Description
Side-channel attacks are often very difficult to achieve, because the attacked device is a
black box towards the attacker. The signals that the chip emits to the outside are often
measured in order to find interesting patterns. As soon as this pattern occurs within the
measured signal, the attack takes place to extract useful information. In this bachelor
thesis, the part of sampling a signal, analyzing it and finding a predefined pattern in it,
is discussed in detail. The main focus is on pattern recognition in order to find, test and
optimize algorithms. These algorithms are optimized so that they can be implemented
in an Field Programmable Gate Array ( FPGA) in order to reduce the computing time
and achieve pattern recognition in real time. In addition, the sampling of a signal, its
adjustment and its analysis are discussed by using the most fundamental methods
of signal processing. To test the effectiveness of the pattern recognition algorithms,
a known real world attack on the ATAES 132A micro chip is used. In a real world
implementation, a Software Defined Radio (SDR ) would be used to record and sample
the emitted signal by for example an ATAES 132A micro chip. The built in FPGA of the
SDR will perform the real time pattern recognition. Right after the pattern has been
found, the SDR generates a trigger which is used to start a laser fault injection attack.
If this hypothetical attack extracts the secret key in a real implementation, the concept
of black box triggering works.
|
Thesis topic in co-operation with the Fraunhofer Institute for Applied and Integrated Security AISEC, Garching