The present invention relates to random signal generators, and more particularly, though not exclusively, to a random binary signal generator for smart cards.
A random binary signal generator allows a smart card, for example, during a terminal authentication procedure, to send a random binary sequence comprising 16 or 32 bits, for example. The card and terminal apply a secret key authentication function to the sequence. Thereafter, the terminal transmits the obtained result to the card, and the card compares its computed result with the one received from the terminal. If both results match, the terminal is presumed to be authentic and the card accepts the requested transaction.
Random generators are known in the art, and are implemented as state machines having a finite number of internal states. Such state machines comprise, for example, shift registers of which certain bits are fed-back through an XOR gate as an input to the machine. Starting from an initial internal state, the state machine is enabled by a clock signal, and a random bit is extracted from the state machine at each clock pulse.
A drawback of such state machines, however, is that they generate deterministic binary sequences that are highly repetitive, as well as an output statistical bias influencing the distribution of logic ones and logic zeros. To overcome this drawback, there is a need for state machines having a large number of internal states (ideally, the machine would have an infinite number of internal states), but this approach opposes the greater simplicity, lower cost and lower consumption desired for random signal generators.
A random signal generator can also be implemented by an electronic noise generator, such as the noise generated by an avalanche-connected diode or a biased diode, or also the thermal noise in a resistor. However, these noise sources are of very small amplitude. It is necessary to use a high gain amplifier with at least a 60 dB gain. This implies using power-consuming bipolar transistors. However, the presence of such transistors is undesired in many integrated circuits, in particular, integrated circuits used in smart cards.
In view of the foregoing background, an object of the present invention is therefore to provide a random signal generator that does not have the drawbacks of known generators.
This and other objects, advantages and features in accordance with the present invention are based on the observation that folded MOS transistors, such as having a zigzag-shaped channel, are known to be unusable when the technology employed is pushed to its limits. Folded MOS transistors are unusable in these situations because the drain-source current has an increasingly strong random component when the size of the zigzag-shaped channel decreases. A concept of the present invention is to take advantage of this drawback to provide a random signal generator that can be integrated within an integrated circuit, particularly, an integrated circuit for smart cards.
Thus, the present invention includes a random signal generator using an electronic noise source. The electronic noise source may comprise a folded MOS transistor whose drain-source current has a random component. The random signal generator also comprises means or a circuit for generating a random binary or digital signal from the random component.
According to one embodiment, the folded MOS transistor comprises an S-shaped or zigzag-shaped channel, in which the sizes are at the resolution limit allowed by the transistor manufacturing technology. According to another embodiment, the generator further comprises a reference transistor. The reference transistor receives a gate voltage and a bias current that are the same as those applied to the folded transistor for extracting the random component.
According to yet another embodiment, the generator comprises means or a circuit for comparing the random component to a detection current. The generator may comprise means or a circuit for amplifying the random component, and the generator may comprise means for sampling the random binary signal to obtain a random digital signal.
According to another embodiment, the generator comprises a logic circuit for generating random binary numbers from the random digital signal. The generator may automatically maintain the gate voltage of the folded transistor within a determined range of values that ensures delivery of an equally probable output signal.
According to another embodiment, the generator comprises a plurality of electronic noise sources generating a current that includes a random component. Each source is coupled to means or a circuit for generating a random binary signal from the random component generated by the source, respectively. The generator additionally may comprise means for combining the random binary signals from the sources to generate random binary signals.
The present invention also relates to an integrated circuit comprising a random binary signal generator according to the invention, and means or a circuit for connecting the generator output to other components in the integrated circuit. According to one embodiment, the integrated circuit is arranged on a support for forming a smart card or any other equivalent portable electronic item.
According to another embodiment, the integrated circuit comprises a processor unit including means or a circuit for receiving a random number generated by the generator, means or a circuit for transmitting the random number to an external terminal, means or a circuit for subjecting the random number to a secret key authentication function. The integrated circuit may further comprise means or a circuit for comparing the result of the function to a result provided by the terminal in response to the transmission of the random number, and means or a circuit for allowing a transaction with the terminal if the result provided by the terminal matches the result computed by the processor unit.
The present invention also relates to a method for generating a random number using an electronic noise source. The method comprises providing a folded transistor having an S-shaped or zigzag-shaped channel with sizes chosen to be at the resolution limit allowed by the transistor manufacturing technology, and extracting a random current component at the terminals of the folded MOS transistor. The method further comprises generating a binary or digital signal as a function of the random component, and sampling the binary signal.
According to one embodiment, the method further comprises amplifying the random component and subtracting therefrom a reference value before converting the random component into a binary signal. The method may also comprise the step of adjusting the gate voltage of the folded transistor as a function of the random binary signal obtained in the sampling step. The method may further comprise generating binary numbers from the binary signal.
These and other objects, features and advantages of the invention will become clear from the following description of a generator according to the invention given by way of a non-limiting example, with reference to the enclosed drawings, in which:
a is a detailed view of transistor Mc used as the noise source in the generator shown in
b is a detailed view of the component shown in
The generator 1 comprises two transistors M1, M2 arranged as a current mirror, which bias both transistors Mref and Mc by applying the same current I1 to the drain of transistors M1, M2. The source of transistor Mc is grounded through a balancing resistor R formed by the drain-source resistance of transistor Mc. Because the drain-source current in transistor Mc includes a random component, the drain current is I1+ΔI.
The generator 1 further comprises two transistors M3, M4, also arranged as a current mirror, with the source and gate of transistor M3 connected to the drain of transistor Mc. In this manner, transistor M3 measures the current difference ΔI=Ic−Iref between the drain currents of transistors Mc and Mref. The current I2, at the source of transistor M4, corresponds to the current ΔI multiplied by a gain α, namely I2=α*ΔI. Transistor M4 is oversized relative to transistor M3, so that the amplification ratio or gain α has a value greater than 1, such as 2, for example.
The logic signal is sampled by a D flip-flop 4 having a D input connected to the output of the inverter gate 2, and having a Q output delivering the random binary signal. The clock input CK of flip-flop 4 is driven by a clock signal Clk having a predetermined frequency. The clock signal Clk is output by an oscillator 3 that may be internal or external to the generator 1.
For the current I2 to oscillate randomly about the current IS, the signal at the Q output of the flip-flop 4 is fed-back to the gate of transistors Mref and Mc through an integrating circuit for automatically adjusting the gate voltage VG. The gate voltage VG thus ranges between values VGmin and VGmax, as shown in
Here, the integrating circuit comprises a PMOS transistor M5 and an NMOS transistor M6. The source of transistor M5 is connected to the drain of transistor M6 through two series-connected current sources S1, S2. The signal provided by output Q of the flip-flop 4 is applied to the gates of both transistors M5 and M6, and the junction node between the current sources S1, S2 is connected to the gates of transistors Mc and Mref. The gate of each transistor Mc and Mref is also connected to ground through a capacitor C, which through steady current charging and discharging, adds to the relatively high gate capacitances of transistors Mref and Mc for integrating the gate voltage VG.
Thus, when the signal at the output Q of flip-flop 4 is a 0, transistor M6 is off and transistor M5 is on. Current source S1 then applies a current to the connection node between the gates of transistors Mref, Mc and capacitor C. Conversely, if the logic level of the signal applied to the integrating circuit is a 1, transistor M5 is turned off and transistor M6 is on. The current source S2 then applies a current to the connection node between the gates of transistors Mref, Mc and capacitor C. Thus, the integrator device continuously tracks current changes within transistor Mc so that the gate voltage VG is kept in the range VGmin to VGmax, and bits set to a 1 and a 0 at the generator output have equal probabilities.
a-3b show transistors Mref and Mc in more detail. In
The polysilicon layer is coated with a metal contact providing the gate connection. The ends of each channel 14 are provided with contact pads 13, 19 and are series connected by metal contacts 17 shown in broken lines. Two contact pads 13, 19 at the ends of the chain-pattern thus obtained are provided with metal contacts 12, 18, respectively forming the drain and source connectors of the transistor. In other words, transistor Mref is formed by a plurality of series-connected transistors having a common gate.
In
Channel 24 as a whole is overlaid with a gate 21 composed of a thin electrically insulating layer, for example, polycrystalline silicon. The insulating layer is overlaid with a metal contact serving as the gate connection. Both ends of the channel include contacts 23, 29 connected to respective metal contacts 22, 28 providing the drain and the source of the transistor, respectively.
The drain-source current includes a random component when the channel dimensions are chosen to be close to the minimum resolution allowed by the manufacturing technology in use. The dimensions may even be slightly less than this minimum because defects may otherwise appear when developing the channel implantation mask. Therefore, as shown in the enlarged view in
The random component ΔI of the drain-source current increases as the number of bends in such a transistor increases. On the other hand, if the number of bends increases too much, the drain-source current decreases and has to be amplified with a large gain. Consequently, there is an optimum number of folds, which depends on the integration scale of the component. For example, there were 10 folds in the embodiment tested by the inventors.
The generator 1 may be coupled to a logic circuit 5 as shown in
Alternatively, as shown in
The generator illustrated in
The communications module 47, which is of the contact-type or contactless-type, is designed to cooperate with the corresponding communications module 57 provided in terminal 51. The terminal 51 comprises a processor unit 52, a memory 54 and an internal bus 53 allowing the processor unit to communicate with the memory and communications module 57.
The random generator 45 is used in an authentication procedure of terminal 51 as performed by the chip 41 for authorizing a transaction. For this purpose, the processor unit 42 uses generator 45 for generating a random number A, which is transmitted to terminal 51. In parallel, processor unit 42 computes a result R=FKS(A) obtained by transforming the random number A according to an authentication function FKS that uses a secret key KS. Terminal 51, which knows the secret key KS, performs the same computation and transmits the transformation result R′ to chip 41. If the chip finds that both results R and R′ match, it proceeds to the transaction requested by the terminal 51.
Preferably, terminal 51 also comprises a random generator 55, according to the present invention, for authenticating the chip 41. In this case, the terminal 51 executes a procedure similar to the one described above by sending a random number to the chip 41, and compares the computed result with the result that has been computed and transmitted by the chip.
Those skilled in the art will recognize that the random generator according to the present invention can be subject to many embodiments, modifications and applications within the scope to the above teaching.
Number | Date | Country | Kind |
---|---|---|---|
00 15309 | Nov 2000 | FR | national |
This application is a divisional of Ser. No. 09/995,258 filed Nov. 27, 2001, which is now U.S. Pat. No. 7,706,529, the entire disclosure of which is hereby incorporated herein by reference.
Number | Name | Date | Kind |
---|---|---|---|
4123771 | Uchida | Oct 1978 | A |
4862237 | Morozumi | Aug 1989 | A |
4905176 | Schulz | Feb 1990 | A |
5856688 | Lee et al. | Jan 1999 | A |
5961577 | Soenen et al. | Oct 1999 | A |
6209123 | Maziasz et al. | Mar 2001 | B1 |
6324558 | Wilber | Nov 2001 | B1 |
6756633 | Wang et al. | Jun 2004 | B2 |
6818936 | Lin et al. | Nov 2004 | B2 |
6882572 | Wang et al. | Apr 2005 | B2 |
6906389 | Depetro et al. | Jun 2005 | B2 |
7692188 | Cho et al. | Apr 2010 | B2 |
7706529 | Marinet et al. | Apr 2010 | B2 |
7922084 | Ishibashi et al. | Apr 2011 | B2 |
20040080356 | Hareland et al. | Apr 2004 | A1 |
20040214396 | Wang et al. | Oct 2004 | A1 |
20060108650 | Cho et al. | May 2006 | A1 |
Number | Date | Country |
---|---|---|
1209808 | May 2002 | EP |
Number | Date | Country | |
---|---|---|---|
20100180117 A1 | Jul 2010 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 09995258 | Nov 2001 | US |
Child | 12731808 | US |