Patent Grant
7201326
1. Field of the Invention
The present invention relates to an information processing device, and more particularly to a tamper proof device such as an integrated circuit card (IC card) having a high degree of security.
2. Description of the Related Art
The IC cards are intended for holding information that must not be tampered by encrypting data with secret encryption keys and decrypting the encrypted text. An IC card has no internal power source and becomes operable only when inserted into a card reader/writer by which it is powered. When it becomes operable, the IC card receives commands from the card reader/writer and transfers data as commanded. The general descriptions of IC cards can be found in books such as IC Card, Junichi Mizusawa, The Institute of Electronics, Information and Communication Engineers, published by Ohm.
As shown in
The structure of an IC card chip is basically the same as that of a microprocessor. As shown in
The memory unit includes a read-only memory (ROM), a random access memory (RAM), and an electrically erasable programmable read-only memory (EEPROM). ROM is not modifiable, and mainly stores program code. RAM is rewritable, but its contents are lost when power is off. RAM therefore cannot be used to retain data after the IC card is withdrawn from the reader/writer such that its power supply is stopped. EEPROM is rewritable, and it retains its contents even without power. EEPROM is used to store data that must sometimes be rewritten and must be retained even when the IC card is removed from the reader/writer. EEPROM is used, for example, in a prepaid card that retains data indicating the amount of use, which has to be rewritten at every use and must be retained after the card is withdrawn from the card reader/writer.
IC cards store programs and data inside an enclosed IC card chip so as to store important information and perform cryptographic processing. The degree of difficulty in deciphering cryptographic processing in IC cards has been considered to be similar to the difficulty of deciphering cryptographic algorithms. It is suggested, however, that there is a risk that information being cryptographically processed in IC cards and the cryptographic keys used for such processing may be inferred through observation and analysis of current consumption during the cryptographic processing, which is easier than deciphering cryptographic algorithms. The current consumption can be observed by measuring current that is supplied from the card reader/writer. Such risks are described in ‘8.5.1.1 Passive protective mechanisms’ p. 263 of Smart Card Handbook written by W. Rankl & W. Effing, John Wiley & sons Co.
The CMOS circuits in an IC card chip consume current when their output changes from ‘1’ to ‘0’, and vice versa. The data bus 203 has a particularly large electrical capacitance such that it draws a large current when the value placed on it changes from ‘1’ to ‘0’, or vice versa. This suggests the possibility that observation of the current consumption can reveal the operations inside the IC card chip.
Therefore, it is possible to infer which component is operating or what kind of data is being processing from the current consumption.
As countermeasures against such risks, the prior art provides two general methods: one method keeps the values of current consumption constant; the other method changes the current consumption while performing the same processing. An example of the former method provides a positive data bus, a negative data bus and a plurality of arithmetic units, which perform dummy and real operations concurrently to keep the current consumption constant regardless of the input data and operational results (PCT WO 99/67766). This method, however, raises problems of increased hardware scale, such as a doubling of the bus width and a quadrupling of the number of arithmetic units. As an example of the latter method, a method for encrypting data transferred on the bus or stored in memory has been suggested (JP-A-5731/2001). This method imposes a programming restriction, however, because the difference in life time of a plurality of data sharing the same key information places a limitation on the timing of updating of the encryption key.
An object of the present invention is to reduce the correlation between data being processed and current consumption in an IC card microprocessor chip without adding substantial hardware scale or programming restrictions.
According to one aspect of the invention, the information processing device includes a memory unit; an arithmetic unit; first encryption means for encrypting data written into/read from the memory unit or data input into/output from the arithmetic unit with a first cryptographic algorithm into first data including first key data and first encrypted data; and transfer means for transferring the first data into/from the memory unit or the arithmetic unit such that only encrypted data is transferred thereby.
According to another aspect of the invention, the information processing device includes an arithmetic unit; first encryption means for encrypting data to be input into the arithmetic unit into first data including first key data and first encrypted data; at least one decoder for decrypting the first data; at least one encoder for encrypting output of the arithmetic unit into second data including second key data and second encrypted data; transfer means for transferring data into/from the arithmetic unit such that only encrypted data is transferred thereby. The decoder and the encoder are disposed close to the arithmetic unit so as to reduce current consumption therebetween.
According to a third aspect of the invention, the information processing device includes a memory unit; an arithmetic unit; first encryption means for encrypting data written into/read from the memory unit or data input into/output from the arithmetic unit with a first cryptographic algorithm into first data including first key data and first encrypted data; and second encryption means for encrypting the first data with a second cryptographic algorithm into second data including second key data and second encrypted data to be stored in at least one of ROM/EEPROM and RAM of the memory unit.
The foregoing and additional features and characteristics of the present invention will become more apparent from the following detailed description considered with reference to the accompanying drawings in which like reference numerals designate like elements and wherein:
There are two methods of reducing the correlation between the data being processed and the current consumption: (1) changing the current consumption in an unpredictable way even if the value of data being processed is the same, and (2) keeping the current consumption constant even if the value of the data to be processed varies.
The present invention reduces the correlation between current consumption and data by the first method (1). Basically, this method makes the correlation between the true value of the data and its electrical expression unpredictable by encrypting the data and changing the encryption key at short time intervals. In this case, if the encryption key is shared among a plurality of data items, changing the encryption key entails the necessity of rewriting all of the data items sharing it. Therefore, the amount of data that shares an encryption key should be minimized, and more preferably should be the same as the size of the access unit. For example, keys may be provided per byte. The number of bits of a key may be 1 bit at the minimum. A 1-bit key can also be considered to be a selection bit for selecting two types of keys. A simple exemplary method is to invert the data when the key bit is a ‘1’, and not to invert the data when the key bit is a ‘0’. The cryptography used in this method can be considered to be a kind of Vernam cipher, equivalent to the use of 255 as a key for a key bit of ‘1’, and the use of 0 as a key for a key bit of ‘0’. The Vernum cipher performs encryption by an exclusive-or (xor) operation using a secret key K, and performs decryption by xoring the encrypted data again with the secret key K that was used for encryption.
Encrypted data=(key bit x 255) xor plaintext data (Eq. 1)
Suppose the decimal number 63 is to be encrypted and the key bit is ‘1’. The encryption equation can be expressed in binary notation as follows.
(1x11111111(b)) xor 00111111(b)=11000000(b) (Eq. 2)
If the key bit is added as the most significant bit, then 63 is encrypted to 448 as follows.
Encrypted data=111000000(b)=448 (Eq. 3)
The encrypted data can then be decrypted by taking the most significant bit as the key bit and the lower 8 bits as the encrypted data as follows.
Plaintext data=(key bit x 255) xor encrypted data (Eq. 4)
This can be expressed in binary notation as follows.
(1x11111111(b)) xor 11000000(b)=00111111(b)=63 (Eq. 5)
Other key values can be selected by the bits in (Eq. 1) and (Eq. 4). In this case, the two key data items selected by the one key bit should become all ‘1’s when xored. This is because two key data items having bits of the same value constantly produce the same encrypted value of the bit to be processed. The values 0 and 255 satisfy this condition. Therefore, if the data to be processed is 8-bit data, the value of the encryption key is K, and the key bit is 1 bit, the encryption and decryption equations are the following.
encrypted data=(not (key bit x 255) xor key value) xor plaintext data (Eq. 6)
and
plaintext data=(not (key bit x 255) xor key value) xor encrypted data (Eq. 7)
The key bit may be stored by adding it at the highest or any other bit position. Suppose the key value is 170 (=10101010(b)) and the plaintext data value is 123. If the key bit is ‘1’, then
If the key bit is added to the highest position, then the encrypted value becomes 209+256=465. Similarly, if the key bit is ‘0’, then
The value of the key bit is ‘0’, so the value obtained by adding the key bit is also 46. The applicable register is able to store encrypted data with the encryption bit. The encrypted data is stored as is and is decrypted only when the true value is necessary, such as at the time of arithmetic operations.
Next, in decryption, since 465 includes the key bit of ‘1’ and the encrypted data value of 209,
Similarly, since 46 includes the key bit of ‘0’ and the encrypted data value of 46,
The correct value is obtained by decryption in both cases.
If there is only one key bit, current consumption differs depending on whether the value of the key bit is ‘1’ or ‘0’. In a bus of the pre-charged type, the voltage of which is first raised, then decreased according to the value of the bit, current consumption varies depending on the value of the key bit and on whether discharging is performed after charging. Therefore, it is preferable to use a plurality of bits to express the key bit, to prevent variations in current consumption depending on the key bit value. It is assumed that two bits are used to express the key bit. If two bits (key bit 1 and key bit 2) are used to express key bits ‘0’ and ‘1’, the ‘1’ state of the key bit is expressed by (key bit 1=1, key bit 2=0), and the ‘0’ state of the key bit is expressed by (key bit 1=0, key bit 2=1).
An information processing device according to the present invention provides encryption and decryption units at positions such that a link to conventional memory units and buses is possible. More specifically, the information processing device disposes the decryption unit at the point from which data is output to the outside, and the decryption unit decrypts the data and then outputs the decrypted data to a conventional bus or memory unit. The encryption unit is disposed at the point from which data is input from the outside, and the encryption unit encrypts data from a conventional bus or memory unit and then passes the encrypted data into the information processing device.
There are two methods of performing arithmetic operations on encrypted data. One is to place a decryption unit at the input interface of the arithmetic unit and an encryption unit at the output interface from which operational results are obtained. The other is to construct an arithmetic circuit capable of operation on the encrypted data as is. A method of implementing an encrypted full adder that receives encrypted data as input and outputs encrypted arithmetic results is described by taking a 1-bit full adder as an example.
First, with an ordinary full adder that is not encrypted, suppose the two inputs are A′ and B′, the carry bit from the lower-order position is C′in, the carry bit to the higher-order position is C′out, and the addition result is R′. The relationships of A′, B′, C′in, C′out, and R′ can be expressed as follows.
R′=A′ xor B′ xor C′in (Eq. 12)
C′out=A′ and B′ or (A′ xor B′) and C′in (Eq. 13)
Next, an encrypted full adder is assumed. It is assumed that the first encrypted input data is A, the key bit of the first input data is Ak, the second encrypted input data is B, the key bit of the second input data is Bk, the carry bit produced to the higher-order position is Cout, the carry bit from the lower-order position is Cin, the addition result is R, and the encryption bit carrying Cout, Cin, and R is Rk.
In this case, in an encrypted full adder that inverts bits when the encryption bit is ‘1’, the relationship between A′, B′, C′in, C′out, and R′ of the input and output of the unencrypted full adder, and A, k, B, Bk, Cin, Cout, R, and Rk of the input and output of the encrypted full adder can be expressed as follows.
A′=A xor Ak
B′=B xor Bk
C′in=Cin xor Rk
C′out=Cout xor Rk
R′=R xor Rk (Eq. 14)
Substituting (Eq. 14) with equations (Eq. 12, Eq. 13) yields the logic expressions of the encrypted full adder as the following.
R=(A xor Ak xor B xor Bk xor Cin xor Rk) xor Rk=A xor B xor Cin xor Ak xor Bk (Eq. 15)
Cout=((A xor Ak) and (B xor Bk) or ((A xor Ak) xor (B xor Bk)) and (Cin xor Rk) xor Rk=((A xor Ak) and (B xor Bk) or (A xor B xor Ak xor Bk)) and (Cin xor Rk)) xor Rk (Eq. 16)
Using De Morgan's theorem,
X or Y=not ((not X) and (not Y)) (Eq. 17)
(Eq. 16) may be written as the following.
Cout=(((A xor Ak) nand (B xor Bk)) nand ((A xor B xor Ak xor Bk) nand (Cin xor Rk))) xor Rk (Eq. 18)
Equations (Eq. 15) and (Eq. 18) can be expressed in the logic circuit shown in
Next, it is assumed that an encrypted full adder that reverses bits when the encryption bit is ‘0’. The bits Cin and Cout are assumed to be inverted when the encryption bit is ‘1’. Then the relationships between A′, B′, C′in, C′out, and R′ and A, Ak, B, Bk, Cin, Cout, R, and Rk can be expressed by equation (Eq. 19).
A′=A xor not Ak
B′=B xor not Bk
C′in=Cin xor Rk
C′out=Cout xor Rk
R′=R xor not Rk (Eq. 19)
By substituting (Eq. 19) with equations (Eq. 12) and (Eq. 13) and applying De Morgan's theorem, the logic expression of the encrypted full adder that inverts bits when the encryption bit is ‘0’ can be determined in the following way.
R=(A xor (not Ak) xor B xor (not Bk) xor Cin xor Rk) xor (not Rk)=not (A xor B xor Cin xor Ak xor Bk) (Eq. 20)
Cout=(((A xor not Ak) nand (B xor not Bk)) nand ((A xor B xor not Ak xor not Bk) nand (Cin xor Rk))) xor Rk=(((A xor Ak) or (B xor Bk)) nand ((A xor B xor Ak xor Bk) nand (Cin xor Rk))) xor Rk (Eq. 21)
For example, an encrypted full adder for performing encryption with 0x55 when the encryption bit is ‘0’ and encryption with 0x AA when the encryption bit is ‘1’ can be implemented by an array of eight encrypted full adders expressed alternately by the logic circuit in
There are two ways to implement an instruction decoder that interprets and executes encrypted data: one is to connect a decryption unit at the point at which instructions from the instruction decoder are received into the information processing device; the other is to provide a decoder with a many-to-one correspondence between instruction codes and instructions such that the encrypted data can be directly interpreted and executed. Encrypting instruction codes yields a number of encrypted values equal to the number of different encrypted bit values. All of the values obtained through encryption are made to correspond to the instruction that corresponds to the original instruction code. A correspondence between all the instruction codes and instructions is established in this way, and a table showing the many-to-one correspondence between the encrypted instruction codes and instructions is generated. Constructing a decoder according to the table makes it possible to implement a decoder that can interpret encrypted instruction codes without decrypting them.
0x5C xor 0x55+0=0x009 (when key bit=0)
0x5C xor 0xAA+0x100=0×1F6 (when key bit=1)
Providing a many-to-one correspondence table of instruction codes and instructions by which the two values can be made to correspond to BSR can implement a decoder capable of interpreting and executing instructions without the need for decryption thereof.
An embodiment of the encrypted full adder 501 for encryption that inverts bits when the encryption bit is ‘0’ is shown in
According to the embodiments of the present invention, it is possible to provide information processing devices with higher security. It is also possible to provide IC card components and information processing systems with higher security.
The principles, preferred embodiments and modes of operation of the present invention have been described in the foregoing specification. However, the invention which is intended to be protected is not limited to the particular embodiments disclosed. The embodiments described herein are illustrative rather than restrictive. Variations and changes may be made by others, and equivalents employed, without departing from the spirit of the present invention. Accordingly, it is expressly intended that all such variations, changes and equivalents which fall within the spirit and scope of the present invention as defined in the claims, be embraced thereby.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2001-350936 | Nov 2001 | JP | national |
This application is a Continuation application of nonprovisional U.S. application Ser. No. 10/217,565 filed on Aug. 14, 2002 now U.S. Pat. No. 6,691,921. Priority is claimed based upon U.S. application Ser. No. 10/217,565 filed on Aug. 14, 2002 which claims the priority date of Japanese application 2001-350936 filed on Nov. 16, 2001.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4278837 | Best | Jul 1981 | A |
| 4465901 | Best | Aug 1984 | A |
| 4630201 | White | Dec 1986 | A |
| 4764959 | Watanabe et al. | Aug 1988 | A |
| 4827113 | Rikuna | May 1989 | A |
| 4956863 | Goss | Sep 1990 | A |
| 5224166 | Hartman, Jr. | Jun 1993 | A |
| 5323323 | Gilham | Jun 1994 | A |
| 5764772 | Kaufman et al. | Jun 1998 | A |
| 5850450 | Schweitzer et al. | Dec 1998 | A |
| 5889622 | Wille et al. | Mar 1999 | A |
| 6182104 | Foster et al. | Jan 2001 | B1 |
| 6233339 | Kawano et al. | May 2001 | B1 |
| 6363210 | Owashi et al. | Mar 2002 | B1 |
| 6408075 | Ohki et al. | Jun 2002 | B1 |
| 6414558 | Ryan et al. | Jul 2002 | B1 |
| 6419159 | Odinak | Jul 2002 | B1 |
| 6691921 | Endo et al. | Feb 2004 | B2 |
| 20010003540 | Pomer et al. | Jun 2001 | A1 |
| Number | Date | Country |
|---|---|---|
| 6-112937 | Jun 1993 | JP |
| 7-182304 | Dec 1993 | JP |
| 2001-005731 | Jun 1999 | JP |
| 2001-92654 | Sep 1999 | JP |
| 2001-237825 | Nov 2000 | JP |
| WO 9967766 | Jun 1999 | WO |
| Number | Date | Country | |
|---|---|---|---|
| 20040136530 A1 | Jul 2004 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10217565 | Aug 2002 | US |
| Child | 10745540 | US |
