Security Issues in QCA Circuit Design - Power Analysis Attacks - Field-Coupled Nanocomputing: Paradigms, Progress, and Perspectives - page 200

in Fig. 5 (a). Thus, these can be considered as static power. The second part of
the power consumption is dependent on the changes to the cell's polarisation,
which can be considered as dynamic power. It is clear from Fig. 5 (b)thatthe
dynamic power is significantly larger than the static power in a QCA cell. Since
the dynamic power is dependent on the polarisation changes, the total power
loss in a QCA cell can possibly provide information about its inputs.
Note that the absolute value of the power consumption is on the order of a
few hundredths of an electron volt, i.e., eV, which is extremely low. However,
a power analysis attack exploits the difference between the static power and
dynamic power rather than the absolute value of the power.

2.3 Power Dependence on Hamming Distance in CMOS
and QCA Gates

The fundamental of power analysis attack is based on the fact that the power
consumption of cryptographic circuits is dependent on the input data. The power
consumption of CMOS circuits is mainly divided into two parts: static power
consumption and dynamic power consumption [ 32 ]. The N-type and P-type MOS
transistors are used to build the CMOS cells so that the pull-up network and
pull-down network are not conducting at the same time. When there is no change
from the input signal, there is only a little leakage current flowing through the
turned-off transistor, which is termed the static power consumption. The case
when only static power is consumed in CMOS is shown in Fig. 6 for an inverter
that has a constant input.
While the input of the inverter is changing from '1' to '0' or from '0' to '1',
dynamic power is also consumed. The dynamic power is from two parts: the
first one is the load capacitance which needed to be charged and the second
one is a short circuit current occurs due to switched output signal. As shown
in Fig. 7 (a), when the input signal changed from '1' to '0', the inverter draw a
charging current from the power supply to the load capacitance. In the other
case as shown in Fig. 7 (b) when input is changing from '0' to '1', the previously
charged load capacitance dissipates energy. The power consumption of a CMOS
inverter under four cases are summarized in Table 1 . It is clear that the power
consumption of an CMOS inverter is dependent on the Hamming Distance (HD)
of its input. The HD between two inputs X 1 and X 2 is the Hamming Weight
(HW) of X 1 ⊕X 2 , where HW is the number of binary '1's and

⊕ represents XOR:

HD ( X 1 ,X 2 )= HW ( X 1 ⊕ X 2 ) . (9)

The power dependence on processed data of QCA circuits is studied here with
an inverter and a majority gate. The power consumption of a QCA inverter, as
shown in Fig. 8 , is provided in Table 2 under different tunneling energy levels.
The power consumption does not vary as the temperature range is small, from
1 . 0K to 8 . 0 K. It can be seen from the table that in the case of transitions from
0

→ 0 and 1 → 1, little static power is dissipated, while for transitions from 0

→ 1 and 1 → 0 significantly more power is consumed. This is similar to that of