# A Survey of Electromagnetic Side-Channel Attacks and Discussion on their Case-Progressing Potential for Digital Forensics

## Full Text

This Markdown is an automated text extraction for search and reading tools. The [hosted PDF](SurveyEMSideChannelsForensics.pdf) is the authoritative publication source.

Source PDF: [SurveyEMSideChannelsForensics.pdf](SurveyEMSideChannelsForensics.pdf)
Source SHA-256: `c7d4d95c6379c656e1fd45f06ab842edf78aa85ea55098d3303e57691549c498`
Generated with `pdftotext -nopgbrk`; formatting and reading order may differ from the PDF.

---

Digital Investigation 29 (2019) 43e54

Contents lists available at ScienceDirect

Digital Investigation
journal homepage: www.elsevier.com/locate/diin

A survey of electromagnetic side-channel attacks and discussion on
their case-progressing potential for digital forensics
Asanka Sayakkara, Nhien-An Le-Khac, Mark Scanlon*
Forensics and Security Research Group, University College Dublin, Ireland

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 10 October 2018
Received in revised form
1 March 2019
Accepted 11 March 2019
Available online 15 March 2019

The increasing prevalence of Internet of Things (IoT) devices has made it inevitable that their pertinence
to digital forensic investigations will increase into the foreseeable future. These devices produced by
various vendors often posses limited standard interfaces for communication, such as USB ports or WiFi/
Bluetooth wireless interfaces. Meanwhile, with an increasing mainstream focus on the security and
privacy of user data, built-in encryption is becoming commonplace in consumer-level computing devices, and IoT devices are no exception. Under these circumstances, a significant challenge is presented to
digital forensic investigations where data from IoT devices needs to be analysed.
This work explores the electromagnetic (EM) side-channel analysis literature for the purpose of
assisting digital forensic investigations on IoT devices. EM side-channel analysis is a technique where
unintentional electromagnetic emissions are used for eavesdropping on the operations and data
handling of computing devices. The non-intrusive nature of EM side-channel approaches makes it a
viable option to assist digital forensic investigations as these attacks require, and must result in, no
modification to the target device. The literature on various EM side-channel analysis attack techniques
are discussed e selected on the basis of their applicability in IoT device investigation scenarios. The
insight gained from the background study is used to identify promising future applications of the
technique for digital forensic analysis on IoT devices e potentially progressing a wide variety of currently
hindered digital investigations.
© 2019 Elsevier Ltd. All rights reserved.

Keywords:
Electromagnetic side-channel attacks
Internet-of-things (IoT)
Digital forensics
Data encryption

Introduction
Digital forensics is the field where legal investigations are
assisted by analysing digital sources of evidence. In contrast,
cybersecurity is the domain where the concern is to ensure the
security of digital data and the privacy of their owners. In today's
modern world, technology is becoming increasingly prevalent in
everyday life and many people stay almost always connected to the
Internet (Nie and Erbring, 2000). While various social networks
facilitate their users to share their life events to the rest of the world
intentionally, every computer-based device they interact with in
everyday life leaves unintentional traces of their activities. Such
sources of forensic information include computer hard disks,
network activity logs, removable media, internal storage of mobile
phones and many others (Soltani and Seno, 2017).

* Corresponding author.
E-mail addresses: asanka.sayakkara@ucdconnect.ie (A. Sayakkara), an.lekhac@
ucd.ie (N.-A. Le-Khac), mark.scanlon@ucd.ie (M. Scanlon).
https://doi.org/10.1016/j.diin.2019.03.002
1742-2876/© 2019 Elsevier Ltd. All rights reserved.

Internet of Things (IoT) is an emerging trend started as a narrow
research domain called wireless sensor networks, which evolved
into Internet-connected everyday objects. IoT ecosystem includes a
wide variety of devices, such as smart-watches, smart TVs, CCTV
cameras, medical implants, fitness wearables, etc. The increasing
availability of IoT devices across society makes it inevitable to find
them in modern crime scenes and digital forensic investigations.
Most of these devices comes with limited data processing and
storage capabilities and they usually possess limited standard interfaces to the outside world, such as USB ports or WiFi/Bluetooth
wireless interfaces, unlike their PC counterparts (Stojkoska and
Trivodaliev, 2017).
Due to the increasing concerns regarding security and privacy
among communities, modern digital devices, such as computer
systems, mobile devices, etc., are designed and shipped with builtin security. Popular smartphones, such as iOS and Android based
devices, encrypt their internal storage in order to protect user data
from third parties (Ahmad et al., 2013). Each of the mainstream PC
operating systems, such as Mac OS, Windows, and Linux, provide
built-in
hard
disk
encryption.
Meanwhile,
network

44

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

communications, both wired and wireless, commonly employ
strong packet encryption mechanisms (van de Wiel et al., 2018).
Modern computer hardware has made the automated handling of
encrypted data an everyday possibility in consumer, industrial and
military applications (Fritzke, 2012). Computer devices seized at a
crime scene containing encrypted data poses a significant challenge
to the investigation (Lillis et al., 2016; Sayakkara et al., 2018a). The
IoT device ecosystem is no exception for this data encryption trend
making the challenge of digital forensic investigations on IoT devices even more complex.
Side-channel analysis attacks have been proven to be useful to
breach security on computer systems when standard interfaces,
e.g., network interfaces and data storage devices, are sufficiently
protected (Spreitzer et al., 2018; Dhem et al., 1998; Zhang et al.,
2014; O'Malley and Choo, 2014). In order for a side-channel
attack to be effective in practical scenarios for a security breach, it
has to be executable without having physical access to the device
being attacked (Wakabayashi et al., 2017). In the case of digital
investigation, the investigator has the freedom to handle the device, and ideally, any investigative activity must not affect or change
the digital information in the device (Du et al., 2017). Electromagnetic (EM) Side-channel Attacks is one approach that has shown
promising results. It requires minimum physical manipulations to
the device being inspected (Hayashi et al., 2013). EM emissions of a
device can be passively observed to infer both the internal operations being performed and the data being handled (Sayakkara et al.,
2018a). This condition is ideal for a digital investigator who attempts to ensure that the device does not go though any physical
changes due to its investigation. It is worth noting that hardware
manufacturers are continuously trying to circumvent EM sidechannel attack vulnerabilities through EM shielding and operation obfuscating enabled firmware.
This paper discusses the possibility for EM side-channel analysis
as a potential case-advancing possibility for digital forensic analysis
of IoT devices. A comprehensive analysis of the literature is provided identifying some promising avenues for research and their
future potential. EM side-channel attacks for the recovery of
cryptographic keys and other forms of important information are
evaluated for potentially overcoming the encryption problem in
digital forensics on IoT devices. Since the nature of EM emission
phenomena is associated with the power consumption of
computing devices (Callan et al., 2015a), the literature that focuses
on power analysis attacks are also discussed where appropriate.
The contribution of this work can be summarised as follows:

 A comprehensive literature review and a comparative study of
the research that has been carried out in EM side-channel
analysis is provided and recent advances are summarised.

 The scenarios where different EM side-channel attacks in the
literature are relevant and applicable in digital forensic investigations are identified.

 Light is shined on several new avenues of research that are
possible to achieve in digital forensic investigations and cybersecurity through the adoption of EM side-channel analysis
techniques.

 The shortage of reliable tools and frameworks available to utilise
EM side-channel analysis for digital forensic investigations on
IoT devices is identified and the recommendations are made to
overcome it.
The rest of this paper is organised as follows. Section 2 presents
an overview of side-channel attacks. Sections 3, 4, and 5 explores
approaches for acquisition, unique identification, and information
leakage EM emissions relevant to digital forensics. In Section 6, the
advancements in wireless communication technologies and

standardisation, and the legal background relevant to EM sidechannels are discussed. Section 7 provides insights of possible
future ethical directions of this technique. Finally, Section 8 concludes the paper.
Side-channel attacks
The topic of side-channel attacks spans a wide variety of techniques. Each side-channel attack on a computer system focuses on
one specific unintentional leakage of information from either
hardware or software (Spreitzer et al., 2018). Some of such information leaking side-channels are listed below.

 The memory and cache spaces shared between different
software.

 The amount of time a program takes to respond to different
inputs.

 The sounds different components of computer hardware make.

 The amount of electricity a computer system draws.

 The EM radiation a computer hardware emits.
Computer programs contain conditional branches and loops in
order to handle inputs and produce the intended output.
Depending on the input values, the execution path of a program can
differ, which may result in a different program execution time. It
has been shown that the execution time of encryption algorithms
can reveal information regarding the input values provided to it,
which includes the encryption key (Dhem et al., 1998). For example,
the square and multiplication segment in the RivestShamirAdleman (RSA) algorithm checks whether a key bit is 0 or 1 before
moving into multiplication operations. Therefore, the observation
of large number of execution times with the same key and different
input data can lead to uncovering the key bits effectively (Dhem
et al., 1998; Brumley and Boneh, 2005; Kocher, 1996).
In environments where multiple virtual machines (VMs) run on
the same hardware, such as cloud infrastructure, cache-based sidechannel attacks are possible (Zhang et al., 2014). While each VM has
its own virtual resources, many of them are mapped into shared
physical resources including shared cache memories. It has been
shown that an attacker running a VM on a virtualised environment
can spy on a victim VM through the shared cache storage. This can
lead to the extraction of sensitive information, including cryptographic keys (Liu et al., 2015).
It has been shown that acoustic emanations from various
components and peripherals of computer systems can be used to
exfiltrate information (O'Malley and Choo, 2014). Genkin et al.
showed that it is possible to distinguish between CPU operations by
listening to acoustic emanations resulting in an attack on the
cryptographic keys of the RSA algorithm (Genkin et al., 2014).
Computer displays and their video cables have also been identified as an eavesdroppable EM source, which can leak the image
being displayed on the display. Such leakages from CRT based
displays have been known for several decades (Van Eck, 1985).
Video information provided to a computer display has synchronisation information to recognise between different lines of pixels
and different frames, which are called horizontal and vertical
synchronisations. By recognising this synchronisation information
in the EM emissions, an attacker can reconstruct the images being
displayed (Hongxin et al., 2009; Elibol et al., 2012).
Kocher et al. were the first to introduce power consumption
based side-channel attacks; simple power analysis (SPA) and differential power analysis (DPA) (Kocher et al., 1999). SPA collects power
consumption variation (in mA) over time with a high sample rate,
such as twice the clock frequency of target cryptographic device.
The waveform of the power consumption, when plotted against

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

time, contained patterns that corresponded to the instructions of
the data encryption standard cryptographic algorithm (DES). If SPA
can reveal the sequence of operations, it follows that this sequence
depends on the data being handled by the algorithm (due to conditional branching). Designing code to minimise data dependent
branching, which does not show characteristic power consumption
patterns for specific operations, can prevent attackers from recognising what is executing on the device (Zankl et al., 2018).
DPA is a technique that can be custom tailored for specific
encryption algorithms. Kocher et al. used the DPA technique against
DES (Kocher et al., 1999). The technique was able to guess the
encryption key accurately, given sufficient cipher texts and power
traces for those encryption operations. The authors claim that they
have used DPA to reverse engineer various unknown algorithms
and protocols on devices. The authors state that it may be possible
to automate this reverse engineering process. Kocher et al. hints
that these techniques (SPA, DPA) might be useable with EM emissions too in addition to power consumption.
While various side-channel attacks are possible on computer
systems, it is possible to increase the advantages achievable by
combining multiple side-channels that leak different kinds of information together (Agrawal et al., 2003). For example, power
analysis and EM analysis can be performed together in order to
reduce the errors and improve the accuracy of inferring the leaked
information from a computer system.
Unintentional electromagnetic emissions
EM radiation is the underlying technology for numerous of
wireless communication. Meanwhile, it is a well documented fact
that electronic devices generate EM radiation on unintended frequencies as a side effect of their internal operations (Getz and
Moeckel). Such unintended EM radiation are regulated by government agencies, such as Federal Communications Commission (FCC) in
the USA, due to the possible interference they can make on legitimate wireless communication and the potential health issues they
can cause to the users of these devices. However, it is not possible to
entirely avoid such emissions. Equipment manufacturers attempt
to minimise it as much as possible (Ott, 2011). This section discusses how EM signals are generated from different components of
a computer system, what kind of information they may carry, and
what types of methods and tools can be used to capture these
signals.
Hardware that causes electromagnetic emissions
As derived from Maxwell's equations, EM waves can be generated by electric currents varying over time. Characteristics of the
EM waves being generated, such as frequency, amplitude, and
phase, depends on the nature of the time varying electric current
(Maxwell, 1865). Based on this principle, modern communication
systems generate oscillating currents on antennas that generate EM
waves that propagate over free space. They can be captured by
another antenna with appropriate properties. Modern digital
computer systems have a large number of components that depend
on electric pulses or alternating currents for their operations. That
leaves the space for EM waves to be generated at unexpected frequencies without the intention of the system manufacturer.
There are multiple computer components that operate in a coordinated, sequential fashion according to clock signals. Among
them, both the CPU and RAM are of particular interest. The CPU
performs a cycle of fetching instructions, decodes and executes
them, while RAM maintains the data and corresponding instructions when a computing device is powered on. EM emission
signals from these components contain a significant amount of

45

side-channel information regarding the events related to software
execution and data handling. On most IoT devices, the CPU and
RAM are included in microcontroller (MCU) chips making it the
most important EM source on-board.
Sampling electromagnetic emissions
The EM emission frequencies of a target device is unpredictable
due to its dependability on various hardware characteristics.
Therefore, it is difficult to have a universal purpose device that can
be used to observe EM emissions from a target device and interpret
side-channel information. It has been shown that small magnetic
loop antennas can be used for the purpose of detecting EM emissions from computing devices (Peeters et al., 2007). When EM
signals are captured by a loop antenna, it requires digital sampling
before the data can be used for analysis. Theoretically, the sample
rate of the equipment should be twice that of the maximum EM
frequency required to be captured e referred to as Nyquist frequency (Smith, 1997). For this reason, EM signal sampling equipment must have a very high sample rate. The most commonly used
equipment to capture EM signals are oscilloscopes and spectrum
analysers with high sample rates. The digitised data these devices
capture can be subsequently analysed in signal analysis software.
However, access to such devices for information security professionals is not very common (Wolfe, 2003).
Software defined radios (SDRs) are getting increasingly popular
among wireless hackers, hobbyists, and security enthusiasts who
are interested in access to the radio frequency (RF) spectrum. An
SDR consists of a minimal hardware component, which can be
tuned to a range of RF frequencies and then digitise it with a fast
analogue-to-digital converter (ADC). The processing of digitised RF
data is handled entirely on software (Tuttlebee, 2003). A wide variety of SDR hardware and software platforms are available (Cass,
2013; Ossmann; Ettus and Braun, 2015; Blossom, 2004). Due to
the enhanced flexibility provided by software, SDR platforms have
become a perfect candidate for EM side-channel attack analysis
research. A SDR can be used to scan through a wide range of frequencies to locate potential EM emissions from a computer system.
A simple set-up with an SDR platform can be used to demonstrate the unintentional EM signals. An Arduino Leonardo prototyping board is loaded with a simple program to blink an LED
connected to it via the general purpose I/O pins. An antenna connected to an RTL-SDR dongle is placed close to the Arduino board in
order to receive unintentional EM signals emitted from the board.
The Arduino board consists of a microcontroller chip that operates
at 16 MHz. However, the RTL-SDR dongle cannot be tuned to frequencies below 22 MHz. Therefore, a GNURadio script was programmed to tune the RTL-SDR dongle to the first harmonic of the
Arduino clock, i.e., 32 MHz. Fig. 1 illustrates the spectrograms of
three different EM signal samples gathered.
When two different LED blinking patterns are performed by two
different programs separately, the Arduino emitted two completely
different EM signal patterns. The spectrogram shown in Fig. 1 (a)
illustrates the EM signal observed from a simple LED blinking
program, while the spectrogram in Fig. 1 (c) illustrates the EM
signal of a more complex LED blinking program. The spectrogram
depicted in Fig. 1 (b) shows the EM signal observed during the reprogramming stage of the Arduino device from the first program
to the second.
Connection between CPU instructions and electromagnetic
emissions
As a result of executing instructions in different combinations by
the CPU, EM signal patterns are emitted at various frequencies and

46

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

Therefore, it is clear that the target device's system clock is the main
source of EM radiation. The design of the printed circuit board
(PCB), and characteristics of the electronic components provide
variations to this strong signal. Meanwhile, subfigures (a), (b), and
(c) of Fig. 1 clearly show that the instruction sequence, i.e., the
program being executed on the CPU, has a significant influence to
the EM emission pattern.
Electromagnetic emissions as a hardware signature

Fig. 1. Spectrograms of AM demodulated EM emissions acquired from an Arduino
device where (a) running a simple LED blink program, (b) reprogramming the device,
and (c) running a complex LED blink program are depicted.

amplitudes. Depending on the sequence of instructions, i.e., the
exact program being executed, the output of EM noise from the CPU
varies significantly. Due to this, systematically modelling and predicting possible EM signal characteristics of a computer processor is
a difficult task. In order to identify unintentional EM emissions of a
computer processor, the most practical method is scanning a large
frequency spectrum for suspected EM signals and subsequently
trying to interpret these identified signals for potential sidechannel information. This arduous approach is a time consuming
task that requires manual inspection by a human user.
In 2014, Callan et al. introduced a metric called SAVAT (Signal
AVailability for an ATtacker) that measures the EM signal power
emitted when a CPU is executing a specific pair of instructions (A
and B). The authors show that different selections of A/B instruction
pairs emit different SAVAT values, i.e., signal power (Callan et al.,
2014; Zajic and Prvulovic, 2014). An improvement to the SAVAT
technique is a method called Finding Amplitude-modulated Sidechannel Emanations (FASE). The key idea behind the FASE technique
relies on the phenomena that when a program activity is alternating at a frequency (falt ) that affects any periodic EM signal
originating from any source at a frequency fc, it is possible to
observe two side-band signals at fc  falt and fc þ falt between the fc
signal. Further improvements to SAVAT technique enabled the
possibility of identifying both amplitude and frequency modulated
EM emissions from CPUs (Callan et al., 2015b; Prvulovic et al., 2017;
Yilmaz et al., 2018). While it is evident from existing studies that
EM side-channel leakage is available across various type of CPUs,
further studies are necessary to identify the effect of different CPU
architectures to the produced EM emissions.
Electromagnetic emissions as a signature
When a computing device running a program generates EM
emissions, the patterns observable depend on the precise settings
of the device. In the EM emission spectrum of the Arduino device in
Fig. 1, it is clearly evident that both the hardware and software
settings have influenced the EM emission patterns. The signal
captured at the first harmonic frequency of the Arduino device's
system clock, i.e., 32 MHz, is showing a varying patterns in the
spectrogram view according to the changes made to the device.

Despite the software components available on a computing
device, it is important to investigate whether the hardware alone
can provide a recognisable EM emission pattern. Such a capability
can lead to profiling of hardware devices and components uniquely.
It has been shown that a simple EM signal acquisition device called
RTL-SDR, which demonstrated the capability to capture EM emissions from an Arduino device, can be used to profile computing
devices uniquely. Laput et al. used a similar device to acquire EM
signals that were successfully applied to a support vector machine
(SVM) classifier to uniquely distinguish the EM source device
(Laput et al., 2015). This possibility has led to the idea that EM
emissions from an electronic device owned by a person can be used
as an authentication token of the person instead of relying on
conventional methods, such as Radio Frequency Identification
(RFID) tags (Bianchi and Oakley, 2016; Yang and Sample, 2016).
This uniquely distinguishable EM emission patterns of a known
electronic device can help to identify any potential alteration that
may have applied to it. For example, a known electronic device can
be altered at the hardware fabrication level for malicious purposes,
such as accessing stored data or eavesdropping on users’ activities.
Such hardware modifications result in a changed EM emission
pattern that can be used to identify it (Yang et al., 2017). Similarly, a
genuine electronic device can be replaced by a counterfeit electronic device for a malicious objective. It has been shown that even
when counterfeit hardware attempts to follow the design of the
genuine device, it still creates distinguishably different EM emission patterns compared to their original product (Ahmed et al).
Electromagnetic emissions as a software signature
When software runs on different computing devices, it is clear
that the hardware EM emissions are influenced by the software
instructions being executed. It is important to consider this influence from two different aspects. The first aspect is how uniquely
the EM emissions of different software running on the same
hardware platform can be recognised. This can be used to pin point
to the exact software running on a device. The second aspect is how
unique the same software program is when it is running across
various hardware platforms. It facilitates the unique detection of a
specific piece of software.
When software systems are being developed, requirements
arise to debug their behaviour or find ways to increase their performance. Instrumenting software by applying logging events and
break points are the most common ways to identify where complex
software is not performing as expected. These developmental options affect the performance of the software being inspected in
addition to the overhead of their placement each time a copy of the
software needs to be inspected. It has been shown that unintended
EM emissions of the CPU can be used to inspect software execution
sequences without having to instrument the software (Callan et al.,
2016; Callan, 2016; Han et al., 2017; Espitau et al., 2017). Even when
the same program is running on different devices, the ability to
identify the instruction execution sequence can help to uniquely
identify the software itself.
The capability to detect software code execution sequence has

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

opened up the opportunity to identify when a computing device is
running software code not intended by the manufacturer or the
owner. One possible scenario can be software bugs or hardware
faults that cause an IoT device to execute unexpected instruction
sequences. Another possible scenario can occur when an IoT device
is under an attack causing it to run malware or an unintended part
of the device's genuine software. Stone et al. (Stone and Stone,
2015; Stone et al., 2015) and Nazari et al. (2017) showed that
such abnormal deviations of software code executions on
computing devices can be detected using the corresponding EM
emission patterns from its execution.
Even though it has been identified that EM emission patterns
are associated with both the hardware and software characteristics
of the source device, the format of a captured EM signal used for
this identification can vary. Instead of directly using time-domain
EM signal traces, one such alternative format is RF-DNA fingerprinting. This is a technique to fingerprint the physical layer of RF
transmitting devices, which includes WiFi, Bluetooth, Zigbee, GSM
devices, and even RADAR antennas. This technique has been used to
identify rogue devices in a deployment using their RF signals
without physically inspecting them (Reising, 2012; Dubendorfer,
2013; Danev et al., 2012; Lukacs et al., 2015). Deppensmith et al.
showed that RF-DNA technique can be applied to unintentional EM
emission fingerprinting on computing devices reliably
(Deppensmith and Stone, 2014). However, the evaluations performed by Stone et al. on microcontroller based IoT devices indicates that further studies are necessary to conclude the most
reliable format to represent unintentional EM signals (Stone and
Stone, 2016).
Fig. 2 illustrates the structure of an RF-DNA fingerprint. When
calculating it, the EM signals emitted from a device on a selected
frequency is captured, filtered, and amplified appropriately to
achieve a clean trace. From this acquired time-domain EM trace, the
three signal characteristics; Amplitude, Phase and Frequency, are
separately considered for further processing. Each signal characteristic is broken into N equally sized regions and then for each
region, four statistical metrics; standard deviation, variance, skewness, and kurtosis is calculated. The standard deviation and variance
metrics measure the spread of data samples, while the skewness
and kurtosis metrics measure the symmetry and the sharpness of a
data sample. Furthermore, the signal itself is again considered as a
one entire region, i.e., the ðN þ 1Þth region, to calculate the same
statistical metrics. As Fig. 2 illustrates, each of these calculated

47

statistical metrics are arranged in a single vector, which becomes
the RF-DNA fingerprint of the originally acquired EM trace from the
device.
Information leaking electromagnetic emissions
This section dives into the question of what information is
contained in an EM emission trace of a particular computing system. From a digital forensic perspective, both the kind of software
running on IoT devices and the data being handled by each software
application are potentially of significant interest. Even if EM sidechannel analysis cannot reveal all of data being handled by an IoT
device platform, extracting critical information, e.g., cryptographic
keys, can help progress forensic analysis.
Observable electromagnetic spectrum patterns
While there exists a wide variety of microcontroller chips used
on IoT devices, Sohaib et al. has shown that it is still viable to
perform EM side-channel attacks on them (Sohaib ul Hassan).
When considering information leakage from an EM emission trace,
visually inspecting the time-domain signal is the first observational
technique. This approach is called simple electromagnetic analysis
(SEMA), which evolved from the simple power analysis (SPA)
introduced by Kocher et al. (1999). Another way of performing visual observations is by transforming the EM trace into the frequency domain and plotting it as a spectrogram. This enables
observation of different signal patterns distributed over multiple
frequencies.
Multiple published works have demonstrated the effectiveness
of the SEMA approach in extracting critical data from computers,
including cryptographic keys. The El Gamal and RSA algorithms
implemented using GnuPG library were attacked by observing
critical CPU operations (Genkin et al., 2015a). Furthermore, elliptic
curve based cryptographic algorithms (ECC), such as Elliptic Curve
based Diffie Hellman (ECDH) and Elliptic Curve based Digital
Signature Algorithm (ECDSA), are identified to be vulnerable to EM
side-channel attacks with the SEMA approach (Goubin, 2003;
Genkin et al., 2016a, 2016b). Due to the low computational overhead in ECC algorithms, many mobile devices and IoT platforms
tend to employ ECC algorithms to secure data. This indicates that
such devices can be inspected through EM side-channels to access
cryptographically protected data.
Differential electromagnetic analysis (DEMA)

Fig. 2. The RF-DNA fingerprinting process.

If a large number of EM traces from a computing system can be
observed executing specific software, it is possible to identify the
data bits involved in the operations that appear across the large
number of EM traces. While being impractical to perform under
real-world attack scenarios due to the challenge of collecting such a
large number of EM traces from a target computer, this may be the
only resort for EM side-channel analysis when it is not possible to
extract information from visual observation of a single EM trace.
DEMA, a variant of Differential Power Analysis (DPA), uses the
variation of EM emissions of a CPU to discover variables used in an
executing program, such as encryption algorithms (Kocher et al.,
1999, 2011). When a bit in a CPU register is flipped from 0 to 1 or
vice versa, it consumes an amount of energy, which is reflected in
the corresponding EM emission. Some CPUs may emit a higher EM
signal when switching a register bit from 0 to 1 than vice versa
since that operation can lead to a higher energy consumption
(Peeters et al., 2007). Due to this, when the contents of a complete
CPU register are modified, it is possible to identify the hamming
distance between the previous and new state using the resulting

48

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

EM emission. Since every instruction running on a CPU affects the
values on different registers, this means that attackers can identify
instructions being executed and intermediate variables used based
on the EM observation.
Fig. 3 illustrates how differential EM analysis is used to identify
the key of a simple XOR-cipher. Initially, a large set of input data
bytes and all possible key bytes are used to perform XORoperations and the hamming distances of the resulting values are
stored in a matrix called Hypothetical Matrix, as shown in Fig. 3. The
objective is to find the hypothetical key of the matrix that generates
matching hamming distances for the same input values. For this,
the input values are fed to software running on the CPU, where the
XOR operation is performed with the unknown key. EM emissions
are sampled for each input value to generate the Real Matrix, where
each column contains an EM signal sample for its corresponding
input. Calculating cross-correlations between rows of the two
matrices can find the corresponding row in the Hypothetical Matrix
that provides the best correlation of hamming distances. The hypothetical key corresponding to this row is likely the encryption
key used for the XOR operation inside the CPU.
Standard encryption algorithms, such as the Data Encryption
Standard (DES) and the Advanced Encryption Standard (AES),
employ XOR operations at various stages in their functionality using chunks of the encryption key and input data. Therefore, DEMA
attacks are possible by attacking each chunk of the key being used
with the XOR operations. Such an attack reveals parts of the
encryption key, which have to be combined at the end. However, in
a real-world setting, the attacker may not have enough EM emission samples of the encryption operations to calculate the correct
part of the key used. This results in lists of possible key chunks for
each segment of the encryption key used in the algorithm. The
problem of identifying the correct parts of the key to build the
complete encryption key is called the Key Enumeration Problem,
which can be solved within a reasonable computational overhead
(Bogdanov et al., 2015).
Quisquater et al. practically demonstrated that EM analysis is a
viable option to the aforementioned power analysis attack on
computer CPUs (Quisquater and Samyde, 2001). By precisely
moving the EM probe over a microcontroller, the authors were able
to build an accurate 3-dimensional EM signature of the chip

running an idle loop. It was shown that the radiation spectrum of
each processor was sufficiently unique to use as a distinguishable
feature for processor identification. These experiments were performed in a Faraday cage to minimise the external noise effects and
the EM emissions were captured using a small magnetic loop antenna (diameter z 3 mm). An oscilloscope digitised the signal for
analysis. Gandolfi et al. (2001) applied DEMA to extract encryption
keys from three different chips used on smartcards namely;
COMP128, DES, and RSA. According to this study, the SNR of EM
emissions from these chips is higher than the SNR of power consumption analysis. This results in the extraction of more information from the DEMA technique as compared to simple power
analysis attacks (Messerges et al., 1999). Asymmetric key encryption, such as RSA, can also be attacked by identifying the individual
modular exponentiation operations performed within the algorithm through EM emissions (Witteman et al., 2011).
Due to the modern electronic multimedia distribution model,
e.g., music, movies, and ebooks, end-users can often become the
attackers. These users might have the malicious intention of
breaking encryption or sharing Digital Rights Management (DRM)
protected data. Since the victim device is owned by the attacker,
unlimited physical access to the hardware and software becomes
available for the attacker through various side-channel attacks.
White Box Cryptography (WBC) was introduced as a solution to
this; whereby cryptographic algorithms and keys are combined
with random codes and random data to create an obfuscation that
makes side-channel attacks more difficult. However, it has been
shown that EM side-channel attacks, such as DEMA, are still
capable of extracting encryption keys despite of the application of
WBC techniques (Sanfelix et al., 2015).
Recently, Camurati et al. made an important discovery that
extended the previously known capabilities of EM side-channel
analysis of cryptographic operations on IoT devices (Camurati
et al., 2018). It was shown that mixed-signal processors, such as
system-on-chips (SoCs), that contains a radio transceiver and a CPU
on the same silicon die, can cause long distance EM leakages. This
occurs when the CPU noise gets modulated into the radio transceiver's emission e extending the range of the CPU EM sidechannel. As the usage of SoCs is getting increasingly popular on
IoT devices, this latest type of EM side-channel leakage, called
screaming channels, has significantly increased the potential attack
surface.
Analysis on wireless-powered devices

Fig. 3. EM analysis of XOR-Cipher algorithm to extract the encryption key.

Unlike traditional computing devices that have their own power
source to run CPU operations, wireless-powered devices, e.g., passive RFIDs, depend on an external RF field provided by the device's
reader for power (Calari and Lampkin, 1997). IoT devices are ideal
candidates to be powered by wireless means. EM side-channel
analysis on such devices is challenging due to the presence of a
strong RF field from the reader, which obfuscates the weak EM
emissions of the devices themselves. However, RFID based devices
are being used in critical systems, such as secure access control to
buildings and electronic payments, where cryptographic operations are performed on-board. Therefore, investigating the EM sidechannel capability on such devices is important from both security
and forensic standpoints.
Hutter et al. demonstrated the capability to perform EM sidechannel analysis on RFID based devices using a custom-made
RFID tag as a proof of concept (Hutter et al., 2007). Using this
custom set up, the authors were able to recover the AES key used in
a challenge response protocol between the RFID tag and the reader.
In order to avoid the disturbance from the RF field of the reader
device, the RFID circuitry was placed outside the reader's RF field,

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

while the power harvesting antenna is kept inside the reader's RF
field. The two components were connected through a sufficiently
long wire. This enabled the measurement of the EM emissions from
the RFID circuitry without interference. However, it is not possible
to follow a similar approach in a regular RFID tag as the antenna and
RFID circuitry are inseparable by any reasonable means.
Kasper et al. performed EM side-channel attacks on RFID based
smart-cards in a more realistic setting. This research employed
commercially available smart-cards and performed the attacks
within the RF field of the RFID reader (Kasper et al., 2009). When
the RFID smart-card is consuming more energy, the amplitude of
the RFID readers RF field becomes lower. Similarly, when the RFID
smart-card is consuming less energy, the amplitude of the RFID
readers field is higher. This means, the power consumption of the
RFID tag is reflected in the amplitude of the RFID readers carrier
frequency. Kasper et al. used this signal as the EM side-channel to
attack internal operations of the RFID reader. A computercontrolled USB-oscilloscope and a computer-connected custommade RFID reader was used to attack the RFID tag while capturing
EM fluctuations using a small RF loop probe. This set up was used to
perform a correlation power analysis (CPA) attack to extract the
symmetric keys used in DES and 3-DES implementations on the
smart-card successfully.
Recent research by Xu et al. demonstrated that RFID based
smart-cards that employ side-channel attack mitigation techniques, such as head and tail protection, are not effective enough
against EM side-channel analysis attacks (Xu et al., 2018). In their
work, encryption keys used for the 3DES algorithm were demonstrated to be recoverable. In light of this attack vector, it is important to note that wireless-powered IoT devices are also susceptible
to threats from EM analysis based attacks.
Many smart-card-based fraud, such as stealing credit card details, involve malicious devices, such as card skimmers, that can read
and store data from the cards. Investigators face the challenge of
identifying victims of such skimming devices due to the fact that
card details are encrypted when stored on these devices. Souvignet
and Frinken demonstrated that correlation power analysis, which is
a variant of DPA attack, can be used to extract the details of victim
smart-cards from such skimmer devices by physically tapping into
a seized device (Souvignet and Frinken, 2013). The success of their
work indicates that EM side-channel analysis can be even more
promising in extracting such evidence without requiring any
physical alterations to the device itself.
Countermeasures to electromagnetic side-channels
As EM side-channel analysis has been shown to be successful on
recovering data from computing devices, various countermeasures
have been explored to counteract it on both software and hardware
levels (Zankl et al., 2018). Masking variables by using random
values alongside the operations is a basic software-based countermeasure that has been proven to not be effective enough against
EM side-channel attacks (Kim et al., 2008; Chari et al., 2002).
Various other approaches including the randomisation of the
operation sequences or lookup tables of algorithms (Saputra et al.,
2003; Kim et al., 2016), avoiding instructions pairs executing
adjacently that are known to emit distinguishable EM patterns
(Callan et al., 2014; Zajic and Prvulovic, 2014), and accessing critical
data using pointers instead of values (Witteman and Oostdijk,
2008) require further studies to see how effective they are
against EM side-channel attacks.
Quisquater et al. suggests several hardware design countermeasures to these attacks (Quisquater and Samyde, 2001). Actions
that can be taken by hardware designers includes minimising metal
parts in a chip to reduce EM emissions, the use of Faraday cage like

49

packaging, making the chip less power consuming (which leads to
less unintentional emissions), asynchronism (i.e., design the chip
not to use a central system clock and instead operate asynchronously), and the use of dual line logic (i.e., using two lines that in
combination of two bits represents a state instead of a single line
that simply represent 0 or 1 states). Furthermore, it has been shown
that it is possible to mathematically model an electronic chip
during the design phase to identify and avoid potential information
leakages through EM side-channels (Ishai et al., 2003; Standaert
et al., 2009).
Standards and tools
EM side-channel attacks are not currently commonly being used
for digital forensics purposes. Therefore, it can be too early to find
any existing standards or tools on EM side-channel analysis for
digital forensics. However, in order for future establishment of
standards and tools, it is important to review the relevant standards
and tools in both hardware and software security domains.
The concerns of electromagnetic wave emissions from IoT devices from the software perspective are mostly concentrated towards the wireless communication technologies, such as WiFi,
Bluetooth, and proprietary IoT protocols, e.g., Zigbee (Golmie et al.,
2003; Iyer et al., 2015). Meanwhile, unintentional EM emission
minimisation is generally left to those involved in the hardware
design and manufacturing process. The term electromagnetic
compatibility (EMC) refers to a device's unintentional EM emissions
that can affect the functionality of other devices and the health of
humans who are exposed to it (Getz and Moeckel). The Federal
Communications Commission (FCC), the Food and Drug Administration (FDA), the International Electrotechnical Commission (IEC),
and the European Union (EU) are examples of authorities concerned
with EMC (Barron, 2007; Hayashi, 2016). However, regarding the
question of EM side-channel information leakage from general
purpose electronic devices, there are no such rules to govern the
manufacturers. Instead, only guidelines exist, which may or may not
be followed (ISO/IEC 17825:2016; ISO/IEC TS 30104:2015).
Once a hardware device's design is completed and
manufacturing commences, it is a challenging task to apply mitigation steps should the EMC tests reveal that it does not meet requirements. In the worst case scenario, the minimisation of EM
emissions may require a complete reworking of the PCB used in the
device or a replacement of a critical electronic component. Due to
the potential for costly manufacturing disruption, the minimisation
of EM emissions is necessary from the initial hardware design
phase. Due to the fact that EM side-channel information leakage is
not a problem limited to hardware manufacturers, a joint effort by
both hardware and software developers to establish standards is
necessary.
In order to facilitate the assessment on EM side-channel security
threats, standardised tools and frameworks are highly necessary.
Test vector leakage assessment (TVLA) is a technique that can be used
to assess the resistance of cryptographic implementations against
hardware side-channel attacks (Becker et al., 2013). TempestSDR is
a software tool that can be used with a large variety of hardware
platforms, e.g., the universal software radio peripheral1 (USRP) or
HackRF,2 to eavesdrop on computer monitors by capturing the EM
signals emitted by the video cables (Marinov, 2018; Sayakkara et al.,
2018b). Multiple commercial and open source products exist that
can be used to break encryption on microcontroller based IoT devices, such as ChipWhisperer (ChipWhisperer embedded ha, 2018;

1
2

https://www.ettus.com/.
https://greatscottgadgets.com/hackrf/.

50

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

O'Flynn and Chen, 2014) and Riscure Inspector (Riscure, 2018;
Ramsay and Lohuis). Blanco et al. presented a side-channel trace
acquisition framework called SCAP, which is targeted at general
purpose computing devices (including mobile devices). While the
framework does not currently perform side-channel attacks, the
objective is to provide a platform to build future analysis tools
(Blanco et al., 2017). Such tools enable IoT system developers to test
the robustness of their hardware against physical side-channel
attacks and identify information leakage.
Discussion
Having discussed the scientific literature related to EM sidechannel analysis attacks, it is important to identify the future
impact it may cause in the domain of digital forensics on IoT devices. This section highlights some of the potential ways this impact
may occur in the future under different themes. Fig. 4 illustrates the
avenues for future research in this direction. Many of these future
potentials are already starting to be realised and others are ambitious predictions that can prove significantly beneficial to digital
forensics.
Recent versions of mobile operating systems, e.g., Android and
iOS, secure their internal storage using encryption. Critical information necessary for digital forensic investigation can be inaccessible due to being stored in an encrypted form (Casey and Stellatos,
2008; Zdziarski, 2008; Hoog, 2011). This includes encrypted emails,
encrypted instant messenger applications, encrypted files, and
encrypted storage partitions. While the increasing application of
cryptographic protection on computing devices poses a challenge

to traditional digital forensics, it can open up new opportunities to
EM side-channel attacks (Lillis et al., 2016). EM side-channel attacks
require a large number of traces acquired from a victim device
while the device is performing cryptographic operations using a
single key. The most common encryption operations occurring on
older systems are the secure socket layer (SSL) based web traffic
(Fahl et al., 2013). With encrypted data storage becoming
commonplace, EM side-channel attacks can potentially be performed by observing cryptographic operations during live data
forensic analysis procedures (Hay et al).
Instead of using a single side-channel attack in isolation, combinations of multiple side-channel attacks directed towards a single
computer system can prove more fruitful. It has been proven that
power and EM side-channel analysis can be combined to achieve
better results (Agrawal et al., 2003). There can be some operations
of the CPU that are more clearly reflected in the device's power
consumption than in the EM emission and vice versa. Similarly,
malware running on a victim computer can aid an EM side-channel
attacker to extract additional information (over the EM sidechannel alone) by intentionally modulating data into the EM
emission of the CPU or the monitor (Cheddad et al., 2010; Yang and
Sample, 2017; Sayakkara et al., 2018b).
The unintentional EM emissions from computing devices can
cause interference to other radio signals in the vicinity. This phenomena is evident in laptop computers, which have been shown to
modulate signals from commercial AM radio stations (Entriken,
2018). IoT devices already use this interference phenomena to
communicate purposefully with other devices by modulating the
ambient RF signals. This is called backscatter communication

Fig. 4. Altering the traditional IoT digital evidence acquisition process (Du et al., 2017), EM side-channel analysis can help in various ways for live analysis of computing devices.

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

technology (Liu et al., 2017). There are various carrier wave sources
that have been tested in the literature including TV transmission
stations and WiFi access points (Liu et al., 2013; Kellogg et al., 2014;
Bharadia et al., 2015; Zhang et al., 2016). The potential of using this
backscatter phenomena to eavesdrop on internal CPU operations of
IoT devices by listening to ambient RF sources warrants further
exploration.
Devices deployed in wired networks, such as routers and
switches, are known EM noise sources. It has been already showed
that MAC addresses in Ethernet frames can be extracted by performing SEMA analysis on the EM emissions from wired routers
(Schulz et al., 2016). When it is required to perform an investigation
on a live wired network, it is necessary to be connected in order to
inspect packets (Corey et al., 2002). In situations like this, the EM
emissions of routers and switches might be able to provide an
approximate picture of the workload and traffic on the network
(Goudos et al., 2008).
Recent advances that have been made in the area of artificial
intelligence (AI) (incorporating machine learning (ML) and deep
learning (DL)) have demonstrated promising applications to many
other domains across computer science. Various tasks where human intuition was required to perform decision making are now
being replaced with ML/DL powered algorithms. Software libraries
and frameworks are becoming increasingly available in order to
assist the building of applications that have intelligent capabilities.
Examples include the automated detection of malicious programs
(Kolter and Maloof, 2004), image manipulation (Saboia et al., 2011),
and anomaly detection in network traces (Mukkamala and Sung,
2003).
EM side-channel analysis techniques that previously required
human intervention can be automated through the development of
AI algorithms. Recent work by Wang et al. (2018) applied deep
learning algorithms including multi-layer perceptron (MLP) and
long short-term memory (LSTM) to detect anomalies in the code of
simple IoT devices, e.g., Arduino and Raspberry Pi, through the
power consumption side-channel. Therefore, it is potentially
possible to extract better information from EM traces than the
current manual observations are capable of achieving. Several examples that were discussed in previous sections already leverages
AI techniques to recognise EM trace patterns, which strongly hints
the future role that can be played by AI algorithms in EM sidechannel analysis for digital forensics (Laput et al., 2015; Lerman

51

et al., 2011; Callan et al., 2016; Callan, 2016; Nazari et al., 2017;
Stone and Stone, 2016).
Conclusions
Traditionally, digital forensics focuses on analysing traces left
behind by suspects on digital devices by inspecting file storage, log
files, network traces, etc. Live data forensics can also be performed
on systems that require more sophisticated investigative techniques and skills. As computing systems transform from less privacy and security concerned platforms into hardened platforms
that are designed with security in mind from their inception, the
typical work conducted by digital forensic investigators must
change accordingly. Cryptographically protected storage systems is
one of the largest challenges hindering efficient digital forensic
analysis. EM side-channel analysis has been demonstrated as a
potential door-opener for cryptographically protected data storage
and communications from a security perspective, which can be
built upon and adopted for digital forensic purposes.
This paper comprehensively analysed the literature on EM sidechannel attacks with the goal of applying the technique to assist
digital forensic investigations on IoT devices. While various mitigation techniques have been suggested and applied to counter
against EM side-channel attacks, existing literature demonstrates
that such attempts have not been successful in reducing the prevalence of this attack vector. EM side-channel analysis is still in its
infancy for digital forensic applications, which demands courtadmissible, forensically-sound processing when used not only for
obtaining security keys but also for the detection of unintentional
data leakage. However, this technique has significant potential to
have a substantial impact on the field and enable the progression of
otherwise stalled investigative cases involving both IoT devices,
and encrypted computing devices in general.
Conflicts of interest
None.
APPENDIX
A categorisation of the literature of EM side-channel attacks and
related areas are listed in Table 1.

Table 1
Categorisation of the literature on EM side-channels.
Objective

Technique

References

Acquisition of EM
Emissions

SAVAT
FASE
SDR Signal Acquisition
General
Time/Frequency/Hilbert
Transform Domains

(Callan et al., 2014; Zajic and Prvulovic, 2014; Prvulovic et al., 2017)
(Callan et al., 2015b; Prvulovic et al., 2017)
(Tuttlebee, 2003; Cass, 2013; Ossmann; Ettus and Braun, 2015; Blossom, 2004)
(Getz and Moeckel; Ott, 2011; Jabbar and Rahman, 1991; Peeters et al., 2007; Sohaib ul Hassan)
(Yang and Sample, 2016; Stagner, 2013; Ahmed et al; Stone and Stone, 2015; Stone et al., 2015; Callan et al.,
2016; Callan, 2016)
(Nazari et al., 2017; Clark et al., 2013)
(Laput et al., 2015; Reising, 2012; Dubendorfer, 2013; Lukacs et al., 2015; Deppensmith and Stone, 2014)
(Bianchi and Oakley, 2016; Yang et al., 2017; Stone and Stone, 2016)
(Hayashi et al., 2013; Van Eck, 1985; Elibol et al., 2012; Kocher et al., 1999; Agrawal et al., 2003; Genkin et al.,
2015a, 2015b, 2016a)
(Genkin et al., 2016b; Belgarric et al., 2016)
(Kocher et al., 1999, 2011; Peeters et al., 2007; Bogdanov et al., 2015; Quisquater and Samyde, 2001; Gandolfi
et al., 2001; Messerges et al., 1999; Witteman et al., 2011)
(Sanfelix et al., 2015; Hutter et al., 2007; Kasper et al., 2009; Chari et al., 2002; Camurati et al., 2018)
(Fritzke, 2012; Quisquater and Samyde, 2001; Ishai et al., 2003; Standaert et al., 2009)
(Callan et al., 2014; Zajic and Prvulovic, 2014; Kim et al., 2008, 2016; Chari et al., 2002; Saputra et al., 2003;
Witteman and Oostdijk, 2008; Standaert et al., 2009)
(Cass, 2013; Ossmann; Ettus and Braun, 2015; Blossom, 2004; Marinov, 2018; ChipWhisperer embedded ha,
2018; O'Flynn and Chen, 2014; Riscure, 2018)
(Ramsay and Lohuis; Blanco et al., 2017)

EM as a Signature

Cryptographic Key
Extraction

RF-DNA
General
SEMA/Spectrum Observation

DEMA

Countermeasures for
EM Side-Channels

Hardware Countermeasures
Software Countermeasures

Tools and Frameworks

Various

52

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

References
Agrawal, D., Rao, J.R., Rohatgi, P., 2003. Multi-channel attacks. In: International
Workshop on Cryptographic Hardware and Embedded Systems (CHES).
Springer, pp. 2e16.
Ahmad, M.S., Musa, N.E., Nadarajah, R., Hassan, R., Othman, N.E., 2013. Comparison
between android and ios operating system in terms of security. In: 8th International Conference on Information Technology in Asia (CITA). IEEE, pp. 1e4.
Ahmed, M.M., Hely, D., Barbot, N., Siragusa, R., Perret, E., Bernier, M., Garet, F., 2017.
Radiated electromagnetic emission for integrated circuit authentication, IEEE
Microw. Wirel. Compon. Lett. 27 (11), 1028e1030.
Barron, M.R., 2007. Creating consumer confidence or confusion? the role of product
certification in the market today. Marquette Intellect. Prop. Law Rev. 11 (2), 413.
Becker, G., Cooper, J., DeMulder, E., Goodwill, G., Jaffe, J., Kenworthy, G.,
Kouzminov, T., Leiserson, A., Marson, M., Rohatgi, P., et al., 2013. Test vector
leakage assessment (tvla) methodology in practice. In: International Cryptographic Module Conference, vol. 1001, p. 13.
Belgarric, P., Fouque, P.-A., Macario-Rat, G., Tibouchi, M., 2016. side-Channel analysis
of weierstrass and koblitz curve ECDSA on android smartphones. In: Cryptographers Track at the RSA Conference. Springer, pp. 236e252.
Bharadia, D., Joshi, K.R., Kotaru, M., Katti, S., 2015. BackFi: high throughput WiFi
backscatter. Comput. Commun. Rev. 45 (4), 283e296.
Bianchi, A., Oakley, I., 2016. Wearable authentication: trends and opportunities. IT
Inf. Technol. 58 (5), 255e262.
Blanco, A.B., de Fuentes, J., Manzano, L.G., Encinas, L.H., Munoz, A.M., Oliva, J.R.,
Garcıa, I.S., 2017. A framework for acquiring and analyzing traces from cryptographic devices. In: 13th EAI International Conference on Security and Privacy
in Communication Networks (SecureComm).
Blossom, E., 2004. GNU radio: tools for exploring the radio frequency spectrum.
Linux J. (122), 4, 2004.
Bogdanov, A., Kizhvatov, I., Manzoor, K., Tischhauser, E., Witteman, M., 2015. Fast
and memory-efficient key recovery in side-channel attacks. In: International
Conference on Selected Areas in Cryptography. Springer, pp. 310e327.
Brumley, D., Boneh, D., 2005. Remote timing attacks are practical. Comput.
Network. 48 (5), 701e716.
Calari U., Lampkin M.C., RFID reader, US Patent 5,621,199 (Apr. 15 1997).
Callan, R.L., 2016. Analyzing Software Using Unintentional Electromagnetic Emanations from Computing Devices. Ph.D. thesis. Georgia Institute of Technology.
Callan, R., Zajic, A., Prvulovic, M., 2014. A practical methodology for measuring the
side-channel signal available to the attacker for instruction-level events. In:
47th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO). IEEE, pp. 242e254.
Callan, R., Popovic, N., Daruna, A., Pollmann, E., Zajic, A., Prvulovic, M., 2015.
Comparison of electromagnetic side-channel energy available to the attacker
from different computer systems. In: IEEE International Symposium on Electromagnetic Compatibility (EMC). IEEE, pp. 219e223.
Callan, R., Zaji

c, A., Prvulovic, M., 2015. FASE: finding amplitude-modulated sidechannel emanations. In: ACM SIGARCH Computer Architecture News, vol. 43.
ACM, pp. 592e603.
Callan, R., Behrang, F., Zajic, A., Prvulovic, M., Orso, A., 2016. Zero-overhead profiling
via em emanations. In: Proceedings of the 25th International Symposium on
Software Testing and Analysis. ACM, pp. 401e412.
Camurati, G., Poeplau, S., Muench, M., Hayes, T., Francillon, A., 2018. Screaming
channels: when electromagnetic side channels meet radio transceivers. In:
Proceedings of the 25th ACM Conference on Computer and Communications
Security (CCS), CCS ’18. ACM.
Cass, S., 2013. A $40 software-defined radio. IEEE Spectrum 50 (7), 22e23.
Chari, S., Rao, J.R., Rohatgi, P., 2002. Template attacks. In: International Workshop on
Cryptographic Hardware and Embedded Systems (CHES). Springer, pp. 13e28.
Cheddad, A., Condell, J., Curran, K., Mc Kevitt, P., 2010. Digital image steganography:
survey and analysis of current methods. Signal Process. 90 (3), 727e752.
ChipWhisperer Embedded Hardware Security Toolchain. https://newae.com/tools/
chipwhisperer/, 2018-01-21.
Clark, S.S., Mustafa, H., Ransford, B., Sorber, J., Fu, K., Xu, W., 2013. Current events:
identifying webpages by tapping the electrical outlet. In: European Symposium
on Research in Computer Security (ESORICS). Springer, pp. 700e717.
Corey, V., Peterman, C., Shearin, S., Greenberg, M.S., Van Bokkelen, J., 2002. Network
forensics analysis. IEEE Internet Computing 6 (6), 60e66.
Danev, B., Zanetti, D., Capkun, S., 2012. On physical-layer identification of wireless
devices. ACM Comput. Surv. 45 (1), 6.
Deppensmith, R.D., Stone, S.J., 2014. Optimized fingerprint generation using unintentional emission radio-frequency distinct native attributes (rf-dna). In:
Aerospace and Electronics Conference, NAECON 2014-IEEE National. IEEE,
pp. 327e330.

, P., Quisquater, J.-J., Willems, J.-L.,
Dhem, J.-F., Koeune, F., Leroux, P.-A., Mestre
1998. A practical implementation of the timing attack. In: International
Conference on Smart Card Research and Advanced Applications. Springer,
pp. 167e182.
Du, X., Le-Khac, N.-A., Scanlon, M., 2017. Evaluation of digital forensic process
models with respect to digital forensics as a service. In: Proceedings of the 16th
European Conference on Cyber Warfare and Security (ECCWS 2017). ACPI,
Dublin, Ireland, pp. 573e581.
Dubendorfer, C.K., 2013. Using RF-DNA Fingerprints to Discriminate ZigBee Devices
in an Operational Environment. Master’s thesis. Air Force Institute of

Technology, Wright-Patterson Air Force Base, Ohio.
Casey, E., Stellatos, G.J., 2008. The impact of full disk encryption on digital forensics.
ACM SIGOPS - Oper. Syst. Rev. 42 (3), 93e98.
Elibol, F., Sarac, U., Erer, I., 2012. Realistic eavesdropping attacks on computer displays with low-cost and mobile receiver system. In: Proceedings of the 20th
European Signal Processing Conference (EUSIPCO). IEEE, pp. 1767e1771.
Entriken, W.. System Bus Radio. https://github.com/fulldecent/system-bus-radio,
2018-01-26.

rard, B., Tibouchi, M., 2017. Side-channel attacks on bliss
Espitau, T., Fouque, P.-A., Ge
lattice-based signatures: exploiting branch tracing against strongswan and
electromagnetic emanations in microcontrollers. In: Proceedings of the 2017
ACM SIGSAC Conference on Computer and Communications Security. ACM,
pp. 1857e1874.
Ettus, M., Braun, M., 2015. The Universal Software Radio Peripheral (Usrp) Family of
Low-Cost Sdrd, Opportunistic Spectrum Sharing and White Space Access. The
Practical Reality, pp. 3e23.
Fahl, S., Harbach, M., Perl, H., Koetter, M., Smith, M., 2013. Rethinking SSL development in an appified world. In: Proceedings of the 2013 ACM SIGSAC Conference on Computer & Communications Security. ACM, pp. 49e60.
Fritzke, A.W., 2012. Obfuscating against Side-Channel Power Analysis Using Hiding
Techniques for Aes. Master’s thesis. Air Force Institute of Technology, WrightPatterson Air Force Base, Ohio.
Gandolfi, K., Mourtel, C., Olivier, F., 2001. Electromagnetic analysis: concrete results.
In: International Workshop on Cryptographic Hardware and Embedded Systems (CHES). Springer, pp. 251e261.
R. Getz, B. Moeckel, Understanding and eliminating EMI in Microcontroller Applications, National Semiconductor.
Genkin, D., Shamir, A., Tromer, E., 2014. RSA key extraction via low-bandwidth
acoustic cryptanalysis. In: International Cryptology Conference. Springer,
pp. 444e461.
Genkin, D., Pachmanov, L., Pipman, I., Tromer, E., 2015. Stealing keys from PCs using
a radio: cheap electromagnetic attacks on windowed exponentiation. In: International Workshop on Cryptographic Hardware and Embedded Systems
(CHES). Springer, pp. 207e228.
Genkin, D., Pipman, I., Tromer, E., 2015. Get your hands off my laptop: physical sidechannel key-extraction attacks on pcs. Journal of Cryptographic Engineering 5
(2), 95e112.
Genkin, D., Pachmanov, L., Pipman, I., Tromer, E., 2016. ECDH key-extraction via
low-bandwidth electromagnetic attacks on PCs. In: Cryptographers Track at the
RSA Conference. Springer, pp. 219e235.
Genkin, D., Pachmanov, L., Pipman, I., Tromer, E., Yarom, Y., 2016. ECDSA key
extraction from mobile devices via nonintrusive physical side channels. In:
Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security. ACM, pp. 1626e1638.
Golmie, N., Rebala, O., Chevrollier, N., 2003. Bluetooth adaptive frequency hopping
and scheduling. In: Military Communications Conference (MILCOM’03), vol. 2.
IEEE, pp. 1138e1142.
Goubin, L., 2003. A refined power-analysis attack on elliptic curve cryptosystems.
In: International Workshop on Public Key Cryptography. Springer, pp. 199e211.
Goudos, S.K., Rekanos, I.T., Sahalos, J.N., 2008. EMI reduction and ICs optimal
arrangement inside high-speed networking equipment using particle swarm
optimization. IEEE Trans. Electromagn Compat. 50 (3), 586e596.
Han, Y., Etigowni, S., Liu, H., Zonouz, S., Petropulu, A., 2017. Watch me, but don't
touch me! contactless control flow monitoring via electromagnetic emanations.
In: Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security. ACM, pp. 1095e1108.
Hayashi, Y.-i., 2016. State-of-the-art research on electromagnetic information security. Radio Sci. 51 (7), 1213e1219.
Hayashi, Y.-i., Homma, N., Mizuki, T., Shimada, H., Aoki, T., Sone, H., Sauvage, L.,
Danger, J.-L., 2013. Efficient evaluation of em radiation associated with information leakage from cryptographic devices. IEEE Trans. Electromagn Compat.
55 (3), 555e563.
Hay B., Bishop M., Nance K., Live analysis: Progress and challenges, IEEE Security &
Privacy 7 (2).
Hongxin, Z., Yuewang, H., Jianxin, W., Yinghua, L., Jinling, Z., 2009. Recognition of
electro-magnetic leakage information from computer radiation with svm.
Comput. Secur. 28 (1e2), 72e76.
Hoog, A., 2011. Android Forensics: Investigation, Analysis and Mobile Security for
Google Android. Elsevier.
Hutter, M., Mangard, S., Feldhofer, M., 2007. Power and EM attacks on passive 13.56
MHz RFID devices. In: International Workshop on Cryptographic Hardware and
Embedded Systems (CHES), vol. 7. Springer, pp. 320e333.
Ishai, Y., Sahai, A., Wagner, D., 2003. Private circuits: securing hardware against
probing attacks. In: Annual International Cryptology Conference. Springer,
pp. 463e481.
ISO/IEC 17825:2016. Testing Methods for the Mitigation of Non-invasive Attack
Classes against Cryptographic Modules. https://www.iso.org/obp/ui/#iso:std:
iso-iec:17825:ed-1:v1:en, 2018-01-21.
ISO/IEC TS 30104:2015. Physical Security Attacks, Mitigation Techniques and Security Requirements. https://www.iso.org/standard/56890.html, 2018-01-21.
Iyer, V., Hermans, F., Voigt, T., 2015. Detecting and avoiding multiple sources of
interference in the 2.4 GHz spectrum. In: 12th European Conference on Wireless Sensor Networks (EWSN). Springer, pp. 35e51.
Jabbar, M., Rahman, M.A., 1991. Radio frequency interference of electric motors and
associated controls. IEEE Trans. Ind. Appl. 27 (1), 27e31.

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54
Kasper, T., Oswald, D., Paar, C., 2009. EM Side-Channel Attacks on Commercial
Contactless Smartcards Using Low-Cost Equipment. Information Security Applications, pp. 79e93.
Kellogg, B., Parks, A., Gollakota, S., Smith, J.R., Wetherall, D., 2014. Wi-Fi backscatter:
internet connectivity for RF-powered devices. In: ACM SIGCOMM Computer
Communication Review, vol. 44. ACM, pp. 607e618.
€ffer, M., Moon, S., 2008. Differential side channel analysis attacks on
Kim, C., Schla
FPGA implementations of ARIA. ETRI J. 30 (2), 315e325.
Kim, T., Lee, S., Choi, D., Yoon, H., 2016. Protecting secret keys in networked devices
with table encoding against power analysis attacks. J. High Speed Netw. 22 (4),
293e307.
Kocher, P.C., 1996. Timing attacks on implementations of diffie-hellman, rsa, dss,
and other systems. In: Annual International Cryptology Conference. Springer,
pp. 104e113.
Kocher, P., Jaffe, J., Jun, B., 1999. Differential power analysis. In: Advances in Cryptology (CRYPTO ‘99). Springer, 789e789.
Kocher, P., Jaffe, J., Jun, B., Rohatgi, P., 2011. Introduction to differential power
analysis. Journal of Cryptographic Engineering 1 (1), 5e27.
Kolter, J.Z., Maloof, M.A., 2004. Learning to detect malicious executables in the wild.
In: Proceedings of the 10th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM, pp. 470e478.
Laput, G., Yang, C., Xiao, R., Sample, A., Harrison, C., 2015. Em-sense: touch recognition of uninstrumented, electrical and electromechanical objects. In: Proceedings of the 28th Annual ACM Symposium on User Interface Software &
Technology. ACM, pp. 157e166.
Lerman, L., Bontempi, G., Markowitch, O., 2011. Side channel attack: an approach
based on machine learning. In: Proceedings of 2nd International Workshop on
Constructive Side-Channel Analysis and Security Design (COSADE). Schindler
and Huss, pp. 29e41.
Lillis, D., Becker, B., O'Sullivan, T., Scanlon, M., 2016. Current challenges and future
research areas for digital forensic investigation. In: The 11th ADFSL Conference
on Digital Forensics, Security and Law (CDFSL 2016). ADFSL, Daytona Beach, FL,
USA, pp. 9e20.
Liu, V., Parks, A., Talla, V., Gollakota, S., Wetherall, D., Smith, J.R., 2013. Ambient
backscatter: wireless communication out of thin air. Comput. Commun. Rev. 43
(4), 39e50.
Liu, F., Yarom, Y., Ge, Q., Heiser, G., Lee, R.B., 2015. Last-level cache side-channel
attacks are practical. In: IEEE Symposium on Security and Privacy. IEEE,
pp. 605e622.
Liu, W., Huang, K., Zhou, X., Durrani, S., 2017. Full-duplex backscatter interference
networks based on time-hopping spread spectrum. IEEE Trans. Wirel. Commun.
16 (7), 4361e4377. https://doi.org/10.1109/TWC.2017.2697864.
Lukacs, M.W., Zeqolari, A.J., Collins, P.J., Temple, M.A., 2015. rf-dna fingerprinting for
antenna classification. IEEE Antennas Wirel. Propag. Lett. 14, 1455e1458.
Marinov, M.. TempestSDR Remote Video Eavesdropping Using a Software-Defined
Radio Platform. https://github.com/martinmarinov/TempestSDR, 2018-02-01.
Maxwell, J.C., 1865. A dynamical theory of the electromagnetic field. Phil. Trans. Roy.
Soc. Lond. 155, 459e512.
Messerges, T.S., Dabbish, E.A., Sloan, R.H., 1999. Investigations of power analysis
attacks on smartcards. Smartcard 99, 151e161.
Mukkamala, S., Sung, A.H., 2003. Identifying significant features for network
forensic analysis using artificial intelligent techniques. International Journal of
Digital Evidence 1 (4), 1e17.
Nazari, A., Sehatbakhsh, N., Alam, M., Zajic, A., Prvulovic, M., 2017. EDDIE: EM-based
detection of deviations in program execution. In: Proceedings of the 44th
Annual International Symposium on Computer Architecture. ACM,
pp. 333e346.
Nie, N.H., Erbring, L., 2000. Internet and Society, vol. 3. Stanford Institute for the
Quantitative Study of Society, pp. 14e19.
Ossmann, M.. Software Defined Radio with Hackrf, Great Scott Gadgets. https://
greatscottgadgets.com/sdr.
Ott, H.W., 2011. Electromagnetic Compatibility Engineering. John Wiley & Sons.
O'Flynn, C., Chen, Z.D., 2014. ChipWhisperer: an open-source platform for hardware
embedded security research. In: International Workshop on Constructive SideChannel Analysis and Secure Design. Springer, pp. 243e260.
O'Malley, S.J., Choo, K.-K.R., 2014. Bridging the air gap: inaudible data exfiltration by
insiders. In: 20th Americas Conference on Information Systems (AMCIS). Association for Information Systems.
Peeters, E., Standaert, F.-X., Quisquater, J.-J., 2007. Power and electromagnetic
analysis: improved model, consequences and comparisons, Integration. the
VLSI Journal 40 (1), 52e60. https://doi.org/10.1016/j.vlsi.2005.12.013.
Prvulovic, M., Zaji

c, A., Callan, R.L., Wang, C.J., 2017. A method for finding frequencymodulated and amplitude-modulated electromagnetic emanations in computer
systems. IEEE Trans. Electromagn Compat. 59 (1), 34e42.
Quisquater, J.-J., Samyde, D., 2001. Electromagnetic Analysis (EMA): Measures and
Counter-measures for Smart Cards. Smart Card Programming and Security,
pp. 200e210.
Ramsay, C., Lohuis, J.. White Paper: TEMPEST Attacks against AES Covertly Stealing
Keys for 200 Euros, Tech. Rep., Fox-IT, Netherlands. https://www.fox-it.com/nl/
wp-content/uploads/sites/12/Tempest_attacks_against_AES.pdf.
Reising, D.R., 2012. Exploitation of RF-DNA for Device Classification and Verification
Using GRLVQI Processing. Master’s thesis. Air Force Institute of Technology,
Wright-Patterson Air Force Base, Ohio.
Riscure. Inspecture SCA: Side-Channel Analysis Tool for Embedded Systems. https://
www.riscure.com/security-tools/inspector-sca/, 2018-01-21.

53

Saboia, P., Carvalho, T., Rocha, A., 2011. Eye specular highlights telltales for digital
forensics: a machine learning approach. In: 18th IEEE International Conference
on Image Processing (ICIP). IEEE, pp. 1937e1940.
Sanfelix, E., Mune, C., de Haas, J., 2015. Unboxing the White-Box Practical Attacks
against Obfuscated Ciphers, Black Hat Europe.
Saputra, H., Vijaykrishnan, N., Kandemir, M., Irwin, M.J., Brooks, R., Kim, S.,
Zhang, W., 2003. Masking the energy behavior of des encryption. In: Proceedings of the Conference on Design, Automation and Test in Europe-Volume
1. IEEE Computer Society, p. 10084.
Sayakkara, A., Le-Khac, N.-A., Scanlon, M., 2018. Electromagnetic side-channel attacks: potential for progressing hindered digital forensic analysis. In: Proceedings of the International Workshop on Speculative Side Channel Analysis
(WoSSCA 2018). ACM, Amsterdam, Netherlands. https://doi.org/10.1145/
3236454.3236512.
Sayakkara, A., Le-Khac, N.-A., Scanlon, M., 2018. Accuracy enhancement of electromagnetic side-channel attacks on computer monitors. In: Proceedings of the
13th International Conference on Availability, Reliability and Security, ARES
2018. ACM, New York, NY, USA. https://doi.org/10.1145/3230833.3234690, 15:
1e15:9. http://doi.acm.org/10.1145/3230833.3234690.
Schulz, M., Klapper, P., Hollick, M., Tews, E., Katzenbeisser, S., 2016. Trust the wire,
they always told me!: on practical non-destructive wire-tap attacks against
ethernet. In: Proceedings of the 9th ACM Conference on Security & Privacy in
Wireless and Mobile Networks. ACM, pp. 43e48.
Smith, S.W., 1997. The Scientist and Engineer's Guide to Digital Signal Processing.
California Technical Publishing. Ch. 28.
N. Sohaib ul Hassan, Em side channel analysis on complex soc architectures, MSc
thesis, Tampere University of Technology.
Soltani, S., Seno, S.A.H., 2017. A survey on digital evidence collection and analysis.
In: 7th International Conference on Computer and Knowledge Engineering
(ICCKE). IEEE, pp. 247e253.
Souvignet, T., Frinken, J., 2013. Differential power analysis as a digital forensic tool.
Forensic Sci. Int. 230 (1e3), 127e136.
Spreitzer, R., Moonsamy, V., Korak, T., Mangard, S., 2018. Systematic classification of
side-channel attacks: a case study for mobile devices. IEEE Communications
Surveys & Tutorials 20 (1), 465e488.
Stagner, C.B., 2013. Detecting and Locating Electronic Devices Using Their Unintended Electromagnetic Emissions. Missouri University of Science and
Technology.
Standaert, F.-X., Malkin, T., Yung, M., 2009. A unified framework for the analysis of
side-channel key recovery attacks. In: Eurocrypt, vol. 5479. Springer,
pp. 443e461.
Stojkoska, B.L.R., Trivodaliev, K.V., 2017. A review of internet of things for smart
home: challenges and solutions. J. Clean. Prod. 140, 1454e1464.
Stone, B.D., Stone, S.J., 2015. Radio frequency based reverse engineering of microcontroller program execution. In: National Aerospace and Electronics Conference (NAECON), 2015. IEEE, pp. 159e164.
Stone, B., Stone, S., 2016. Comparison of radio frequency based techniques for device
discrimination and operation identification. In: 11th International Conference
on Cyber Warfare and Security: ICCWS2016. Academic Conferences and Publishing Limited, p. 475.
Stone, S.J., Temple, M.A., Baldwin, R.O., 2015. Detecting anomalous programmable
logic controller behavior using rf-based hilbert transform features and a
correlation-based verification process. International Journal of Critical Infrastructure Protection 9, 41e51.
Tuttlebee, W.H., 2003. Software Defined Radio: Enabling Technologies. John Wiley
& Sons.
van de Wiel, E., Scanlon, M., Le-Khac, N.-A., 2018. Enabling non-expert analysis of
large volumes of intercepted network traffic. In: Peterson, G., Shenoi, S. (Eds.),
Advances in Digital Forensics XIV. Springer International Publishing, Cham,
pp. 183e197.
Van Eck, W., 1985. Electromagnetic radiation from video display units: an eavesdropping risk? Comput. Secur. 4 (4), 269e286.
Wakabayashi, S., Maruyama, S., Mori, T., Goto, S., Kinugawa, M., Hayashi, Y.-i., 2017.
Poster: is active electromagnetic side-channel attack practical?. In: Proceedings
of the 2017 ACM SIGSAC Conference on Computer and Communications Security. ACM, pp. 2587e2589.
Wang, X., Zhou, Q., Harer, J., Brown, G., Qiu, S., Dou, Z., Wang, J., Hinton, A.,
Gonzalez, C.A., Chin, P., 2018. Deep learning-based classification and anomaly
detection of side-channel signals. In: Cyber Sensing 2018, vol. 10630. International Society for Optics and Photonics, p. 1063006.
Witteman, M., Oostdijk, M., 2008. Secure application programming in the presence
of side channel attacks. In: RSA Conference, vol. 2008.
Witteman, M.F., van Woudenberg, J.G., Menarini, F., 2011. Defeating RSA multiplyalways and message blinding countermeasures. In: Cryptographers Track at
the RSA Conference (CT-RSA), vol. 6558. Springer, pp. 77e88.
Wolfe, H., 2003. Setting up an electronic evidence forensics laboratory. Comput.
Secur. 22 (8), 670e672.
Xu, R., Zhu, L., Wang, A., Du, X., Choo, K.-K.R., Zhang, G., Gai, K., 2018. Side-channel
attack on a protected rfid card. IEEE Access, 1e1.
Yang, C., Sample, A.P., 2016. Em-id: tag-less identification of electrical devices via
electromagnetic emissions. In: IEEE International Conference on RFID (RFID).
IEEE, pp. 1e8.
Yang, C.J., Sample, A.P., 2017. EM-comm: touch-based communication via modulated electromagnetic emissions. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 1 (3), 118.

54

A. Sayakkara et al. / Digital Investigation 29 (2019) 43e54

Yang, K., Hicks, M., Dong, Q., Austin, T., Sylvester, D., 2017. Exploiting the analog
properties of digital circuits for malicious hardware. Commun. ACM 60 (9),
83e91.
Yilmaz, B.B., Callan, R.L., Prvulovic, M., Zaji

c, A., 2018. Capacity of the em covert/
side-channel created by the execution of instructions in a processor. IEEE
Trans. Inf. Forensics Secur. 13 (3), 605e620.
Zajic, A., Prvulovic, M., 2014. Experimental demonstration of electromagnetic information leakage from modern processor-memory systems. IEEE Trans. Electromagn Compat. 56 (4), 885e893.
Zankl, A., Seuschek, H., Irazoqui, G., Gulmezoglu, B., 2018. Side-channel attacks in

the internet of things: threats and challenges. In: Solutions for Cyber-Physical
Systems Ubiquity. IGI Global, pp. 325e357.
Zdziarski, J., 2008. iPhone Forensics: Recovering Evidence, Personal Data, and
Corporate Assets. O'Reilly Media, Inc.
Zhang, Y., Juels, A., Reiter, M.K., Ristenpart, T., 2014. Cross-tenant side-channel attacks in paas clouds. In: Proceedings of the 2014 ACM SIGSAC Conference on
Computer and Communications Security. ACM, pp. 990e1003.
Zhang, P., Bharadia, D., Joshi, K., Katti, S., 2016. Hitchhike: practical backscatter using
commodity wifi. In: Proceedings of the 14th ACM Conference on Embedded
Network Sensor Systems. ACM, pp. 259e271.