of the previous stage are performed (909 or 914). First, SinputX3P obtained by transposing the result of the S box is XORed with the result of the previous stage (2002). X3P and the previous disturbance data X are XORed (2003). This XOR processing is the same as the XOR processing (1701) of the selective transposition E processing result and the encryption key.

  In DES, IP-1 processing (915) is finally performed. This process is shown in FIG. The IP-1 process is a process of exchanging bit sequences similar to the IP process, and uses the IP-1 table instead of the IP table (2102). The processing result so far is IP-1 processed, and similarly, the disturbance data X is IP-1 processed (2103).

  Finally, in order to return to the correct processing result, reverse deformation processing is performed (916). The reverse deformation process is shown in FIG. The correct result is obtained by XORing the IP-1 processing result with the disturbance data X subjected to the IP-1 processing. Here, for the first time, a correct processing result that is not deformed is obtained.

  Although the concealment of data to be processed has been described so far, it may be necessary to conceal the disturbance data. The basic idea is that the disturbance data is exclusive-ORed with the disturbance data XR for disturbance and transformed. However, XR is fixed, and XRo for reverse deformation is obtained in advance by calculating the replacement of bit positions and the like. Then, when the disturbance data is required, the original disturbance data is obtained using XRo, so that the deformation and reverse deformation of the disturbance data are performed and the efficiency is improved. First, description will be made by taking the disturbance data for the encryption key as an example. The process of FIG. 30 is a transformation operation for exclusive ORing the disturbance data for the encryption key with the disturbance data XR for disturbance. After the disturbance data X2 is generated in the deformation process b (3601), the deformation process of the disturbance data in FIG. 30 is performed. The disturbance data X2 for the encryption key undergoes PC-1 processing, LS processing, and PC-2 processing. Since these are replacements of predetermined bit positions in advance, the disturbance data XRo when the processing up to the PC-2 processing is executed is obtained and recorded with respect to the predetermined value XR (from 3102). 3105). After PC-2 processing, XOR the disturbance data XRo recorded in PC-1 and LS generated in 1403 and PC-2 processed disturbance data X2PC1LSPC2 (3202), and the original PC-1 and LS and PC-2 processed disturbance data X2PC1LSPC2 can be obtained. By doing so, the disturbance data can also be concealed. Here, the disturbance data XR for disturbance and the processed disturbance h urine data XRo may be the same data.

  Although the embodiment in DES is encryption, the DES algorithm is almost the same in decryption, and therefore, this embodiment can be applied with almost no change. In addition, since cryptographic algorithms other than DES also use a lot of transposition processing, substitution processing, and remainder operation, it is difficult to apply the present invention to transform the data and estimate the original data from the current waveform. Can do.

IC card hardware configuration. Hardware configuration in the IC card chip. Current consumption waveform. Data transformation procedure using one disturbance data. Nested data transformation procedure using two disturbance data. Data transformation procedure using two disturbance data in succession. A data transformation procedure in which the data processing of disturbance data is calculated in advance. A data transformation procedure integrating the inverse transformation of two disturbance data. DES overall process flow. Flow of processing of f function of DES. Deformation process a. IP processing. PC-1 processing. PC-2 processing. LS processing. Selective transposition E processing. XOR processing of selective transposition E processing result and encryption key. S box processing. Transposition P process. Transposition P processing and XOR processing of the previous result. IP-1 processing. Reverse deformation process. Creation of table of modified S box. i-th modified S box table creation routine. The first S box table. A table obtained by modifying the data of the first S box table. A table obtained by modifying the arrangement of the first S box table. Selective transpose E table. Transposed P table. Encryption processing of disturbance data. Deformation calculation processing of encryption data for disturbance data. Decoding processing for disturbance data Example of original table. The table which deform | transformed the content of FIG. The table which changed arrangement | positioning of the table of FIG. Deformation process b. IP transposition table. PC-1 selective transposition table.

Claims (6)

A computing device for computing data;
A storage device for storing data and programs;
The arithmetic unit has a data bus for accessing the storage device,
The arithmetic unit is
A disturbance data generation process for generating the first disturbance data;
A disturbance data generation process for generating the second disturbance data;
A first data transformation process for encrypting input data using the first disturbance data and outputting modified data;
A first modified data process for performing a first process, which is a predetermined process, on the modified data and outputting first processed modified data as a processing result;
A first disturbance data process for performing the first process on the first disturbance data and outputting first processed disturbance data as a processing result;
A second data modification process for further encrypting the first processed modified data using the second disturbance data and outputting second processed modified data;
A first data inverse deformation process for decoding the second processed deformation data and outputting a first processed deformation data as a processing result using the first processed disturbance data;
Performing a second process, which is a predetermined process, on the first processed modified data, and outputting a second processed modified data as a processing result;
A second disturbance data process for performing the second process on the second disturbance data and outputting the second processed disturbance data as a processing result;
Using the second processed disturbance data, further decoding the second processed modified data and outputting processed data as a processing result;
An information processing apparatus characterized by performing The information processing apparatus according to claim 1,
In the first and second disturbance data generation processing, a random number is used as the first and second disturbance data. The information processing apparatus according to claim 1,
The first and second data transformation processes are processes for performing an exclusive OR operation with the disturbance data,
The information processing apparatus according to claim 1, wherein the first and second data inverse transformation processes are processes for performing an exclusive OR operation with the processed disturbance data. The information processing apparatus according to claim 1,
The first and second data transformation processes are processes for performing addition / subtraction with the disturbance data,
The information processing apparatus according to claim 1, wherein the first and second data inverse transformation processes are processes for performing addition and subtraction with the processed disturbance data. The information processing apparatus according to claim 1,
The first and second data transformation processes are processes for performing multiplication and division operations with the disturbance data.
The information processing apparatus according to claim 1, wherein the first and second data inverse transformation processes are processes for outputting a result of a multiplication / division operation with the processed disturbance data as the processed data . The information processing apparatus according to claim 1,
The first and second data transformation processes are processes for performing an exclusive OR operation with the disturbance data,
The first and second processes are transposition processes or substitution processes,
The information processing apparatus according to claim 1, wherein the first and second data inverse transformation processes are processes for performing an exclusive OR operation with the processed disturbance data.

Images