Buy ebook Cyber Security in Critical Infrastructures A Game Theoretic Approach Stefan Rass cheap price
Buy ebook Cyber Security in Critical Infrastructures A Game Theoretic Approach Stefan Rass cheap price
Buy ebook Cyber Security in Critical Infrastructures A Game Theoretic Approach Stefan Rass cheap price
com
https://textbookfull.com/product/cyber-security-
in-critical-infrastructures-a-game-theoretic-
approach-stefan-rass/
https://textbookfull.com/product/cross-layer-design-for-secure-and-
resilient-cyber-physical-systems-a-decision-and-game-theoretic-
approach-quanyan-zhu/
textbookfull.com
https://textbookfull.com/product/making-eden-how-plants-transformed-a-
barren-planet-first-edition-beerling/
textbookfull.com
https://textbookfull.com/product/research-first-edition-philip-kerr/
textbookfull.com
https://textbookfull.com/product/understanding-sustainable-
development-john-blewitt/
textbookfull.com
https://textbookfull.com/product/finance-and-the-behavioral-prospect-
risk-exuberance-and-abnormal-markets-1st-edition-james-ming-chen-auth/
textbookfull.com
Arrhythmia Essentials Brian Olshansky
https://textbookfull.com/product/arrhythmia-essentials-brian-
olshansky/
textbookfull.com
Advanced Sciences and Technologies for Security Applications
Stefan Rass
Stefan Schauer
Sandra König
Quanyan Zhu
Cyber-Security
in Critical
Infrastructures
A Game-Theoretic Approach
Advanced Sciences and Technologies for
Security Applications
Series Editor
Anthony J. Masys, Associate Professor, Director of Global Disaster Management,
Humanitarian Assistance and Homeland Security, University of South Florida,
Tampa, USA
Advisory Editors
Gisela Bichler, California State University, San Bernardino, CA, USA
Thirimachos Bourlai, Statler College of Engineering and Mineral Resources,
West Virginia University, Morgantown, WV, USA
Chris Johnson, University of Glasgow, Glasgow, UK
Panagiotis Karampelas, Hellenic Air Force Academy, Attica, Greece
Christian Leuprecht, Royal Military College of Canada, Kingston, ON, Canada
Edward C. Morse, University of California, Berkeley, CA, USA
David Skillicorn, Queen’s University, Kingston, ON, Canada
Yoshiki Yamagata, National Institute for Environmental Studies, Tsukuba, Ibaraki,
Japan
Indexed by SCOPUS
The series Advanced Sciences and Technologies for Security Applications com-
prises interdisciplinary research covering the theory, foundations and domain-
specific topics pertaining to security. Publications within the series are peer-
reviewed monographs and edited works in the areas of:
– biological and chemical threat recognition and detection (e.g., biosensors,
aerosols, forensics)
– crisis and disaster management
– terrorism
– cyber security and secure information systems (e.g., encryption, optical and
photonic systems)
– traditional and non-traditional security
– energy, food and resource security
– economic security and securitization (including associated infrastructures)
– transnational crime
– human security and health security
– social, political and psychological aspects of security
– recognition and identification (e.g., optical imaging, biometrics, authentication
and verification)
– smart surveillance systems
– applications of theoretical frameworks and methodologies (e.g., grounded theory,
complexity, network sciences, modelling and simulation)
Together, the high-quality contributions to this series provide a cross-disciplinary
overview of forefront research endeavours aiming to make the world a safer place.
The editors encourage prospective authors to correspond with them in advance of
submitting a manuscript. Submission of manuscripts should be made to the Editor-
in-Chief or one of the Editors.
Cyber-Security in Critical
Infrastructures
A Game-Theoretic Approach
Stefan Rass Stefan Schauer
Universitaet Klagenfurt Austrian Institute of Technology GmbH
Klagenfurt, Austria Wien, Austria
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
. . . to our families. . .
Contents
Part I Introduction
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1 What are Critical Infrastructures? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Security Challenges for Critical Infrastructures . . . . . . . . . . . . . . . . . . . . 5
1.2.1 Natural and Physical Threats. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.2 Cyber Threats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 Advanced Persistent Threats (APT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.1 Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.2 Life-Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4 Selected Real-Life Incidents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.4.1 The Blackout in Italy (2003) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.4.2 The Transportation Gridlock in Switzerland (2005) . . . . . . 13
1.4.3 The Attack on the Ukrainian Power Grid (2015) . . . . . . . . . 14
1.4.4 The WannaCry and NotPetya Malware Infections
(2017) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4.5 The Blackout in Venezuela (2019) . . . . . . . . . . . . . . . . . . . . . . . . . 17
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2 Critical Infrastructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.1 Examples and Definitions of Critical Infrastructures . . . . . . . . . . . . . . . 21
2.1.1 What Makes an Infrastructure “Critical”? . . . . . . . . . . . . . . . . . 21
2.1.2 Threats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2 Cyber Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.1 Hacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.2 Malware and Ransomware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3 Physical Security of Critical Infrastructures . . . . . . . . . . . . . . . . . . . . . . . . 27
2.3.1 Eavesdropping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.3.2 Jamming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.3.3 Terrorist Attacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4 Cyber-Physical Security of Critical Infrastructures. . . . . . . . . . . . . . . . . 28
vii
viii Contents
Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
Part I
Introduction
Chapter 1
Introduction
Abstract This chapter opens the book by introducing the characteristics and
particularities of critical infrastructures. Their existence and interplay forms a vital
pillar of contemporary societies, and their protection is a top duty of governments
and security research. Recent years have shown a paradigm shift of cyber-attacks
from specific individual threat and attack scenarios, to a modern combination of
various attack types and strategies to what we call an advanced persistent threat
(APT) today. This term describes a diverse class of attacks that all share a set of
common characteristics, which presents new challenges to security that demand
urgent and continuous action by practitioners, researchers and every stakeholder
of a critical infrastructure. The main focus of the book is describing game theory as
a tool to establish security against APTs, and to this end, the introduction here starts
with the abstract characteristics of an APT, showcasing them with a set of selected
real-life documented cases of APTs that ends the chapter.
In today’s life, society is using and relying on numerous services, which satisfy the
basic needs of people and guarantee a smooth flow of the everyday life. Among
those services are the supply with basic resources (e.g., electricity, communication,
heating, etc.), vital supplies (e.g., water, food, medicine and health care, etc.) as
well as industrial goods (e.g., oil, gas, etc.) and general services (e.g., transportation,
cash and financial services, etc.). The organizations and companies providing these
services are called Critical Infrastructures (CIs) and represent the backbone of
today’s society. A core characteristic of a CI is that any failure of a CI, either in
er ity
t ic
wa ct
r
e
el
u-
r t. m
mn
po o
s co ti
an le a
tr te nic
g
in ce al
a nk an d ic ly
b fin e p
msup
&
y t
nc en
ge ces m s
r rn ce
e vi ve vi
em ser go ser
part or as a whole, will have considerable impact not only on the infrastructure itself
but also on the social well-being of people [6].
Over the last decades, the interrelations among CIs have increased such that CIs
have become more and more dependent on each other. These dependencies emerged
due to the fact that national and international supply chains have become more
important and many infrastructures depend on the resources other infrastructures
provide (cf. Fig. 1.1). The most prominent example is power supply; almost every
modern organization relies on a continuous availability of electricity such that CI
can provide their main services to full capacity. A shortage in power distribution or
even a blackout will either reduce the general output of the depending CIs or even
make it impossible for the CIs to provide its service. Figure 1.1 shows a schematic
illustration of such dependencies among CIs in multiple domains, which go far
beyond the supply with electric power.
Further, CIs also heavily rely on the exchange of information due to the uprising
digitalization of those infrastructures. Hence, a smooth operation of the communi-
cation infrastructure and a uninterrupted connection to the Internet has become a
core requirement. For example, CIs are using Industrial Control Systems (ICSs)
and Supervisory Control and Data Acquisition (SCADA) systems to remotely
monitor physical systems within their premises and control multiple (even highly
important) processes. If communication is interrupted, those systems might shut
down immediately or simply cannot be reached any more (cf. real-life examples in
Sect. 1.4) and the CI might have to operate the respective systems manually (if this
is possible at all).
Due to these facts, the CIs within a nation or even within a region have
evolved into a highly complex and sensitive infrastructure network with a variety
of interdependencies among them, as illustrated in Fig. 1.1. Thus, incidents within
a single CI can have far-reaching consequences, not only causing damages and
financial losses within that CI but also affecting multiple other infrastructures as
well as society as a whole. For example, a local power outage can cause a cellular
antenna tower to shut down and interrupt the mobile communication network in that
1.2 Security Challenges for Critical Infrastructures 5
area. Such indirect consequences are often referred to as cascading effects and they
impose a major problem. In general, they are hard to identify and thus difficult to
assess as well as to prepare for. Therefore, during a risk analysis or the general risk
management process of a CI, a particular focus lies on cascading effects. Chapter 2
provides more detailed information on how to model and describe the cascading
effects among CIs and how to integrate them into a CI’s risk analysis. To better
illustrate how cascading effects can manifest themselves in real-life, a few examples
of documented incidents over the last two decades follow in Sect. 1.4.
In general, CIs have to face a broad variety of security challenges and threats in
their daily business. These are ranging from natural disasters of various kinds over
technical faults and human failure up to intentional attacks from the physical and the
cyber domain. Depending on the geographical location of a CI, its business domain
and the applied technical equipment, these threats can vary. In other words, a CI
located at the sea (e.g., a sea port) has to deal with different natural disasters than
an infrastructure located in the mountains (e.g., a dam). Further, an infrastructure
from the chemical industry has to deal with other technical threats than an airport or
a telecommunication provider.
Therefore, CI operators and the respective security officers need to become
aware of their individual threat landscape, adapt to the requirements towards risk
management and implement a large collection of security measures to be prepared
against these threats or counter them. In this context, threats stemming from natural
disasters and technical failures are quite static and change slowly (if at all) over time.
However, threats from the cyber domain have increased massively, in particular over
the last decade, and are considered to be highly dynamic due to the rapidly changing
software landscape. In addition to the increasing digitization in CIs, the effects of
a cyber attack can be tremendous (as briefly sketched in the following Sects. 1.4.3
and 1.4.4), such that cyber threats have become a major concern for CI operators.
Section 1.2.1 provides a short overview on the security challenges stemming
from natural disasters and technical failures. A strong focus is here on the security
challenges stemming from cyber threats in Sect. 1.2.2. The information presented
in both sections is coming mainly from well-established threat catalogs, i.e., the
European Anion Agency for Cybersecurity (ENISA) Threat Landscape [20] and the
German Federal Office for Information Security (BSI) IT-Grundschutz Catalog [2].
6 1 Introduction
In the context of the physical domain, natural disasters represent threats with a
potentially very high impact on an entire CI as well as on its environment. Examples
are extreme weather conditions, storms or floods or very high temperatures but
also fires or lightning strikes need to be considered. Whether an infrastructure is
particularly exposed to some specific natural disaster depends on its geographical
location and the environmental conditions.
With a lot of communication services using the Internet or at least being reachable
from the outside world, shutting down such services by generating a high load of
traffic is nowadays a common attack vector. This is achieved by creating massive
amounts of connection requests targeted at a specific device, which is then no longer
able to perform it tasks due to the continuous handling of the requests. Additionally,
the current trend towards the Internet of Things (IoT), i.e., connecting small devices
to the Internet with low or no security measures installed, makes it much easier for
adversaries to take over control of a huge number of devices and thus creating a
Distributed Denial of Service (DDoS) attack.
In recent years, malicious code and software, i.e., “malware”, has been on the
rise and becoming a major threat for cyber systems. The variety of malware, its
sophistication and the number of systems targeted by it are huge. Hence, it is
very difficult to protect systems from being infected by malware. Additionally, a
wide range of functions can be implemented by malware such that it can tamper
cyber systems on the communication as well as on the operation level. Hence, the
proper functioning of a CI can be affected (as described in Sect. 1.4.4), leading to
a multitude of different effects. In the context of malware, these effects can range
from the retrieval of information by a malicious party up to an injection of fraudulent
information or commends into the communication channel. Ransomware on the
other hand follows a different principle, i.e., it encrypts the entire infected system
or crucial data stored on it and demands a ransom (often some amount of Bitcoin)
in exchange for the decryption key.
Another threat vector that has gained popularity over the last years are phishing
mails. Such emails appear to the recipient as genuine mails but contain either a
link pointing to a malicious website or an attachment (e.g., a Portable Document
File (PDF) or Office document) equipped with some malicious code. In both
cases, the goal of the attacker is to trick the recipient into clicking on the link or
opening the attachment. In this case, a malware or ransomware is loaded onto the
recipient’s system and is then causing additional damage (see also examples in the
following section). In general, such phishing mails are not targeted and sent to a
large number of recipients, which makes it easier for an observant user to identify
the attempt. However, some phishing attacks are more elaborated and targeted to
a certain organization or even to one specific user or a specific user group within
8 1 Introduction
an organization. Such a targeted attack is then called spear phishing attack. In that
case, it is much harder even for a trained user to identify the mail as a phishing
attempt. Still, to craft such an attack, a lot more effort is required. In particular,
social engineering (see the following paragraph) is applied as an important tool in
the preparation phase to tailor the mail and thus the attack to perfectly fit to the
recipients.
With all the technical means available for protecting cyber systems today, the main
cause of incidents in the cyber domain is human failure. Hence, social engineering
approaches [12, 19] try to exploit this concept and trick the person operating a
system to introduce some malicious software into the system. In this way, an
adversary is able to surpass all technological protection mechanism (e.g., firewalls
or intrusion protection systems) and get direct access to the internal networks of an
organization. Since control and protection mechanisms are usually not as strict in
the internal network as they are on the perimeter, this makes social engineering such
a highly critical threat.
1.3.1 Characteristics
Among today’s security challenges for CIs, one of the most complex and severe are
Advanced Persistent Threats (APTs). These attacks combine different techniques
and attack strategies from the physical and the cyber domain and are very difficult
to detect. Although there is no common consensus on how to specifically define an
APT due to their variety and diversity, some shared characteristics are commonly
accepted as distinguishing features in relation to other attacks. Among these, an
APT has at least the following characteristics:
1. Targeted: an APT usually aims at a very specific victim and goal. For the victim,
this means that the attack is often “hand-crafted” to fit the particularities of the
target systems, which in turn means that standard technical security precautions
like firewalls, virus scanners or similar are of limited efficacy. With respect to the
attacker’s goal, we can broadly distinguish two classes of APTs, both of which
require distinct modeling and treatment:
• Gaining control over a running service: the adversary’s goal is to take over
the system for as long as s/he can. However, this does not necessarily mean
that the adversary wishes to shut down the targeted system, in particular, if
the system should act as a platform from which further attacks are mounted
1.3 Advanced Persistent Threats (APT) 9
(like a base camp). For example, the APT could aim at limiting bandwidth or
blocking certain content in a multimedia streaming network. In this case, the
attacker’s goal would be to keep the system alive, but at a level that s/he can
maintain or cut down at will. An example of a game model to counteract this
type of APT is FlipIt, which we discuss in Sect. 9.3.
• Hitting a vital target: the purpose of this attack is to shut down the targeted
system or cause at least some permanent damage. Here, unlike in the previous
case, the time spent in the system is of less importance than the point that
the attacker can reach within the system. For example, if the attacker wants
to disrupt a nuclear power reactor, remaining in the systems for any period
of time, no matter how long or short, that lets the adversary cause a nuclear
meltdown or a comparable catastrophe would be the APT goal here. An
example game model designed for defense against this second type of APT
is Cut-The-Rope, discussed in Sect. 9.4.
2. Stealthy and Slow: an additional goal of an APT is often to stay undetected.
While this is not surprising and can be considered as a goal for many attacks,
the distinguishing fact of an APT is the time scale; an APT is hardly an ad hoc
nor quick strike, but rather follows a sophisticated plan of penetration to never
cause much attention and therefore always remain under the radar. The rationale
is to prepare the attack up to a point where it is hardly – if at all – possible to
counteract if the attack is detected.
3. Unbounded resources: budget limits are usually hard to assume reliably for
an APT. Following contemporary findings, most reported cases of successfully
implemented APTs were driven by well organized teams with very high, up
to presumably unlimited, resources. In this context, we often see that the
achievement of the goal (i.e., the infiltration of a system) weighs much higher
than the investment made for this purpose. This is due to the fact that many of
these groups are – or are supposed to be – state sponsored [8]. In this sense, the
economic reasoning of security to be given if an attack is more expensive than
the revenues upon success, this may not apply for an APT.
1.3.2 Life-Cycle
In the initial step of an APT, an attacker chooses the target system or network of
his attack. Therefore, the attacker starts to “hunt” for potential targets (e.g., IT
systems or pieces of infrastructure with specific vulnerabilities); alternatively, the
attack can already be tailored to a specific organization or sector. If the target system
is identified, the attacker starts to collect as much information available as possible
about it by using open data sources like websites, social media and others. The aim
is to gain detailed knowledge about the target’s infrastructure, e.g., hardware and
software in use, organizational structure, employees and others.
When this information is available, the attacker starts to craft an attack with the aim
to compromise a weak systems within the organization. Therefore, technical as well
as social skills are used. On the technical side, the attackers are looking for known
vulnerabilities in the identified IT systems by searching respective databases, e.g.,
the National Vulnerability Database (NVD) operated by the National Institute of
Standards and Technology (NIST), or crawling the dark web for zero-day exploits
(i.e., vulnerabilities in software, which are yet unknown, and guidelines how to
exploit them). On the social side, the attackers start to identify personnel at the
targeted organization (e.g., security officers, network administrators but also normal
users) and use social engineering (cf. also Sect. 1.2.2) to obtain detailed information
about them. The aim is to find either vulnerable systems within the organization,
which are accessible from the outside, or personnel which can easily be infiltrated.
Therefore, the attackers create emails containing malicious links or documents (i.e.,
spear phishing mails, cf. also Sect. 1.2.2). These emails are specifically tailored to
the targeted person such that there is a high probability that the person will click on
the link or open the document. In this way, a malware is downloaded or an exploit is
created, which can then be used by the attacker to gain access to the targeted system
(cf. also Sect. 1.2.2).
The malware initially loaded onto the target systems in the previous step usually
creates a backdoor for the attacker to install additional software on the system.
This software has some remote administrative functionality, which allows the
attacker to install additional tools (e.g., command and control software) or establish
communication lines (e.g., Virtual Private Network (VPN) tunnels) to the outside
world. The attacker’s communication with the compromised system is in general
stealthy and hard to spot by an administrator, since it usually blends into the normal
network traffic. This makes it more difficult for the organization’s security officers
to detect the attacker’s presence in the internal network.
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
1.3 Advanced Persistent Threats (APT) 11
After having established the presence in the compromised system, the attacker
tries to identify additional user accounts with higher privileges (e.g., domain
administrators or accounts used for maintenance or service). Most commonly, key
loggers or network sniffing tools are used to achieve that; however, passwords from
other users can also be exfiltrated from respective databases. In this way, the attacker
can escalate the privileges gained by the initial compromise and gain access to
additional systems and domains in the organization’s network.
Besides looking for user accounts with a higher security level on the initially
compromised system, the attacker also scans the network for other important
systems, domains or networks. In this way, the attacker obtains an overview on
the network structure and the systems running in the organization. This information
allows him to identify additional vulnerabilities and load malware to exploit them
via the command and control tools installed in Step 2. Since the attacker is now
operating within the organization’s network, he has additional capabilities to exploit
vulnerabilities of systems, which would not be accessible from the outside world
(i.e., from the general Internet).
If the internal reconnaissance was successful and the attacker was able to identify
other vulnerable infrastructure in the network, he will start compromising those
other systems by performing Steps 2, 3 and 4 on them. In short, the attacker will
exploit the detected vulnerabilities to gain a foothold in the new system, then install
command and control software therein to finally escalate privileges on the newly
compromised systems. As one reason for moving laterally is to gain access to more
important systems in the organization, the risk of being detected by an administrator
or security officer can be another reason for an attacker to move to a different system.
With the attacker obtaining control over an increased number of systems in the
organization’s network, he is able to establish multiple communication channels
to the outside world to maintain his connection to the remote command and control
tools. This leaves him with additional capabilities to operate in the network even if
one of the existing channels is discovered and closed or if his presence is detected
and removed from one system. At this point, it is extremely difficult to completely
remove the presence of the attacker from the organization’s network.
12 1 Introduction
When the attacker has reached its final target, i.e., a specific server or (industrial)
control system, and gained control over it, the last phase starts. In case of an
(industrial) control system, the attacker can gain or already has gained enough
privileges on the device due to the malware and exploits installed in the previous
steps to perform any desired task (e.g., switch off a transformer substation as
happened in the example of the Ukraine described in Sect. 1.4.3). In case the attacker
wants to exfiltrate specific data hosted on a server, he can achieve that via the
communication channels established in the previous steps. Since an organization
usually handles a lot of data going in and out of its network, it is difficult to identify
data exfiltration at that point. The attacker has the opportunity to use various tools
to obfuscate the source or destination of the data stream, to encrypt the payload or
hide the data in a standard data stream.
After completing the mission, the attacker erases all traces of his presence in
the system. However, the attacker might also leave some backdoors, tools and
communication channels operational to be able to return and compromise the
organization again.
In the following, we will present a list of a few selected and well documented attacks
on CIs. We neither claim this list to be complete nor covering the most severe cases
that ever existed, and there might be an undocumented black-count. Nonetheless,
the examples in the next sections shall illustrate how past incidents caused failures
within a CI and furthermore point out the direct consequences of those incidents
as well as on the aforementioned cascading effects. We will loosely distinguish
between natural and man-made causes as well as unintentional and intentional
causes. The examples shall give an idea how vulnerabilities within CIs could look
like and which potential cascading effects a single incident might have.
initial event was a tree flashover at the Swiss 380 kV line “Mettlen-Lavorgo”,
which caused the tripping of this important supply line [30]. With the failure of the
“Mettlen Lavorgo” line, other 380 kV power lines took over the load while several
attempts for re-closing the line were executed. As a result of the balancing attempt,
one of those lines, the “Sils-Soazza” line, which is the closest to the “Mettlen-
Lavorgo” line, also suffered an overload. This overload was acceptable for 15
minutes according to operational standards (and according to expert opinions) for
such an emergency. During that period, several countermeasures (i.e., reducing the
power consumption for about 300 MW) were implemented in Italy and Switzerland
to restore the agreed schedule [30]. However, these measures were not sufficient
and about 25 minutes after the outage of the “Mettlen-Lavorgo” line, also the “Sils-
Soazza” line tripped.
The loss of two important power lines had a strong impact on the remaining
connections to France, Austria and Slovenia, causing them to collapse almost
immediately after the outage of the “Sils-Soazza” line [30]. As a result, the Italian
power grid was isolated from the European grid. This affected the network in
Northern Italy as instability phenomena and overloads happened, resulting in an
unsatisfactory low voltage level in the entire Italian network. Although frequency
control measures were set in place automatically, turbine tripping, underfrequency
relay opening, loss of excitation and other incidents caused the outage of several
generating units, which should cushion the effects from the overload [30]. The result
was a drop of the frequency below the 47.5 Hz threshold and the blackout of the
entire network.
The recovery process started immediately, such that the northern part of the
Italian network was back online about 5 hours after the blackout, the central part
about 7 hours and main Italy about 13 hours after the blackout happened. As the last
part, the island of Sicily was energized about 18 hours after the initial events [30].
In 2005, the entire Swiss railway service was shut down for three hours due to
human misjudgment and maintenance work. Over 200.000 commuters were affected
and got stuck either on the trains or at the train stations; the shutdown caused a
financial damage of over five million Swiss Francs [27].
In the late afternoon of June 22nd 2005, two power lines “Amsteg – Steinen”
in the central Swiss Kanton Uri were switched off due to construction work at the
tracks at that time [13]. This switch off was planned and checked about a week
before the incident and according to existing documentation, the remaining power
line was supposed to handle the additional load without problems. However, the two
lines could not be switched back on after the construction work was completed and
the additional third line could not hold the overload.
It has to be noted at this point that the Swiss railway operates their own power
grid, including power plants and the distribution network. Due to the special
14 1 Introduction
topology of the Swiss railway power grid, the power line “Amsteg – Steinen” is
crucial since it connects three of the main power plants of the Swiss railway operator
with the rest of the network. After the third line tripped due to the overload, the
Kantons Uri and Tessin were cut off the remaining grid, which caused an overload
of the power grid and the shutdown of three power plants in this region [13]. Hence,
a blackout occurred in this region and the railway operator started the recovery
process to bring them back online.
On the remaining side of the network, the power loss was supposed to be
compensated via connections to the power grid of the German railway operator.
Due to the resulting heavy load, those connections failed; at this point, the Swiss
railway operator was also selling electricity to neighboring countries. Since the
operators were mainly occupied with recovering from the first blackout in the
Kantons Uri and Tessin, the problems with the connections to the German power
grid stayed unobserved. Those issues combined caused also a blackout in the
remaining network.
The Ukrainian power grid fell victim to a major cyber attack in 2015, when hackers
managed to gain access to the critical systems of three major power distribution
companies (so called “oblenergos”) in the Ukraine. They caused a large power
outage in most parts of the country, which lasted for about six hours and left
approximately 225.000 households without electricity by switching off about 30
substations [14, 18]. The attack is seen as an APT, since it was very well prepared
by the adversaries as well as highly sophisticated (cf. also Sect. 1.3 for details on
the characteristics of an APT). Further, the 2015 attack is also known the first cyber
attack that directly caused power outages.
The attack applied six adversarial techniques over a time period of several
months (cf. also Fig. 1.2) to prepare and execute the attack. With these techniques,
the attackers followed the main steps of the ICS Cyber Kill Chain [1] (which
are also reflected in the individual steps of an APT as described in Sect. 1.3).
The attack started with an analysis of open source information about the three
power distributors followed by a serious of spear phishing attacks on them. Those
attacks were particularly targeted at the power providers by using information from
openly available sources (e.g., on Remote Terminal Unit (RTU) and ICS vendors
used in these companies) and social media (e.g., on senior personnel working in
these companies) [18]. The targets received crafted emails with word documents
including an exploit to install the BlackEnergy 3 malware [29] on the system.
This malware provides several tools to an attacker, starting from simple network
scanning up to key logging and remote desktop functionality, which enable a hostile
takeover of the infected system and the potential to infiltrate the network the system
is connected to.
1.4 Selected Real-Life Incidents 15
spear phishing
tools credential
& tech theft
control
VPN access
& operate
workstation remote
Fig. 1.2 Illustration of the steps during the hack of the Ukranian power grid [18]
Sutcliffe, W. Andante in E.
Tartini, G. Larghetto (G minor). (Jensen, V. 6).
Thomas, Emile. Air de Ballet.
— Danse Lente.
— Sanssouci Valse.
— Polka.
— Danse rustique.
Light salon pieces in simple rhythms.
— Sonatine en Ut (C).
Tschaïkowsky, P. Album of Favourite Pieces:—
1. Mazurka; 2. Sweet Dream; 3. Neapolitan Song; 4.
Autumn Song; 5. Barcarolle; 6. Andante cantabile; 7.
Chant sans paroles; 8. Troïka; 9. Christmas. Edited and
partly arranged by Fr. Hermann. (7590).
Nos. 1, 6, 7, 8 and 9 belong to II.
— Chanson triste. (M.F. 85).
— Chant sans Paroles. (M.F. 27).
Valle de Paz, E. del. Op. 29, No. 2. Minuet.
Veracini, A. Sonata in A minor. (G. Jensen.) (7416).
Veracini, F. M. Minuet. (Jensen, V. 5).
— Sarabanda e Danza rustica. (Moffat.) (7589).
Simonetti, A. Mazurka.
Sinding. Op. 10. Suite. (2477).
— Op. 27. Sonata in E major. (2826).
— Op. 30. Romance. (2827).
— Op. 61. 4 Pieces. 3 Bks. (3050 a-c).
Sinding’s compositions rank among the finest works by
modern Scandinavian composers. The Suite Op. 10,
resembles in style the “Holberg” Suite by Grieg.