100% found this document useful (3 votes)

969 views

Internet-of-Things (IoT) System Architectures, Algorithms, Methodologies PDF

Uploaded by

Diana Jess

Available Formats

Download as PDF, TXT or read online on Scribd

100% found this document useful (3 votes)

969 views

Internet-of-Things (IoT) System Architectures, Algorithms, Methodologies PDF

Uploaded by

Diana Jess

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 102

Dimitrios Serpanos

Marilyn Wolf

Internet-of-Things
(IoT) Systems
Architectures, Algorithms,
Methodologies
Internet-of-Things (IoT) Systems
Dimitrios Serpanos • Marilyn Wolf

Internet-of-Things (IoT)
Systems
Architectures, Algorithms, Methodologies
Dimitrios Serpanos Marilyn Wolf
Electrical & Computer Engineering School of ECE
University of Patras Georgia Institute of Technology
Patras, Greece Atlanta, GA, USA

ISBN 978-3-319-69714-7 ISBN 978-3-319-69715-4 (eBook)

https://doi.org/10.1007/978-3-319-69715-4

Library of Congress Control Number: 2017956194

© Springer International Publishing AG 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims
in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Dimitrios Serpanos:
To Loukia and Georgia for their patience
and understanding…
Marilyn Wolf:
To Alec.
Preface

The Internet of Things is the evolutionary step of the Internet that creates a world-
wide infrastructure interconnecting machines and humans. As the Internet became
public in the early 1990s, the first wave of its exploitation and deployment was
mainly focused on the impact to everyday services and applications that changed
the known models for financial transactions, shopping, news feeding and informa-
tion sharing. It was a revolution that digitized a wide range of services as we knew
them, from banking and retail shopping to face-to-face communication and govern-
ment services. The first two decades of the Internet revolution focused strongly on
consumer services and businesses, but human-centric. New business models
appeared for banking, for online shopping, video communication, etc. for consum-
ers. Business to business models and the cloud have impacted businesses signifi-
cantly, wiping out large sectors of industry that did not adjust to the fast pace of the
revolution. The impact on the economies has been tremendous. Now, more than two
decades later, we witness and experience a new way of life because of the Internet’s
reach to our homes and work environments.
The advances of communication technology that enables the deployment and
success of the Internet at home and work had an additional effect: the development
of sophisticated interconnections among machines in the operational environment;
we contrast the operational technology (OT) environment, which controls physical
machines, to the information technology (IT) environment where humans are using
computers for work. The already automated industrial environment received well
the emerging technologies, adopted the suitable ones and created a, private mostly,
network infrastructure that enables highly productive industrial processes. It has
only been a natural step to evolve the Internet itself to include these processes.
Additionally, the control models of the industrial environment, taking advantage of
the smart devices –i.e. devices that include processing, memory and networking
resources- that are deployed in various environments, have been extended and used
in a wide variety of application domains. Conventional application domains like
transportation, aeronautics, energy production and distribution, manufacturing and
health adopt similar control models, exploiting smart sensors, actuators and devices
that enable control automation for sophisticated applications. Critical infrastructure

vii
viii Preface

of countries is run using these technologies today. This emerging Internet-of-Things

(IoT) is the natural evolutionary step of the Internet revolution that started about
three decades ago. Importantly, IoT is building a worldwide infrastructure that will
influence all facets of our life, from agriculture to mining, from health services to
manufacturing and transportation. Clearly, it will provide the infrastructure over
which the new emerging AI revolution will be based.
This book addresses the fundamental IoT technologies, architectures, applica-
tion domains and directions. Development of a complete IoT system and service
includes several components. The hardware base includes embedded processors,
memories of different types, sensors, actuators, cloud servers, intermediate process-
ing systems, network systems and gateways. The software base includes operating
systems, data bases and control applications for several application domains, to the
very least. The combination of hardware and software components for control
applications constitutes the base for the evolution of cyber-physical systems. VLSI
capabilities play a huge role in the design of IoT systems. Event-driven, distributed
operation shapes the design of architectures and applications. Specialized network
protocols enable efficient communication in this environment, including appropri-
ate machine-to-machine (M2M) communication models. These technologies are
emerging with constraints and restrictions for the IoT environment that are different
from the typical IT environment, because of the requirements for safety, real-time
responses, low power operation, etc. Security, privacy, and safety require particular
attention and special techniques.
IoT is a fast-changing field. This book provides a snapshot of its current state. We
continue to work in this area and hope to create updates to this book as the field
progresses.

Atlanta, GA, USA Marilyn Wolf

Patras, Greece Dimitrios Serpanos
Acknowledgements

We would like to acknowledge M.T. Khan, K. Katsigiannis and C. Koulamas for

their support and their review and comments on the drafts.

ix
Contents

1 The IoT Landscape�� 1

1.1 What Is IoT? �� 1
1.2 Applications�� 2
1.3 Architectures �� 4
1.4 Wireless Networks�� 4
1.5 Devices�� 4
1.6 Security and Privacy �� 5
1.7 Event-Driven Systems�� 5
1.8 This Book�� 6
Reference �� 6
2 IoT System Architectures�� 7
2.1 Introduction�� 7
2.2 Protocols Concepts�� 7
2.3 IoT-Oriented Protocols�� 10
2.4 Databases�� 12
2.5 Time Bases�� 13
2.6 Security �� 13
References�� 14
3 IoT Devices �� 17
3.1 The IoT Device Design Space�� 17
3.2 Cost of Ownership and Power Consumption�� 18
3.3 Cost per Transistor and Chip Size�� 19
3.4 Duty Cycle and Power Consumption�� 20
3.5 Platform Design�� 22
3.6 Summary �� 22
References�� 22
4 Event-Driven System Analysis�� 25
4.1 Introduction�� 25
4.2 Previous Work �� 26
4.3 Motivating Example�� 27

xi
xii Contents

Index�� 91
Chapter 1
The IoT Landscape

1.1 What Is IoT?

The Internet of Things (IoT) has become a common news item and marketing trend.
Beyond the hype, IoT has emerged as an important technology with applications in
many fields. IoT has roots in several earlier technologies: pervasive information
systems, sensor networks, and embedded computing. The term IoT system more
accurately describes the use of this technology than does Internet of Things. Most
IoT devices are connected together to form purpose-specific systems; they are less
frequently used as general-access devices on a worldwide network.
IoT moves beyond pervasive computing and information systems, which con-
centrated on data. Smart refrigerators are one example of pervasive computing
devices. Several products included built-in PCs and allowed users to enter informa-
tion about the contents of their refrigerator for menu planning. Conceptual devices
would automatically scan the refrigerator contents to take care of data entry. The use
cases envisioned for these refrigerators are not so far removed from menu planning
applications for stand-alone personal computers.
Sensor network research spanned a range of configurations. Many of these were
designed for data collection at very low data rates. The collected data would then be
sent to servers for processing. Traditional sensor network research did not empha-
size in-network processing.
Embedded computing concentrated on either stand-alone devices or tightly cou-
pled networks such as those used in vehicles. Consumer electronics and cyber-
physical systems were two major application domains for embedded computing;
both emphasized engineered systems with well-defined goals.
Given the wide range of advocates for IoT technology, no single, clear definition
of the term has emerged. We can identify several possibilities:
• Internet-enabled physical devices, although many devices don’t use the Internet
Protocol

© Springer International Publishing AG 2018 1

D. Serpanos, M. Wolf, Internet-of-Things (IoT) Systems,
https://doi.org/10.1007/978-3-319-69715-4_1
2 1 The IoT Landscape

• Soft real-time sensor networks

• Dynamic and evolving networks of embedded computing devices
This book is primarily interested in IoT systems. We use this term to capture two
characteristics. First, the system is designed for one or a set of applications, rather
than being an agglomeration of Internet-enabled devices. Second, the IoT system
takes into account the dynamics of physical systems. An IoT system may consist
primarily of sensors; in some cases it may include a significant number of actuators.
In both cases, the goal is to process signals and time-series data.
Interest in the Internet of Things has been spurred by the availability of micro-
electromechanical (MEMS) sensors. Integrated accelerometers, gyroscopes, chemi-
cal sensors, and other forms of sensor are now widely available. The low cost and
power consumption of these sensors enables new applications well beyond those of
traditional laboratory or industrial measurement equipment. These sensor applica-
tions push IoT systems toward signal processing.
IoT is also enabled by the very low cost of VLSI digital and analog electronics.
As we will see, IoT nodes do not rely on state-of-the-art VLSI manufacturing pro-
cesses. In fact, they are inexpensive because they are able to make use of older
manufacturing lines; the lower device counts available in these older technologies
are more than sufficient for many IoT systems.
IoT systems must consume very little power. Power consumption is a key factor
in total cost of ownership for IoT systems. Achieving the necessary power levels
requires careful attention to hardware design, software design, and application
algorithms.
Security and safety are key design and operational requirements for IoT systems.
As we have argued elsewhere, safety and security are no longer separable problems.
The merger of computational and physical systems requires us to merge the previ-
ously separate tasks of safe physical system design and secure computer system
design.

1.2 Applications

IoT systems are useful in a broad range of applications:

• Industrial systems use sensors to monitor both the industrial processes them-
selves – the quality of the product – and the state of the equipment. An increasing
number of electric motors, for example, include sensors that collect data used to
predict impending motor failures.
• Smart buildings use sensors to identify the locations of people as well as the state
of the building. That data can be used to control heating/ventilation/air condi-
tioning systems and lighting systems to reduce operating costs. Smart buildings
and structures also use sensors to monitor structural health.
• Smart cities use sensors to monitor pedestrian and vehicular traffic and may inte-
grate data from smart buildings.
1.2 Applications 3

• Vehicles use networked sensors to monitor the state of the vehicle and provide
improved dynamics, reduced fuel consumption, and lower emissions.
• Medical systems connect a wide range of patient monitoring sensors that may be
located at the home, in emergency vehicles, the doctor’s office, or the hospital.
Use cases help us understand the requirements on an IoT system.
Sensor network The system may act strictly as a data gathering system for a set of
sensors.
Alert system Data from sensors may be gathered and analyzed. Alerts are gener-
ated when particular criteria are met.
Analysis system Data from sensors is gathered and analyzed, but in this case, the
analysis is ongoing. Reports on analytic results may be generated periodically –
hourly, daily, etc. – or may be continuously updated.
Reactive system Analysis of sensor data may cause actuators to be triggered. We
reserve the term reactive for systems that don’t implement typical control laws.
Control system Sensor data is fed to control algorithms that generate outputs for
actuators.
We can identify a class of nonfunctional requirements that apply to many IoT
systems. Nonfunctional requirements on the system impose nonfunctional require-
ments on the components.
Event latency Latency from capture of an event to its destination may not be
important for batch-oriented applications but becomes important for online
analysis.
Event throughput The rate at which events can be captured, transported, and pro-
cessed depends on the throughput of the nodes, network bandwidth, and cloud
throughput.
Event loss rate and buffer capacity In the absence of strict upper bounds on event
production rates, the environment may produce more events in an interval than the
system can produce. Event loss rate captures the desired capability, while buffer
capacity is a more pragmatic requirement that can be directly tied to component
capabilities.
Service latency and throughput Ultimately, events will be processed by services.
We can also specify the latency and throughput for services.
Reliability and availability Since IoT systems are distributed, reliability is more
likely to be specified over parts of the network rather than reliability of the complete
system. Availability is commonly used to describe distributed systems.
Service lifetime IoT systems are often expected to have longer lifetimes than we
expect for PC systems. The lifetime of the system or a subset of the system may be
considerably longer than that of a component, particularly if the system uses redun-
dant sensors and other components.
4 1 The IoT Landscape

1.3 Architectures

A key aspect of IoT is event-driven or aperiodic sampling. Traditional digital signal

processing and control assume periodic samples resulting in time-series data.
However, time series consume too much power at the nodes and too much band-
width on the network. Not all applications are amenable to aperiodic data
acquisition.
Constraints on power and bandwidth also encourage distributed computing over
sensor events. Relatively small processors can perform useful processing on many
data streams. Recognizing interesting events using edge processing reduces the
amount of network bandwidth consumed; it also reduces power consumption since
wireless communication requires large amounts of power. Cloud computing-
(centralized servers) or fog computing (servers closer to the edge) can be used to
perform further processing on those extracted events.

1.4 Wireless Networks

Wireless networks are integral to IoT systems. Wireless network connections sim-
plify installation and operation of wireless networks.
However, wireless networks introduce some important problems and restric-
tions. Radio communication requires more power than does wired communication.
Some of the wireless networks used in today’s IoT devices were designed for other
purposes, such as telephony and multimedia. As a result, they are not optimized for
event-driven communication and consume significant amounts of power in the com-
munications protocol.
One of the ironies of IoT is that many edge devices and their wireless networks
don’t operate on the Internet Protocol (IP). IP introduces significant overhead with
an extra level of packetization and associated processing. Many IoT devices avoid
IP and rely on upstream nodes to provide them with an Internet presence.
IoT networks are typically run by noncomputer experts. IoT wireless networks
must be easy to deploy and relatively self-managing.

1.5 Devices

The characteristics of event-driven systems allow IoT nodes to be relatively simple.

The realities of low-power operation also push nodes toward relatively low levels of
integration.
VLSI technology and Moore’s law are key factors in the rise of IoT systems
because they allow nodes to be manufactured extremely cheaply. Very small chips
can provide enough computation, memory, and networking for useful IoT node
1.7 Event-Driven Systems 5

functions. In contrast to traditional microprocessor and consumer electronic appli-

cations, where chip areas range around or even higher, chips of several square mil-
limeters are large enough for many IoT node devices.

1.6 Security and Privacy

Security has finally been recognized as an essential requirement for all types of
computer systems, including IoT systems. However, many IoT systems are much
less secure than typical Windows/Mac/Linux systems. IoT security problems stem
from a range of causes: inadequate security features in hardware, poorly designed
software with a range of vulnerabilities, default passwords, and other security
design errors.
Insecure IoT nodes create problems for the security of the entire IoT system.
Because nodes typically have lifetimes of several years, the large installed base of
insecure devices will create security problems for some time to come.
Insecure IoT systems also cause security problems for the rest of the Internet.
IoT devices are plentiful; insecure IoT nodes are ideally suited to denial-of-service
attacks. The Dyn attack [Sch16] is one example of an IoT-based attack on traditional
Internet infrastructure.
Privacy is related to security but requires specific measures at the application,
network, and device levels. Not only must user data be protected from outright theft,
but the network needs to be designed so that less-private data cannot easily be used
to infer more private data.

1.7 Event-Driven Systems

We believe that the event is a fundamental data type in IoT systems and that event-
driven systems are an important structuring technique for IoT. Many of the building
block technologies used for IoT today show some holdover from traditional,
transaction-oriented systems. Event processing pushes us to treat time as a first-
class concept and to consider the relationship between events in event sequences.
We use the term event more broadly then do simulation engineers. We consider
events as time-value sets. Event-driven system simulation is widely used for model-
ing a wide range of engineering systems. In that context, an event is generally used
to mean a change in the state of a variable. Given the decentralized nature of IoT
systems, we are willing to consider stuttering – the repetition of an event value – as
part of the event model. We also use events to model sampled data and time-series
data. We believe that all these uses of the term event can be unified to create rich
system structures.
6 1 The IoT Landscape

1.8 This Book

The rest of this book describes a range of topics in IoT systems in more detail:
• Chapter 2 studies IoT system architectures, including wireless networks.
• Chapter 3 considers VLSI IoT devices. It describes the relationship between cost
of ownership, power consumption, and duty cycle.
• Chapter 4 introduces analysis methods for event-driven IoT systems. These anal-
ysis methods allow us to study the memory requirements implied by event com-
munication and processing.
• Chapter 5 describes the Industrial Internet of Things and applications of IoT
systems in smart energy systems.
• Chapter 6 studies security and safety issues in IoT systems. Computer and cyber-
physical system security is closely tied to safety in sensor and closed-loop con-
trol systems.
• Chapter 7 describes fuzz testing, a technique for testing the security of IoT sys-
tems. Bugs and crashes can provide exploits for attackers; fuzz testing is designed
to help identify such problems.

Reference

[Sch16] Schneier, B. (2016, October 22). DDoS attacks against Dyn. Schneier on Security.
https://www.schneier.com/blog/archives/2016/10/ddos_attacks_ag.html
Chapter 2
IoT System Architectures

2.1 Introduction

In this chapter, we study architectures for IoT systems. We will study typical com-
ponents used for networks, databases, etc.
Figure 2.1 shows the organization of an IoT system:
• The plant or environment is the physical system with which the IoT system inter-
acts. We will use these two terms interchangeably.
• A set of devices form the leaves of the network. A node may include sensors and/
or actuators, processors, and memory. Each node has a network interface. A node
may or may not run the Internet Protocol.
• Hubs provide first-level connectivity between the nodes and the rest of the net-
work. Hubs are typically run IP.
• Fog processors perform operations on local sets of nodes and hubs. Keeping
some servers nearer the nodes reduces latency. However, fog devices may not
have as much compute power as cloud servers. Fog devices also introduce sys-
tem management issues.
• Cloud servers provide computational services for the IoT system. Databases
store data and computational results. The cloud may provide a variety of services
that mediate between nodes and users.

2.2 Protocols Concepts

Several protocols are used for data services in IoT systems.

Communication protocols may not provide sufficient abstraction for many appli-
cations. IoT systems need multi-hop, end-to-end communication. They also may
exhibit complex relationships between data sources and sinks. Higher-level proto-
cols can provide services that model more closely the needs of IoT systems. Given

© Springer International Publishing AG 2018 7

D. Serpanos, M. Wolf, Internet-of-Things (IoT) Systems,
https://doi.org/10.1007/978-3-319-69715-4_2
8 2 IoT System Architectures

Fig. 2.1 Organization of an IoT system

Fig. 2.2 The publish/subscribe model

the heterogeneous and long-lived nature of most IoT systems, standards are often
used rather than custom protocols. Several different protocols have been proposed
and, to varying degrees, used for IoT systems [Duf13]. The user space has not yet
converged on a single standard for IoT communication services.
Given the prevalence of event-oriented models in IoT systems, a protocol should
support event-style communication.
The HTTP protocol uses a request/response design pattern. A client issues a
request for a hypertext object; the server then replies with the object in response.
A publish/subscribe protocol [Twi11] requires less coupling between the client
and server as illustrated in Fig. 2.2. The server, known as a publisher, classifies mes-
sages into categories. Clients subscribe to the categories of interest to them. Publish/
subscribe systems are typically mediated by brokers which receive published
2.2 Protocols Concepts 9

essages from publishers and send them to subscribers. Messages may be orga-
m
nized by topic; all message of a given topic are distributed by the brokers to the
subscribers for that topic. The broker knows the identities of subscribers but the
publisher does not. Brokers may interact with each other using a bridge protocol. A
bridge allows indirect publication of messages, with a message going from the pub-
lisher to a first broker, then to a second broker, and finally to subscribers who are not
connected to the first broker.
Data Distribution Service (DDS) (http://portals.omg.org/dds/) [Obj16] is a pub-
lish/subscribe software architecture; several implementations of DDS are in use. A
DDS domain maintains a logical global data space; the data is managed over a set
of local stores. Publishers and subscribers are dynamically discovered across the
network. Publishers can specify a number of quality of service parameters that are
enforced by the brokers.
Real-Time Publish/Subscribe Protocol (RTPS) [Obj14] is a so-called wire proto-
col that defines a protocol for communication with DDS and other publish/sub-
scribe systems. RTPS provides QoS properties, fault tolerance, and type safety.
Esposito et al. [Esp09] developed an architecture for time-sensitive publish/sub-
scribe systems that would be scalable to Internet-sized systems. They identified
three major design goals: predictable latency, guaranteed delivery in the presence of
multiple faults, and continued performance under scaling. They identified several
types of fault models for publish/subscribe systems: network anomalies (loss, order-
ing, corruption, delay, congestion, partitioning), link crash, node crash, and churn of
nodes unexpectedly joining and leaving the system. Their architecture has three
abstraction layers: the network layer consists of domains composed of nodes; the
nodes layer consists of clusters, with each cluster’s members belonging to the same
stub domain; and a coordinators layer. The coordination layer routes messages
using a tree-based topology built on top of a distributed hash table. The coordinator
is p-redundant to provide fault-tolerant coordination. To provide fault-tolerant over-
lays, they formulate a model for path diversity that can be computed with limited
knowledge of the network connections.
Kang et al. [Kan12] used a semantics-aware communication mechanism to
reduce overhead and improve reliability. They use state-space estimators at both the
publisher and subscriber to maintain continuity of sensor values in the presence of
network variations. Their state estimator is of the form xk + 1 = Fk + 1xk. The designer
sets a model precision bound δ for each sensor. The bound is used to manage band-
width requirements. Their system also dynamically adjusts the model precision
bound.
Choi et al. [Choi16] combined DDS with the OpenFlow software-defined net-
working protocol to ensure that DDS can implement the QoS parameters. They
added two QoS parameters that could not be easily deduced from the standard DDS
parameters: MINIMUM_SEPARATION and an E2E_LATENCY specified by
subscribers.
10 2 IoT System Architectures

2.3 IoT-Oriented Protocols

We can divide protocols into two major categories: those that are tied to a specific
physical layer and those that are not. Generally speaking, protocols that rely on a
specific physical layer do not use the Internet Protocol, while protocols that are
physical layer agnostic do use IP.
Zigbee [Zig14, Far08] is a mesh network designed for low-power operation. A
variety of derivative application standards specialize the protocol for applications
such as smart homes and utilities. Zigbee is based on the IEEE 802.15.4 PHY and
MAC standards. 802.15.4 operates in three bands: 868 MHz, 915 MHz, and
2.4 GHz. It delivers bit rates from 20 to 250 kbps, depending on the frequency band.
The Zigbee NWK layer sits on top of the 802.15.4 MAC layer and provides data and
management services. The APL layer includes three sections: the application sup-
port sublayer, the Zigbee Device Objects layer, and the application framework.
Zigbee provides two types of network security models: a centralized security
network can be started only by a Zigbee coordinator/trust center; distributed secu-
rity networks do not have a central trust center. Nodes can join either type of net-
work and adapt to the type of network they have joined. Networks are formed by
either coordinators or routers after scanning to select an available channel.
Coordinators form centralized security networks, while routers form distributed
security networks. Network steering is the name for the process by which a node
joins a network. After identifying an open network, the node associates with that
network and receives a network key. Clusters define interfaces for features and
domains.
Bluetooth Low Energy (BLE) (https://www.bluetooth.com/what-is-bluetooth-
technology/how-it-works/low-energy) [Hay13] is a part of the Bluetooth standard
designed for low-power operation such as devices powered from coin cell batteries.
A BLE device can work as a transmitter, receiver, or both. Figure 2.3 illustrates the
Bluetooth Classic protocol stack.
The link layer provides an advertising service; devices can scan to identify nodes
and networks. Devices can act as gateways to the Internet based on network address
translation. The BLE protocol is stateful. BLE includes a number of optimizations
to reduce power consumption.
LoRa (http://lora-alliance.org) [LoR15] is designed for wide-area IoT applica-
tions with a base station covering hundreds of square kilometers. It is designed to
support a network topology with gateways for end devices, with gateways organized
into their own star network. Data rates range from 0.3 to 50 kbps.
MQTT (http://www.mqtt.org) [IBM12, Oas14] is an IoT-oriented protocol with
publish/subscribe semantics. The protocol is designed for low overhead and is
agnostic to the data payload. MQTT provides three levels of quality of service: at
most once provides best-effort service, at least once assures delivery but may incur
duplicates, and exactly once ensures the message is delivered without duplication.
2.3 IoT-Oriented Protocols 11

Fig. 2.3 The Bluetooth

stack

MQTT is based on a publish/subscribe model. A message is given a retention

attribute when it is published; messages with QoS designations of at least once or
exactly once should set the retention flag. A new subscriber to the topic will receive
the last publication on that topic.
When setting up a connection, a client can provide a will to the server to specify
a message to be published if the client is unexpectedly disconnected.
Messages are classified using topic strings similar to hierarchical file names. The
set of topics is organized into a topic tree. Topic names follow the names of the
nodes in the topic tree path, with node names separated by “/”. Subscribers can use
wildcards in the topic string: ‘+’ denotes a wildcard match at one level of the topic
tree; “#” denotes a match at any number of levels of the topic tree.
XMPP (http://xmpp.org) is a protocol for streaming XML. It provides security,
authentication, and information about network availability, and rosters of clients.
XMPP-IoT (http://xmpp-iot.org) is a dialect of XMPP designed for IoT
applications.
REST [Vaq14, Rod15] is widely used for Web services and has received some
use as an IoT service model. REST is a design pattern for stateless HTTP transfers.
It exposes directory-structured form resource indicators. REST can be used to trans-
fer XML or JSON data. Clients access resources using GET, PUT, POST, and
DELETE methods.
CoAP (http://coap.technology) [IET14] is a REST-based Web transfer protocol
designed for IoT devices. It can be used with several types of data payloads, includ-
ing XML and JSON.
Google Cloud Pub/Sub [Goo17A, Goo17B] can be used to provide publish/sub-
scribe service to IoT and other systems. Topics and subscriptions are exposed as
12 2 IoT System Architectures

REST collections. The system is divided into a data plane for messages and a con-
trol plane for allocation to servers known as routers; data plane servers are known
as forwarders. The routers balance consistency and uniformity of data using a con-
sistent hashing algorithm. A message life cycle includes several steps. When a pub-
lisher sends a message, it is written to storage. The subscribers receive the message,
and the publisher receives an acknowledgment. Subscribers acknowledge the mes-
sage to Google Cloud Pub/Sub. The message is deleted from storage once at least
one subscriber for each subscription has acknowledged the message. The system
monitors itself to detect and mitigate service problems.
Amazon Web Services (AWS) IoT [Bar15] is a managed cloud service for IoT
devices, which are termed things. A thing shadow is a cloud model of a thing. A rule
engine transforms messages based on rules and routes the results to AWS services.
The message broker is based on MQTT. A Thing Registry assigns unique identity to
things.
Microsoft Azure (https://azure.microsoft.com/en-us/services/iot-hub/) provides
IoT-oriented services. Its Service Fabric is a middleware communication system
that supports microservices running on a cluster. A microservice may be either
stateless or stateful. It also provides a container model for applications; a container
provides an isolated environment but relies on the operating system, in contrast to a
virtual machine which runs underneath the operating system. It provides databases
using both structured and unstructured approaches. It also provides APIs for artifi-
cial intelligence services.

2.4 Databases

Databases are used for both short-term and long-term storage. Applications may
rely on databases to retrieve data over a time window for analysis. Some use cases
may require archival storage of values.
Unstructured databases, known as noSQL, are used in many IoT systems. A
noSQL database does not have a schema. Simple noSQL databases represent data
as key-value pairs, but other representations are possible. The lack of a schema
allows quick deployment but may cause maintenance problems.
The Amazon Simple Storage Service (Amazon S3) (https://aws.amazon.com/
s3/) is an object store with a Web service interface. Data can be pushed to other,
lower-cost storage services for long-term, infrequent use. Notifications can be
issued when objects operated upon.
Google Cloud Storage (https://cloud.google.com/storage) is an object store for
unstructured data. It provides three different service models at different latency/
latency/price points. Cloud SQL can be used to perform database operations.
Streaming transfers are supported using HTTP chunked transfer encoding.
Time-series data possesses structure that may require special handling to provide
proper database performance. Time series are sometimes stored as blobs in rela-
tional databases to allow specialized algorithms.
2.6 Security 13

Dynamic time warping (DTW) [Rat04, Rak12] is widely used to search over
time-series data. DTW was originally used to compare waveforms for speech pro-
cessing. Correlation provides a direct comparison of two waveforms. By warping
one waveform, non-exact matches can be found. Dynamic programming can be
used to find the minimum warp match between two-time series; a limit on maxi-
mum warping is typically applied to avoid obviously bad matches. Very efficient
algorithms have been developed to provide high-speed search. Among other tech-
niques, these algorithms abandon a warp computation early when partial results
exceed a given bound. Fast DTW algorithms have been used to search very large
databases.

2.5 Time Bases

Many IoT systems require a notion of global time. Several algorithms, starting with
Lamport’s algorithm [Lam78], have been developed for the synchronization of
clocks in a distributed system.
The Network Time Protocol (RFC1305) is used on the Internet for distributed
time synchronization.

2.6 Security

Security is a system property; the system can be only as secure as its weakest com-
ponent. Security features are provided by components at several layers in the IoT
stack: devices, physical networks, and middleware. A unified view of IoT system
security architectures has not yet emerged.
Some, but not all processors for low-power operation, provide security features
such as encryption accelerators and root of trust. The National Security Agency has
developed families of lightweight block ciphers [Sch13]: SIMON targets hardware
implementations, and SPECK is intended for software implementations. Gulcan
et al. [Gul14] developed a low-power implementation of SIMON.
Several networks provide security features. Bluetooth Low Energy provides a
Simple Secure Pairing protocol to protect against passive eavesdropping. It also
provides address randomization. As discussed above, Zigbee provides two network
security models: centralized and distributed. LoRa provides unique network keys,
unique application keys, and device-specific keys.
MQTT does not specifically require encryption, but it can be used with several
different security standards. MQTT and the NIST Framework for Improving Critical
Infrastructure Cybersecurity [Oas14B] describe the relationship between MQTT
and the NIST Cybersecurity Framework.
We will study IoT system security in more detail in Chap. 6.
14 2 IoT System Architectures

References

[Bar15] Barr, J. (2015, October 8). AWS IoT: Cloud services for connected devices. AWS Blog.
https://aws.amazon.com/blogs/aws/aws-iot-cloud-services-for-connected-devices/
[Choi16] Choi, H.-Y., King, A. L., & Lee, I. (2016). Making DDS really real-time with OpenFlow.
2016 international conference on embedded software (EMSOFT) (pp. 1–10). Pittsburgh, PA.
[Duf13] Duffy, P. (2013, April 30) Beyond MQTT: A Cisco view on IoT protocols. Cisco Blogs.
https://blogs.cisco.com/digital/beyond-mqtt-a-cisco-view-on-iot-protocols
[Esp09] Esposito, C., Cotroneo, D., & Gokhale, A.. 2009. Reliable publish/subscribe middle-
ware for time-sensitive internet-scale applications. Proceedings of the third ACM international
conference on distributed event-based systems (DEBS’09). ACM, New York, Article 16, 12
pages.
[Far08] Farahani, S. (2008). Zigbee wireless networks and transceivers. Amsterdam: Newnes.
[Goo17A] Google. (2017, April 19). What is Google Cloud Pub/Sub? https://cloud.google.com/
pubsub/docs/overview
[Goo17B] Google. (2017, April 3). Google Cloud Pub/Sub: A Google-scale messaging service.
https://cloud.google.com/pubsub/architecture
[Gul14] Gulcan, E., Aysu, A., & Schaumont, P. (2015). A flexible and compact hardware archi-
tecture for the SIMON block cipher. In T. Eisenbarth & E. Öztürk (Eds.), Lightweight cryptog-
raphy for security and privacy. LightSec 2014, Lecture Notes in Computer Science (Vol. 8898,
pp. 34–50). Cham: Springer.
[Hay13] Heydon, R. (2013). Bluetooth low energy: The developer’s handbook. Prentice Hall:
Upper Saddle River, NJ.
[IBM12] IBM International Technical Support Organization (2012, September). Building smarter
planet solutions with MQTT and IBM WebSphere MQ telemetry, Redbooks.
[IET14] Internet Engineering Task Force (2014, June). The constrained application protocol
(CoAP), RFC 7252, Shelby, Z., Hartke, K., & Bormann, C.
[Kan12] Kang, W., Kapitanova, K., & Son, S. H. (2012). RDDS: A real-time data distribu-
tion service for cyber-physical systems. IEEE Transactions on Industrial Informatics, 8(2),
393–405.
[Lam78] Lamport, L. (1978). Time, clocks, and the ordering of events in a distributed system.
Communications of the ACM, 21(7), 558–565.
[LoR15] LoRa Alliance (2015, November). LoRaWAN: What is it? A technical overview of LoRa
and LoRaWAN.
[Oas14] Oasis. (2014, 29). MQTT version 3.1.1. Oasis standard.
[Oas14B] Oasis (2014, May 28). MQTT and the NISTG cybersecurity framework version 1.0.
Committee note 01.
[Obj14] Object Management Group. (2014). The real-time publish-subscribe protocol (RTPS)
DDS interoperability wire protocol specification, Version 2.2.
[Obj16] Object Management Group. (2016). What is DDS? http://portals.omg.org/dds/what-is-
dds-3/, accessed May 4, 2017.
[Sch13] Schneier, B. SIMON and SPECK: New NSA encryption algorithms. Schneier on
Security. https://www.schneier.com/blog/archives/2013/07/simon_and_speck.html, retrieved
May 8, 2017.
[Vaq14] Vaqqas, M. (2014, September 23) RESTful web services: A tutorial. Dr. Dobb’s. http://
www.drdobbs.com/web-development/restful-web-services-a-tutorial/240169069
[Rak12] Rakthanmanon, T., Campana, B., Mueen, A., Batista, G., Westover, B., Zhu, Q.,

Zakaria, J., & Keogh, E.. 2012. Searching and mining trillions of time series subsequences
under dynamic time warping. Proceedings of the 18th ACM SIGKDD international conference
on knowledge discovery and data mining (KDD’12) (pp. 262–270). ACM, New York.
[Rat04] Ratanamahatana, C. A., & Keogh, E. (2004, August 22–25). Everything you know about
dynamic time warping is wrong. Third workshop on mining temporal and sequential data, in
References 15

conjunction with the tenth ACM SIGKDD international conference on knowledge discovery
and data mining (KDD-2004). Seattle, WA.
[Rod15] Rodriguez, Alex. (2008, November 6). RESTful web services: The basics. IBM devel-
operWorks, updated February 9, 2015. https://www.ibm.com/developerworks/library/ws-rest-
ful/index.html
[Twi11] Twin Oaks Computing, Inc. (2011). What can DDS do for you?
[Zig14] Zigbee Alliance (2014, December 2). ZigBee 3.0: The open, global standard for the
Internet of Things. http://www.zigbee.org/zigbee-for-developers/zigbee/
Chapter 3
IoT Devices

3.1 The IoT Device Design Space

The design space for IoT devices is very different from that for mobile or cloud
processors. Both mobile and cloud systems require very large chips. IoT devices
should operate at extremely low power levels but often not operate continuously.
They must integrate processors, memory and storage, communication, and sensors.
They will also be sold in quantities that dwarf even those of mobile processors,
which in turn require a very low price. Purchase price is, however, only one compo-
nent of the IoT device cost model. Total cost of ownership will drive many IoT
markets – these devices will be installed for use over a lifetime of several years.
Installation cost is an important element in the decision to purchase and install these
devices. We will see that cost of ownership is directly tied to power consumption.
The sensors and MEMS communities have long been interested in IoT as an
application for integrated sensors and actuators. Many commentators have called
for a trillion sensor world. This goal is in fact very realistic given current industry
capabilities. According to Semi.org [Die16], worldwide manufacturing capacity for
200 mm wafers is expected to be 5.4 million wafers per month in 2018. If all this
capacity is used for IoT, it translates to 678 billion chips per month of size 1 mm2 or
68 billion per month of 10 mm2 chips. That capacity puts the industry within range
of producing a trillion sensors per year. We could reach the trillion sensors per year
mark simply by reallocating existing capacity. Even if production does not com-
pletely reach the trillion sensor mark, the industry can clearly manufacture huge
volumes of sensors.

© Springer International Publishing AG 2018 17

D. Serpanos, M. Wolf, Internet-of-Things (IoT) Systems,
https://doi.org/10.1007/978-3-319-69715-4_3
18 3 IoT Devices

3.2 Cost of Ownership and Power Consumption

Lifetime cost of ownership is a key metric for IoT devices [Wol16]. The cost of an
IoT silicon includes several components: sensing and actuation, computation, net-
working, as well as packaging. The completed IoT device includes power supply
and packaging. However, installation cost is a significant factor in the cost of owner-
ship. The cost of installing a cable drop in an existing building in the USA is, in the
authors’ experience, around $150. That cost overwhelms the cost of hardware.
Eliminating all wiring – both power and networking – substantially reduces instal-
lation cost. The cost of replacing batteries is significant. Our colleague Rajesh
Gupta reported that the computer science building at University of California, San
Diego, requires a full-time employee to replace batteries on electronic door locks
(Rajesh Gupta, personal communication, February 2014). The ability to power
devices entirely by energy harvesting would eliminate that cost but imposes con-
straints on the devices.
The high cost and effort of wired power have encouraged the development of
energy-scavenging (also known as energy-harvesting) technologies. A range of
physical mechanisms can be used to convert energy for use by the environment.
Since most scavenging sources provide varying amounts of power, the harvested
energy is stored for later use. Electric power may be stored in a battery, a capacitor,
or a supercapacitor. On-chip power management circuitry stores harvested energy
and then regulates the power as it is used by the rest of the chip.
Paradiso and Starner [Par05] identified several widely different sources of
energy, including radio frequency, ambient light, thermoelectricity, and heel strikes.
They pointed out that indoor lighting provides much lower ambient light levels than
are available from the sun. Sudevalayam and Kulkarni [Sud11] surveyed energy-
harvesting technologies for sensor nodes. They identified a range of technologies
with different sources, conversion efficiencies, and harvest yield. They reported, for
example, that light converged by solar cells typically provided 15 mW/cm2, wind by
anemometer provided 1200mWh/day, and provided footfalls 5W.
Romani et al. [Rom17] survey power conversion and management architectures
for ambient-powered IoT devices. Their reference architecture for a no-battery
power management system includes several components. A transducer extracts
power from an external power source with efficiency η. Several sources of internal
power consumption further limit the overall system efficiency: power control cir-
cuitry consumes intrinsic power Pint; the storage element leaks power at a rate Pleak;
monitor circuits consume Pvmon. A bootstrap circuit may be used to initialize the
system from discharge. They note that a key challenge of the power management
controller is to match the effective load impedance to the power source’s internal
impedance.
3.3 Cost per Transistor and Chip Size 19

3.3 Cost per Transistor and Chip Size

Commentators have noted that established technology nodes offer cost-effective

manufacturing for many products [Whi15]. One article [Hru12] quotes an NVIDIA
presentation claiming that cost per transistor for 20 and 14 nm nodes was barely
lower than that of the previous node and that 20 nm is “essentially worthless.” Maly
[Mal94] developed an early cost model for the cost of silicon as a function of manu-
facturing node. His model computed cost per transistor as a function of design den-
sity, minimum feature size, wafer area, and wafer cost. An even simpler cost model
is based on the total cost of a manufactured wafer, including the cost of the wafer
itself and all processing.
As shown in Fig. 3.1, that cost will decrease slightly as the manufacturing pro-
cess matures. However, the cost of a manufactured wafer grows significantly at
advanced nodes [Wol17]. Double patterning became required for lithography at
20 nm. This technique uses two masks for each feature, roughly speaking one per
edge; the size of a fabricated feature can be smaller than the size of a feature on
either of the masks. Double patterning requires two masks per step rather than one;
since mask costs are a large part of the cost of the manufactured wafer, double pat-
terning (and the more recent use of triple patterning) substantially increases manu-
factured wafer cost. Increasing the number of masks adds costs beyond those of the
masks themselves: more time must be spent with the wafers in expensive equip-
ment; wafers spend longer in the manufacturing plant.
We can write a formula for the cost per transistor based on the manufactured
wafer cost Cm and the number of working transistors per wafer ntr:

Cm
Ctr = .
ntr

In the standard Moore’s Law scenario, we expect the number of transistors per
wafer to double from one generation to the next. If the cost of processing the wafer
increases by less than that factor, cost per transistor goes down; if not,

Fig. 3.1 Cost of processed

wafers over time and
technology node
20 3 IoT Devices

cost-per-transistor increases. Put another way, if Cm(B)/Cm(A) > rtr for technology

nodes A and B, where rtr is the factor increase in working transistors per wafer from
technology A to B, then cost per transistor increases. The transition from 28 to
20 nm was an inflection point at which the cost per manufactured wafer grew enough
to offset density gains. Given that these costs continued at smaller nodes, the cost-
per-transistor reached a local minimum at 28 nm. It is likely that 28 nm will prove
to be the global minimum of cost-per-transistor; more advanced lithography meth-
ods have their own costs.
Another major component of chip cost is silicon area. The traditional emphasis
in VLSI has been on large chips to maximize functionality. However, silicon tech-
nology has advanced to the point where we can provide interesting functionality on
very small amounts of silicon, thereby providing low-cost chips. This is true even
for technology nodes such as 28 nm, which are large relative to the nodes used for
latest-generation chips but still very dense relative to historical standards.
The transistor counts of early microprocessors provide context for the circuitry
required to provide useful functionality. The IBM PC’s CPU was an Intel 8088 run-
ning at 4.77 MHz [Wik16A]; the 8088 contained approximately 4000 transistors
[Wik16B, Wik16C].
Packaging is another significant component of the cost of integrated circuits. A
wide range of system-in-package technologies have been developed that provide
several improvements over traditional single-die packaging: the ability to combine
chips from several manufacturing processes, each with its own native devices;
reduced inter-die parasitic values; and lower cost.
The DARPA SHIELD chip [Ral16] provides an example of a very low-cost,
highly integrated IoT chip. SHIELD is designed to be attached hardware modules –
chips, boards, etc. – to provide a secure, traceable identifier for that module. Since
each module in a system would require its own identifier, SHIELD targets a very
low manufacturing cost of one cent. The system includes several die combined in a
leadless package: a CMOS module, an RF pickup coil, and a thin-film temperature
sensor. The CMOS module is 100 μm × 100 μm in a 14 nm FinFET technology; it
combines a digital CPU and communication, onetime programmable memory, a
physically unclonable function (PUF), an analog-digital converter, and power con-
version and management circuitry. The device is powered by near-field RF energy;
the RF coil is used for both power delivery and communication.

3.4 Duty Cycle and Power Consumption

The duty cycle model is widely used to analyze IoT devices. As shown in Fig. 3.2,
the model assumes periodic activation of the device. The duty cycle is the percent-
age of time for which the device is on:
3.4 Duty Cycle and Power Consumption 21

Fig. 3.2 The IoT device

duty cycle

O
D= ´100%.
T
Lower duty cycles mean lower energy consumption. We can change the duty
cycle through a combination of changes to the operating time O and the period T.
Reducing the operating time may reduce the device’s functionality; increasing its
period lowers its data rate.
Let the on-state power consumption of the device be Pon. If we assume zero leak-
age, then the power consumption under duty cycle operation is

O
Pideal = Pon .
T
If the device has a leakage power of Poff, then its average power consumption
over the duty cycle is

O æ Oö
Pleak = Pon + ç 1 - ÷ Poff .
T è Tø
We can also solve for fractional duty cycle as a function of on-state and off-state
power and total power consumption:

O Pleak - Poff
= .
T Pon - Poff

This model carries several implications for the design of IoT devices: the device
must be good at idling at low power; it should provide low energy and time to shut
down and to turn back on.
Communication power is a large fraction of the total power consumption of
many IoT devices. Many IoT devices transmit small amounts of data during the on
portion of their duty cycle. In this scenario, the overhead associated with setting up
a communication is a significant part of the total communication power; many com-
munication systems are designed for connection-oriented service that allows setup
costs to be amortized over a longer communication.
Dementyev et al. [Dem13] measured the power consumption of several wireless
protocols. They used their data to determine the optimal period T for each protocol:
14.3 s for Zigbee and 10.0 s for Bluetooth Low Energy (BLE).
22 3 IoT Devices

3.5 Platform Design

Unlike mobile devices, most IoT devices do not operate continuously. Nonetheless,
they need to retain state from activation for a range of purposes: communication
status, DSP filtering, etc. SRAM requires power to retain state and thereby length-
ens the allowable duty cycle. Flash memory must be written in blocks. Emerging
technologies offer the promise of bit-level persistent-state devices that can be used
within the processor, not just as memory.
Soerken et al. [Soe17] developed a programmable logic-in-memory (PLiM)
using resistive RAM (RRAM) devices. An RRAM device has persistent state – it
can be written and retains its state after the power supply is removed – making it
well suited to the duty cycle characteristics of IoT devices. They designed their
processor to take advantage of the majority-logic characteristics of RRAMs. They
developed a compiler to translate Boolean functions into instruction streams for
their processor.

3.6 Summary

IoT systems open up a new horizon for VLSI design. IoT systems require ultra-low
power systems that combine disparate elements – computation, communication,
and sensing – at very low price points. IoT systems emphasize small, capable chips
in contrast to the large chips that have driven the industry for many years. We are at
the early stages in the development of this new category of chip.

References

[Dem13] Dementyev, A., Hodges, S., Taylor, S., & Smith, J. (2013). Power consumption analysis
of Bluetooth Low Energy, ZigBee and ANT sensor nodes in a cyclic sleep scenario. Wireless
Symposium (IWS), 2013 IEEE International, Beijing, 2013, pp. 1–4.
[Die16] Dieseldorf, C. G. (2016). Foundries Take Over 200mm Capacity Fab by 2018. www.
semi.org, January 25, 2016.
[Hru12] Hruska, J. (2012). Nvidia deeply unhappy with TSMC, claims 20 nm essentially
worthless. extremetech.com, http://www.extremetech.com/computing/123529-nvidia-deeply-
unhappy-with-tsmc-claims-22nm-essentially-worthless, March 23, 2012.
[Mal94] Maly, W. (1994). Cost of Silicon Viewed from VLSI Design Perspective. Design
Automation, 1994. 31st Conference on, San Diego, CA, USA, 1994, pp. 135–142.
[Par05] J. A. Paradiso and T. Starner, Energy scavenging for mobile and wireless electronics
IEEE Pervasive Computing, vol. 4, no. 1, pp. 18–27, 2005.
[Ral16] Ralston, P., Fry, D., Suko, S., Winters, B., King, M., & Kober, R. (2016). Defeating
counterfeiters with microscopic dielets embedded in electronic components. Computer, 49(8),
18–26.
References 23

[Rom17] Romani, A., Tartagni, M., & Sangiorgi, E. (2017). Doing a lot with a little: Micropower
conversion and management for ambient-powered electronics. Computer, 50(6), 41–49.
[Soe17] Soeken, M., Gaillardon, P. E., Shirinzadeh, S., Drechsler, R., & Micheli, G. D. (2017).
A PLiM computer for the internet of things. Computer, 50(6), 35–40.
[Sud11] Sudevalayam, S., & Kulkarni, P. (2011, Third Quarter). Energy harvesting sensor nodes:
Survey and implications. IEEE Communications Surveys & Tutorials, 13(3), 443–461.
[Whi15] White, M. (2015). IoT, Cost-per-Transistor Extend Lifetimes of Established

Technology Nodes. Electronic Design, May 15, 2015, http://electronicdesign.com/eda/
iot-cost-transistor-extend-lifetimes-established-technology-nodes
[Wik16A]
Wikipedia. (2016). IBM Personal Computer. https://en.wikipedia.org/wiki/IBM_
Personal_Computer. Accessed October 19, 2016.
[Wik16B] Wikipedia. (2016). Intel 8088. https://en.wikipedia.org/wiki/Intel_8088. Accessed

October 19, 2016.
[Wik16C] Wikipedia. (2016). Transistor count. https://en.wikipedia.org/wiki/Transistor_count,

Accessed October 19, 2016.
[Wol16] Wolf, M. (2016). Ultralow power and the new era of not-so-VLSI. IEEE Design & Test,
33(4), 109–113.
[Wol17] Wolf, M. (2017). The physics of computing. Cambridge MA: Elsevier.
Chapter 4
Event-Driven System Analysis

4.1 Introduction

This chapter describes modeling and analysis methods for Internet of Things (IoT)
system design. IoT systems require new types of analysis because events do not
necessarily result in immediate actions or maintain their order relative to other
events.
Traditional methods such as the distributed control-oriented methods of Thiele
and Ernst consider possibly infinite streams of events or samples, but the lifetime of
an event/sample in the system is relatively short. In contrast, IoT systems must deal
with event lifetimes at multiple time scales: some events may schedule activity only
seconds in the future, while other events may schedule activity days, weeks, or
months ahead. IoT also do not maintain temporal order of causality – one event may
cause an event in the near future, while another event may cause an event in the far
future. We need new analytical methods for multiple time scales and complex cau-
sality relationships.
The primary goal of our analysis is the understanding of the required character-
istics of the IoT platform. We propose a model of the IoT system as a network with
devices as leaf nodes and hubs as non-leaf nodes. Hubs perform routing functions
but for our purposes their key role is to control the timing of event activity through
the use of timewheels. While we assume that events carry key-value pairs, we are
not concerned here with the semantics of events. Instead, we analyze the lifetimes
of event populations. Event populations depend in part on the activity of the envi-
ronment in which the IoT operates. To accommodate a wide range of realistic sce-
narios, we develop models based on both deterministic and stochastic event
timing.

© Springer International Publishing AG 2018 25

D. Serpanos, M. Wolf, Internet-of-Things (IoT) Systems,
https://doi.org/10.1007/978-3-319-69715-4_4
26 4 Event-Driven System Analysis

4.2 Previous Work

Several lines of work have established event-based models for real-time networked
systems. One of the goals of these projects has been to unify the analysis of network-
oriented events and the computation on the network nodes that transform one stream
of events into another.
Chakraborty et al. [Cha03] developed a real-time calculus that models events and
resources. They model an event stream R[s, t) over the prescribed time interval as a
pair of arrival curves: αl(∆) for the lower bound on the number of events in the
interval ∆ and αu(∆) for the upper bound of events in the interval. They show how
to model event streams with jitter. They use β functions to model the service pro-
vided by computational and communication components. They show how to ana-
lyze single streams, multiple interacting streams, and platforms with multiple
computing and communication resources. Maxiaguine et al. [Max04] used work-
load curves to characterize the computational workload of real-time systems. They
showed how to use their methods to analyze both a rate-monotonic system and
streaming architectures.
Henia et al. [Hen05] give definitions and formulas for events and event streams.
Many of their results apply to our model; we summarize some of their applicable
results here.
Event time applies to both generation and release time. An event time includes a
nominal time and jitter:

T,J

A periodic event stream has parameters period and jitter:

P, J

The upper event function ηu(Δt) gives the maximum number of events in the
interval Δt. Similarly, the lower event function ηl(Δt) gives the minimum number of
events in the interval Δt. The upper and lower event functions for a periodic event
stream with jitter are

Dt + J æ Dt - J ö
h Pu + J = , h Pl + J = max ç 0,
P è P ÷ø

They give formulas for the jitter of the output of components that combine event
streams using AND and OR combination methods.
4.4 IoT Network Model 27

4.3 Motivating Example

IoT systems are built for a variety of applications: industrial control, environmental
monitoring, logistics, etc. We will use examples in this paper derived from our
experiments with IoT systems for long-term care [Wol15]. This application pro-
vides us with use cases typical of smart homes (turning on and off lights, energy
management, etc.) as well as use cases associated with health care (scheduling med-
ications, checking on the condition of residents, etc.). Our example IoT system
operates in a home with several residents, as a rotating set of staffers, and visitors.
A variety of sensors monitor activity in the home: cameras, utility sensors, smart
objects, etc. The IoT system is designed to track the activity of residents and staffers
and to alert staffers of situations that may deserve their attention.
The system architecture consists of several elements:
• A set of sensors
• A local hub that monitors the sensors as I/O devices
• A cloud-based node for some analytical functions
A key feature of the local hub is its internal timewheel (Coelho, D., 2014, August
2, private communication). Timewheels are used in event-driven simulation to man-
age simulator event activity; in this case, we use the timewheel to manage events in
the real world as mediated by the I/O devices. Events are timestamped with a time
at which they should occur, which may be later than the time at which the event was
generated. The timewheel is a time-sorted queue; when the clock time equals the
timestamp of the event at the head of the timewheel, that event is dequeued and
processed.

4.4 IoT Network Model

Our IoT network model is oriented toward the analysis of event behavior in the
system. Because events have long lives, memory in the form of timewheel queues
plays a critical role in the model.

4.4.1 Events

The model of computation is based on long-lived events. An event is generated at a

node and stored until it is ready to be completed, at which time it is consumed by a
node.
An event is a 5-tuple:

key, value, dest, gen _ time, release _ time

28 4 Event-Driven System Analysis

The semantics of the event is given by the key-value pair. The destination of the
event is the device that should process the given key-value pair. The modeling meth-
ods described in this paper are not concerned with the semantics of key-value pairs.
We also need to know the temporal behavior of an event, which is given by two
values. The generation time ϑ is the time at which the event was created. The gen-
eration time is useful in our analysis; an implementation may or may not keep track
of this value. The release time ρ of an event allows the IoT system to perform
delayed actions – one event in the environment may not cause an immediate
response but rather one that happens some time later. We refer to the difference
between generation and release time as lifetime of an event is λ = ρ − ϑ. Events may
be generated periodically or aperiodically. Activation or release times may be peri-
odic or aperiodic.

4.4.2 Networks

A network consists of nodes and links. We will discuss nodes in more detail below.
We model communication links are unidirectional. Most physical hubs are full
duplex, but we model links as unidirectional to advance our analysis.

4.4.3 Devices and Hubs

A node may be one of two types: device or hub.

A device appears only as leaves in the network. A device has at most one input
and at most one output link. Logically, a device receives or generates events.
Physically, event generation can be caused either by physical events or by internal
node activity; physical event receipt may cause the physical node to initiate a physi-
cal event or change its internal state.
Hubs are non-leaf nodes in the network. A hub may have more than one input or
output link. Hubs may include computing and storage. However, for analytical pur-
poses, the key role of a hub is to sequence events. A hub maintains a timewheel, a
time-ordered queue of events, and a clock. (The timewheel is a notion borrowed
from an event-driven simulation, but we use it in this case to manage events in
cyber-physical systems.) Events arriving from the hub’s devices are entered into the
timewheel in order of activation time. The head event of the queue is removed from
the timewheel when its activation time equals the clock time. The hub then sends the
event to its destination.
Hubs must keep track of real time in order to dispatch events. Devices may or
may not need to keep track of real time, depending on their functions. Given that
events are dispatched by the hubs, they can rely on the hub’s notion of real time to
initiate events. When scheduling events, they may be able to set the release time for
the event relative to the current time, which avoids the need for the device to directly
keep track of real time. Our model only assumes that hubs are required to know real
4.4 IoT Network Model 29

time, which simplifies analysis. Allowing devices to avoid maintaining real-time

clocks may have some advantage in implementation as well.

4.4.4 Single-Hub Networks

A single-hub network consists of one hub mode and one or more device nodes and
their associated links. The hub manages the exchange of events between its device
nodes.
In a single-hub network, input traffic arrives at the hub from its device nodes,
while output traffic is generated by the timewheel and goes to the devices.
As a simple example, consider scheduling medications for residents of the home.
If a resident receives medicines twice per day, once in the morning and again in the
evening, the device responsible for scheduling the medicines must generate an event
for each administration. The event for the next medicine administration is probably
generated when the current medicine administration is released, giving an event
lifetime of 12 h.
The morning routine of the residents presents a more complex set of events and
more scheduling choices. Each resident’s routine will generate a series of events
(getting up, toileting, eating breakfast, etc.); depending on the activity, all the events
in the routine may be scheduled at once, or some may be scheduled on the comple-
tion of other events. If all residents get up at once, they create both congestion in the
house and congestion in the hubs and their timewheels – the maximum number of
events in the system will be a function of the number of residents as well as the
complexity of their routines. By staggering the timing of their activities, we can
both reduce physical congestion as well as reduce the number of events that the
hubs must deal with at any given time.

4.4.5 Multi-hub Networks

A more general network may contain more than one type of hub. One link or a pair
of links is used to connect the hubs. For the moment, we consider only tree-
structured networks.
In our example system, the in-house system consists of a hub and a set of devices.
The cloud analytics system also uses a hub and timewheel to manage the times at
which events should be processed. For modeling purposes, the analytics engine
itself is a device.
We model event routing as hub-to-hub transfers in which an event is removed
from one timewheel and placed on another. When an event is transmitted to another
hub, we may use additional queue operators to remove the event before it reaches
the head of the queue. We will discuss the effects of event routing in more detail in
the next section.
30 4 Event-Driven System Analysis

4.4.6 Network Models and Physical Networks

The mapping between model nodes/links and nodes/links in the physical network
need not be one-to-one. A single physical device may house several logical nodes.
A single network physical link may be used to transport several logical links.
We can rely on results from parallel computing [Dua02] for techniques for rout-
ing events in multi-hub networks. Physical networks may use separate memories for
queues and buffers on network links.
The network model helps us to understand the behavior of more complex physi-
cal links. We can first separately analyze half-duplex traffic on links and then use
that analysis to understand the characteristics of full-duplex links.

4.5 IoT Event Analysis

The theories for event-based analysis of distributed control networks described in

Sect. 4.2 were designed for transducer networks in which events maintain their time
order. In contrast, our events may be generated in one order but released in another
order. The reordering effects of release times and the timewheel substantially change
the analysis of IoT networks as compared to distributed control. We start with analy-
sis of event populations using simple models of event generation. We then go on to
identify stochastic models that are useful for the analysis of IoT event systems.

4.5.1 Event Populations

Because events in IoT systems are long-lived, we must consider the lifetimes of
event populations. Because events may be released long after they are generated, the
system may need to accommodate a large number of events even if no events are
currently being generated.
The event population is the number of events that are still alive, given by the dif-
ference between the number of generated and released events. We can evaluate
event population over the entire network or over a set of components in the network.
When events are generated and released with jitter, we can write formulas for the
upper and lower population; here we concentrate on a jitter-free form of the analysis
to emphasize basic principles.
A general form for the population count is to enumerate all events from the sys-
tem start time:
t
P ( t ) = ò éëJ ( t ) - r ( t ) ùû dt.
0
4.5 IoT Event Analysis 31

This formulation is cumbersome since it requires the entire system history.

However, without some knowledge of the event lifetimes, we can do no better. And
without a bound on event lifetime, the number of events in the system can increase
without limit.
A practical consequence of this observation is that useful IoT systems must put
an upper bound on the lifetimes of events.
If we have a maximum lifetime L on the lifetime of an event, we can write the
event population as
t
P ( t ) = P ( t - L ) + ò éëJ ( t ) - r ( t ) ùû dt
L
We can also write a version of this equation taking into account event jitter.
We need to know the event population at time t − L because some events may
have been generated that have not yet expired.
If no events are generated in an interval L then we can guarantee that the event
population at the end of that interval is zero.
Event population determines buffer requirements for components. The maxi-
mum population determines the memory requirements of the timewheel queue.
Maximum populations on links help to determine the queue sizes on those links.
As a simple example, consider a single-hub network. The event stream controls
medication dispensing, with medications being dispensed every 12 h. Scheduling
medications may be done separately from dosing them, but let us assume for the
moment that each medication dispensation also causes the next dispensing event to
be scheduled. If we let T = 12 h and assume that one person is in the system, then
γu(T) = γl(T) = 1, αu(T) = αl(T) = 1, and ηu(T) = ηl(T) = 2. The maximum lifetime is
L = 12 h. The event population at t = 12 − ϵ, just before the first set of prescription
dispensed is released and the second set generated is
12 -
P (12 -  ) = ò éëJ ( t ) - r ( t ) ùû dt = 2
0
We can easily generalize this formula to the case of n people.
The maximum population in an interval [t1, t2] is

Pmax ( t1 ,t2 ) = max P ( t )

t [ t1 ,t2 ]

Event generation in many IoT systems is not strictly periodic – some events or event
streams may be activated aperiodically. In this case, the event population depends
on the use case.
We can evaluate event populations when event characteristics are stochastic. For
example, consider an event stream that is generated periodically at a rate of one per
second. The activation times (measured relative to the generation time) are given by
a uniform distribution over the interval [1, 10]. The maximum population is
32 4 Event-Driven System Analysis

Fig. 4.1 A multi-hub

network

Pmax ( t1 ,t2 ) = max ò éëJ ( t ) - r ( t ) ùû dt

t [ t1 ,t2 ]
t1

We will develop in Sect. 4.5.3 techniques to characterize event populations under
several different models of event generation.
For the moment, consider the case of the morning routine for n people. Let us
assume for concreteness that a stream of four events is generated, each 1 min apart,
with events released 5 min after generation. If everyone gets up at once, then the
event population in the first 4 mins is P(4) = 4n and Pmax = 4n. If we stagger the
schedules of the residents so that each gets up 5 mins apart, then the maximum
population reduces to 4.
In the case of a multi-hub system, different nodes may have different event popu-
lations and different maximum event populations. In the smart home, we perform
some operations locally and some in the cloud. We model this with the network
shown in Fig. 4.1: two hubs are in the home, one for input devices and one for out-
put devices (a choice made here for modeling clarity); one hub is in the cloud. The
cloud hub is connected to a single device that performs analysis algorithms. The
analysis algorithms consume events, process them, and then possibly generate out-
put events. (One example of such analysis is tracking [2].)

4.5.2 Stochastic Event Populations

A wide variety of assumptions and stochastic models are possible for events. In this
section, we use some basic models to derive important design metrics. Although no
one to our knowledge has gathered large traces of IoT activity, we can use models
of traffic from related domains to help us understand IoT design.
We can gain some intuition by considering the simpler case of the Poisson distri-
bution. A common model for telephone traffic is that call arrivals and departures are
4.5 IoT Event Analysis 33

each modeled as Poisson processes. In our case, we use the Poisson distribution to
model event generation at a rate λ. If successive events have non-overlapping life-
times, then their maximum population in that interval is 1; if their lifetimes overlap,
then the maximum population is 2, which must be accommodated by buffering. The
probability that two events have overlapping lifetimes L is

P [t < L ] = 1 - e- l L

This simple formulation suggests that λL, the product of event generation rate and
lifetime, is a useful metric for judging maximum event populations.
The Erlang-B distribution provides a more accurate model for event populations.
In the case of IoT events, the event dwell time corresponds to call duration; the
queues correspond to telephone lines. (The Erlang-C distribution models call wait-
ing with queues. In our case, consider the queue as a set of servers consisting of
memory locations. One memory location/server is required for each event to wait
for its release time.)
The offered traffic is in units of erlangs:

E = lL

In our case, λ is the event arrival rate and L is the event lifetime. The probability of
blocking (i.e., dropping an event due to a full queue) is

E m / m!
Pb = B ( E,m ) =
å
m
E i / i!
i =0
where m is the size of the queue.
The offered event traffic in erlangs is a useful rule-of-thumb metric for IoT sys-
tem traffic – both the frequency of events and their dwell times must be considered
to understand their effect on timewheel size.
We can use Pb to design the timewheel capacities of the hubs, either using the
maximum traffic as a guide or evaluating the traffic at different points in time using
the population functions. Given the systemwide offered traffic, we can find Pb for
the entire network. However, in a multi-hub system, we must determine how to
partition the timewheel memory between the hubs.
We can model the traffic hub by hub:
n n

åE = ål L
i i i
i =1 i =1
From this, we can determine the Pbs. However, this approach does not minimize
total system memory. If we assume that all the hubs share the same values for arrival
rate and event lifetime, then E < ∑1 ≤ i ≤ nEi.
We describe in Sect. 4.5.4 how to transfer events between hubs to balance queue
sizes.
34 4 Event-Driven System Analysis

4.5.3 Environmental Interaction Modeling

We can identify three methods for modeling the interaction of the IoT system with
its environment, each with its own degree of accuracy and detail.
The simplest model treats both the device and the user as timed finite-state
machines. Given a path through the user machine that defines a given use case, we
can form the product of the device machine and the user machine path. The result-
ing FSM, along with a timing regimen that is specified by the use case, tells us when
events are generated by the device. That trace can be used to build the event popula-
tion trace.
A more sophisticated model treats the user as a Markov decision process (MDP)
with fixed timing. A Markov decision process is a stochastic model used for optimi-
zation. An MDP is defined by a set of states and possible actions out of each state.
Each action is assigned a reward R. Transitions out of the action to the next state are
assigned probabilities. We can use any of several different algorithms (dynamic
programming, linear algebra) to find the path that maximizes the reward. In this
scenario, we solve for the optimal reward path and form its product with the device
model, using a fixed time model. Figure 4.2 shows an example of a simple device
model and user model. The device model combines the actions of all the component
devices related to the routine into a single state machine for simplicity. The actions
in the user model MDP correspond to states in the device model.
A yet more complex model uses a continuous time Markov decision process
(CTMDP). The most common mathematical form of this model is as an MDP with
the timing of state transitions modeled as a Poisson process [Buc11]. Standard MDP
approaches can be used to solve for the optimal path with timing given by the
Poisson process.

4.5.4 Event Transport and Migration

An event does not necessarily have to be stored on the hub that owns either the
event’s source or destination. In a multi-hub system, we can station events at nonlo-
cal hubs to avoid overflowing a hub’s queue capacity or improve its battery life. If
an event is queued nonlocally, we must factor transmission time into its release to
ensure that it reaches its destination device at the proper time.
For simplicity, we consider the case of no energy cost for transporting events
across the network. Let Pe be the power consumption of storing one event in mem-
ory for a unit time. Given a population of events Π, the energy required to store all
events in the population until their release times is

Epop = å éë r ( e ) - J ( e ) ùû Pe
eP
4.5 IoT Event Analysis 35

Fig. 4.2 Models of the morning routine for devices and residents

We have a set of H hubs each with available battery energy Eh(i). We want to find an
allocation of events to hubs such that

"iH : Epop ( i ) £ Eh ( i )

This is a classic bin-packing problem, although we want to solve it as a distributed
problem without a centralized list of events. In practice, transmission energy reduces
the set of plausible event allocations.
We propose a heuristic algorithm for event migration:
• Find a partial ordering of the hubs such that no two adjacent hubs are in the same
set and all hubs are covered.
36 4 Event-Driven System Analysis

• Hubs proceed in order so that no two adjacent hubs off-load simultaneously.

• Each hub off-loads enough events to meet its battery requirements. Events are
moved to adjacent hubs with the greatest available battery capacity.

Acknowledgment Thanks to the team at Alya Networks for useful discussions on key-value-
based IoT networks.

References

[Buc11] Peter Buchholz (2011). Continuous time Markov decision processes: Theory, applica-
tions, and computational algorithms. TU Dortmund Informatik IV lecture notes.
[Cha03] Chakraborty, S., K¨unzli, S., & Thiele, L. (2003). A general framework for analysing
system properties in platform-based embedded system designs. Proceedings sixth design, auto-
mation and test in Europe (DATE) (pp. 190–195). Munich, Germany.
[Dua02] Duato, J., Yalamanchili, S., & Ni, L. (2002). Interconnection networks. San Francisco:
Morgan Kaufman.
[Hen05] Henia, R., Hamann, A., Jersak, M., Racu, R., Richter, K., & Ernst, R. (2005). System
level performance analysis: The SymTA/S approach. IEE Proceedings: Computer Design
Techniques, 152(2), 148–166.
[Max04] Maxiaguine, A., Kunzli, S., & Thiele, L. (2004). Workload characterization model for
tasks with variable execution demand. Proceedings of the conference on seventh design, auto-
mation and test in Europe (DATE) (pp. 1040–1045). Paris, France.
[Wol15] Wolf, M., van der Schaar, M., Kim, H., & Xu, J. (2015). Caring analytics for adults with
special needs. IEEE Design & Test, 32(5), 35–44.
Chapter 5
Industrial Internet of Things

5.1 Introduction

The Internet of Things (IoT) has already brought a revolution to our understanding
of applications in a wide range of human activity. This trend is expected to increase
in the near future, as the potential economic impact of IoT is expected to be between
900 billion USD and 2.3 trillion USD on a yearly basis up to 2025 [Man13]. IoT
applications are spreading to various sectors including smart energy, manufactur-
ing, agriculture, health, security and safety, smart cities, smart buildings, and smart
environment. All these application areas repeat the same basic model: a large num-
ber of smart devices, interconnected over wired or wireless media, interacting and
coordinating to achieve a goal.
In the industrial environment, the effort for smart factories [Zue10], the Industrie
4.0 strategy [Ind14], the Industrial Internet [GE17], and the European initiative for
the Factories of the Future [FoF] have initiated the adoption of IoT in industry with
the goals of increasing flexibility and productivity, while reducing production cost.
The developing concept is the Industrial IoT (IIoT).
The Industrial Internet of Things is part of the general IoT evolution. However, it
faces challenges that are unique and differentiate it from the other systems and ser-
vices of IoT due to the need to integrate programmable logic controllers (PLC) and
supervisory control and data acquisition systems (SCADA). PLC and SCADA sys-
tems, together with the related industrial networks that interconnect them, constitute
the infrastructure of operational technology (OT), which has traditionally evolved
independently from the typical IT technology, because it addresses the needs of
systems in the field – industrial floor, energy production facilities, energy distribu-
tion networks, etc. – with strong requirements such as continuous operation, safety,
real-time operation, etc. The capabilities offered by the emerging IIoT technology
pose challenges for the integration of these OT systems with the traditional enter-
prise IT systems at many levels, from enterprise management to cyber security. For
example, enterprise resource planning systems (ERP) need to be expanded to

© Springer International Publishing AG 2018 37

D. Serpanos, M. Wolf, Internet-of-Things (IoT) Systems,
https://doi.org/10.1007/978-3-319-69715-4_5
38 5 Industrial Internet of Things

include manufacturing operations, which are managed currently by manufacturing

execution systems (MES) that have grown independently and present significant
interoperability challenges to their integration. Clearly, an integrated system that
manages the complete enterprise/factory hierarchy, from business processes to sen-
sors, provides significant flexibility and presents new opportunities to enterprises.
Industrial technology is not part only of manufacturing or factories. The maturity
of the technology and its cyber-physical control capabilities has spread its use out-
side traditional factory environments, and now they constitute a significant part of
the critical infrastructure at many fronts. Energy production and distribution infra-
structure includes OT systems, which are the indispensable infrastructure on which
modern smart grids are built. Actually, the energy sector is a high priority in the
evolution of IIoT, not only because there is increasing need to consumers for energy,
especially in light of the population growth, but also because energy management is
a critical factor in the industrial sector and the desired low-cost production of goods
and services. In addition to the energy sector, industrial systems are widespread in
many other sectors of critical infrastructure, such as water management and
transportation.
The interoperability challenges to the convergence of IT and OT are only a part
of the challenges in the emerging IIoT. Appropriate architectures need to be devel-
oped to build and manage effective IIoT systems, technologies for the design and
management of cyber-physical systems, sensors and networks need to be devel-
oped, and, importantly, safety and security need to be addressed in a unified way in
the context of IIoT. Safety and security are significant challenges, because, tradi-
tionally, security has been a concern in the IT sector, while safety has been the major
concern in the OT sector. Bringing the two together has brought the realization that
safety cannot be achieved without security, while, at the same time, security needs
to include technologies that combine dependability and meet strong requirements
for real time and low power in many application domains. Although the security
issues of industrial control systems have attracted attention in the last decade and
standards have been evolving at a much faster pace than in the past, e.g., the ISO/
IEC 27000 and the ISA/IEC 62443 families of standards [IEC16, ISA16], there are
still significant challenges at the technology, architecture, and management fronts to
obtain solutions for the unified IIoT.
In this chapter, we present the concepts and evolution of the IIoT starting from
the Industrie 4.0 strategy and proceeding to the Industrial Internet. We describe the
IIoT reference architectures as they evolve from the ITU effort to the Industrial
Internet Consortium. Finally, we describe some representative challenges in the
evolution and implementation of IIoT focusing on the energy sector. As security and
safety constitute a significant challenge in IIoT as well as in IoT, in general, we
focus on this challenge in the following chapter.
5.2 Industrie 4.0 39

5.2 Industrie 4.0

Industrie 4.0 is a strategic initiative in Germany that targets to bring IoT technolo-
gies to the manufacturing and production sectors [Ind14].The goal is to enable
Germany to keep a leading role in manufacturing achieving efficient and low-cost
production with flexible workflows. The means to achieve this goal is the wide-
spread inclusion of cyber-physical systems in the manufacturing and production
processes, in order to insert intelligence in the systems and processes, to enable their
high connectivity and communication, and to achieve their coordination into more
complex but flexible processes that lead to high-quality, low-cost products.
Industrie 4.0 takes its name from the identification of the new, emerging industry
as the fourth revolution of industrial production. It is widely accepted that industrial
production to date has gone through three (3) revolutions. The first industrial revolu-
tion, between the eighteenth and nineteenth century, is the one where mechanized
production facilities were introduced in the production of goods and services, where
the required energy was provided by water and steam. Electrical energy was intro-
duced during the second revolution, which led to mass production, as electricity
boosted productivity. In the post WWII era, the inclusion of electronics and soft-
ware, i.e., industrial information technology, to the mechanical and electrical com-
ponents led to the third revolution that enabled automation at high levels. Currently,
many industrial stakeholders believe that we are at the verge of the next, the fourth,
industrial revolution, through wide adoption and use of cyber-physical systems that
leads not only to even higher levels of automation but enables mass customized
manufacturing and production of goods and services, due to the flexibility offered
by the easily programmable, configurable, and controllable manufacturing lines.
The effort for Industrie 4.0 is based on the widespread deployment and use of
computational and communication resources. The last two decades have been char-
acterized by significant advances in high performance, low-power processors, mem-
ories, and communication components that enable efficient processing and
networking. These advances have brought significant processing capabilities to a
large number of devices that are deployed to consumers or to the field. Smart con-
sumer devices have become norm. Smartphones provide hundreds of applications
and enable services ranging from identifying travel and transportation routes to
mobile banking and health monitoring. Smart televisions combine and provide vari-
ous types of entertainment and network services, from customized TV channel con-
trol and management to Internet gaming and home device management. Smart
home appliances monitor parameters, from environmental temperatures to water
and energy consumption, enabling citizens to manage their homes efficiently and
effectively leading to the required living quality while reducing operational cost at
various fronts.
The large basis of computational resources and connectivity becomes apparent
by the published numbers of embedded processors and components that are cur-
rently produced. According to [Ind14], the vast majority of produced processors,
approximately 98%, are deployed in embedded systems. Deployed semiconductor
40 5 Industrial Internet of Things

memory is also growing and expected to grow at 40% year over year in 2017
[Mic17]. Furthermore, the significant advances of wired and wireless networks in
the last two decades have led to ubiquitous connectivity, approaching 100% in cities
and towns, through different technologies.
The available processing and communication basis leads to an evolving hierar-
chy of embedded systems and services up to the level of the Internet of Things,
Data, and Services. Examples of this evolution can be identified at several applica-
tion areas. In transportation, for example, embedded systems are widespread con-
trolling functions from car entertainment systems to car seat control. At this level,
embedded processors are programmed to control specific, individual parameters,
e.g., height and movement in car seats, based on user commands. However, embed-
ded systems in cars are also networked, either within the car system or with the
environment, providing networked embedded services; automatic toll payment is
one of them where embedded systems in the car and the toll booths communicate
with each other, in order to complete the electronic payment transaction of the toll
passage. Such payment systems from several tolls, for example, can be further com-
bined in a distributed system that enables traffic and toll management at a wider
scale, leading to more effective transportation infrastructure that achieves lower
waiting times and fuel costs for travelers as well as lower operational cost and, thus,
higher income to transportation management authorities. One can even envision an
even higher level of connectivity of such complex transportation systems to smart
cities that combine transportation management with additional services, such as
energy distribution, civil services, emergency services, etc., as required at different
times, locations, and during special events.
The advances of sensor technologies, in addition to the evolution of embedded
systems and communication networks, make all these scenarios realistic.
Importantly, sensors bridge the gap between the physical world and the digital
world, providing increasingly rich information to digital systems and enabling intel-
ligent control of systems and processes. In that respect, manufacturing and indus-
trial automation has been traditionally employing IT technologies with sensors and
electromechanical systems, leading the development and deployment of technolo-
gies and concepts for intelligent control, systems, and services. Thus, the develop-
ment of the Industrie 4.0 strategy and the related initiatives comes as a natural
evolution step of industrial technologies influencing and being influenced by the
advancement of consumer technologies of the Internet of Things.
The smart factory concept embodies the goals of the Industrie 4.0 strategy to a
large degree. The concept is based on the hierarchy of cyber-physical systems men-
tioned above, where smart production systems are interconnected in a multilevel
hierarchy to achieve a high degree of automation, targeting flexibility, efficiency,
autonomy, resilience, safety, and low cost. Smart machines will be interconnected
to establish smart plants, which, in turn, will be combined to provide smart facto-
ries. Considering the typical components of manufacturing process, smart factories
are targeted to automate efficiently all components and stages. Materials and
resources will be managed and introduced in the process efficiently; production
processing will be managed in real time minimizing the used resources for the
5.3 Industrial Internet of Things (IIoT) 41

p roducts and the operations, while reconfiguration and reorganization of production

processes and customization of products will be feasible in real time and with safety
for infrastructure and operators, minimizing environmental impact. Customers will
be able to monitor the progress of the development of ordered products, while man-
ufacturers will be optimizing their logistics chains.

5.3 Industrial Internet of Things (IIoT)

The Industrial Internet of Things (IIoT) has emerged as a general concept of the
application of the Internet of Things to the industrial sector. Effectively, it is a gen-
eralization of Industrie 4.0, which appears to focus more on industrial process effi-
ciency. The IIoT vision includes all aspects of industrial operations, focusing not
only on process efficiency but also on asset management, maintenance, etc.
Considering that IIoT is effectively IoT in the industrial sector and that the
Industrie 4.0 concepts are effectively a subset of IIot, as shown in Fig. 5.1, one
needs to identify the difference between IoT and IIoT. Although the basic concepts
are the same, i.e., interconnected smart devices that enable remote sensing, data col-
lection, processing, monitoring, and control, the parameters that identify the IIoT
subset of IoT are the strong requirements for continuous operation and safety as
well as the operational technology employed in the industrial sector. As an example,
one can consider the difference between a consumer service such as a health moni-
toring application on a smart watch and an industrial service such as the monitoring
of a steam pump. Although both applications collect real-time data, e.g., steps or
body temperature in the health application case and pressure or steam volume in the
steam pump case, transmit the data, identify events, and provide feedback or com-
mands to operators/consumers and subsystems, clearly, continuous operation and
safety place stricter requirements in the steam pump case, where the potential effect
of a failure is significantly more catastrophic and may lead to costly operation down
time and even human injuries or loss of life.
These characteristics of the industrial sector – technology and requirements –
lead to specialized, demanding solutions for technology and services, justifying the
focus of the industrial sector on a specialized IoT concept. This has resulted to the
strong interest of the industrial sector in the development of specialized concepts,
from strategy to application and technology. The conventional business develop-
ment models that include numerous interdependencies between stakeholders, from
supply chains to service promotion, lead also to a strong need for interoperable
solutions at many levels, from the device level to services. Thus, there is need for
coordinated activities in the evolution to IIoT, which is addressed by consortia, such
as the Industrial Internet Consortium [IIC14] that provides significant leadership in
this emerging field.
The General Electric company introduced the term Industrial Internet in 2012,
as a leader of the Industrial Internet of Things, identifying also the technologies of
machine-to-machine communication, SCADA, HMI, industrial data analytics, and
42 5 Industrial Internet of Things

Fig. 5.1 IoT, IIoT, and

Industrie 4.0 relationship

IoT IIoT Industrie 4.0

cybersecurity as the main constituents of the IIoT vision [GE17]. Interestingly, they
also calculate the impact of the Industrial Internet to 46% of the global economy,
while in the energy sector they calculate an impact of 100% on energy production
and 44% on energy consumption globally [GE17].

5.4 IIoT Architecture

The development and deployment of IIoT systems and services require the develop-
ment of architectures that enable efficient and effective operations as well as interop-
erability considering the anticipated end-to-end services and the large number of
stakeholders involved for devices, cyber-physical systems, communication systems
and networks, service providers, and business developers. Thus, significant effort is
being spent to develop standards and reference architectures that will be accepted
and adopted by the various stakeholders. The International Telecommunication
Union (ITU) has addressed this issue, publishing in 2012 the ITU-T Y.2060 recom-
mendation, which introduces a reference architecture for IoT, in general, including
explicitly applications that fall in the context of IIoT, such as smart grid, intelligent
transportation systems, e-health, etc. [ITU12]. The Industrial Internet Consortium
(IIC) has also been working on a reference architecture for IIoT and currently has
published Version 1.7 of the Industrial Internet Reference Architecture [IIC17].
This architecture is an elaborated reference architecture, significantly more detailed
than the ITU one, addressing all important aspects to all categories of stakeholders.
Taking into account the details of both reference models, one can consider the IIC
model as a specialized evolution of the ITU model, addressing in more details the
important issues of IIoT relatively to the more generic ITU reference model that
encapsulates the requirements for the general IoT.
The ITU effort has expanded the communications’ vision to include communica-
tion of “anything” to the communication concepts of “any time” and “any place.”
Importantly, it includes all expected applications, including industrial ones, specifi-
cally mentioning smart grids and intelligent transport systems among others. As
“things,” ITU considers physical and virtual objects that are identifiable and able to
5.4 IIoT Architecture 43

T
T

Communication over CN without Gateway

T
Communication Network
(CN)
Direct communication

T G
Communication over CN through Gateway

Fig. 5.2 Communication methods for IoT devices

connect to communication networks, while they have related information that is

either static or dynamic. Importantly, since communication is a critical part of the
whole IoT concept, physical things need to be attached to “devices” that are con-
nected to networks, so that any analog information can be converted to digital and
transmitted through the networks. Devices can be simply data-carrying communi-
cating and storing data, data-capturing interacting with the physical objects through
reader and writers, sensing and actuating devices, or general-purpose devices with
embedded processing and communication resources, such as machines, appliances,
and consumer electronic products.
An important issue in the ITU reference model is the communication model
among devices. As Fig. 5.2 indicates, the model considers three methods of com-
munication, based on the employment of gateways (G) and the use of the commu-
nication network (CN). Devices can communicate without the use of gateways,
directly, over local networks, and/or over the communication network, or they can
communicate over the communication network exploiting gateways.
The ITU model accommodates fundamental characteristics of IoT that are iden-
tified. These fundamental characteristics are interconnectivity, scale, heterogeneity,
services for things and the dynamic nature of device information, and connectivity.
Interconnectivity is a significant characteristic because “anything” can connect to
the global network for any application. As the number of connected devices increases
dramatically, scaling becomes a significant parameter that needs to be addressed at
all levels of IoT and IIoT; the scaling issue relates not only to communication end
points and number of devices but to the size of produced and communicated data as
well as their management in terms of storage and processing. The dynamic nature
of devices, which turns on and off dynamically or connect and disconnect to net-
works, will make the landscape more complex and demanding. The open nature of
(I)IoT and the large number of stakeholders, in addition to the flexible and long
44 5 Industrial Internet of Things

Fig. 5.3 IoT reference

Application Layer
model by ITU

Management
Service support and Application support Layer

Security
Network Layer

Device Layer

supply chains in conventional service provisioning, leads to the need to accommo-

date heterogeneous “things,” devices, platforms, and services. Services for things
also need to be addressed appropriately, not only because of the limited resources of
many “things” but also because of the requirements of several services for security
and safety, including privacy protection and safe actuation that avoids problems and
accidents.
The fundamental characteristics of (I)IoT lead to requirements that need to be
met by the reference architecture. The main requirements mentioned by ITU include
interoperability, identification-based connectivity, autonomy in networking and ser-
vices, accommodation of location-based services, security and privacy, as well as
capabilities for management of things and services, including plug and play.
Figure 5.3 depicts the ITU IoT reference model, which has been introduced to
meet the above requirements. It is a typical layered model with four hierarchical
layers, specifically device, network, application and service support and finally
application layer, and two vertical layers that are crosscutting the four hierarchical
layers, defining management and security functions and properties to all hierarchi-
cal layers.
The device layer, the lowest in the hierarchy, includes the functionality of devices
and communication gateways. Considering the main interest of ITU in communica-
tions, the layer describes communication-centered functionality for the devices: (a)
devices that transmit and receive information over the communication network
directly, i.e., without using any gateway, (b) devices that communicate information
(transmitting and receiving) through gateways, (c) devices that communicate
directly without the use of the communication network but being able to communi-
cate over local networks or to form ad hoc networks, and (d) devices that are able to
selectively turn on and off functionality in order to save operating power. In regard
to gateways, the device layer includes all relevant communication technologies,
wired and wireless, such as CAN bus, Wi-Fi, Bluetooth, Zigbee, etc. Importantly,
the device layer includes protocol conversion, because devices may implement dif-
ferent protocols, and, thus, needs protocol conversion for interoperability.
The network layer provides encapsulation of device data and related protocol
conversion to network layer protocols. The layer includes functionality for the net-
work and transport layers in the OSI protocol reference model. For networking, they
5.4 IIoT Architecture 45

include control functionality for network connectivity, mobility, authentication,

authorization, and accounting, while for transport they anticipate user traffic trans-
port as well as the transport of control and management information for (I)IoT ser-
vice and applications.
The service support and application support layer includes both generic and ser-
vice/application-specific functionality (capabilities) that enable (I)IoT applications
and services. Considering the distributed nature of (I)IoT services and applications,
there exists generic functionality, such as data processing and storage, as well as
specialized functionality, per application and service, since emerging services have
different requirements, for example, smart grid operation places different privacy
requirements than an intelligent toll management system for transportation
services.
Finally, the application layer, the highest hierarchical layer, includes the (I)IoT
applications and services.
The management vertical, crosscutting layer includes both generic and applica-
tion domain-specific functionality. The generic one refers to the typical manage-
ment for configuration, topology, resource, performance, fault, security, and account
management. The application-specific one refers to functions that meet application
requirements, such as smart meter monitoring in smart grids.
Analogously to the management layer, the security vertical, crosscutting layer
includes both generic and application domain-specific functionality. The generic
functionality refers typically to functions related to authorization, authentication,
integrity and confidentiality at all layers, privacy at the application layer, secure
routing at the network layer, access control at all layers, etc. Application-specific
functionality refers to meeting application-specific requirements.
The ITU reference model document presents also a set of business models for
IoT, considering the large number of stakeholders in the area and their different
interests and goals. Importantly, these business models are developed based on the
view of network operators. The business models are based on five main business
roles that the stakeholders may have: (a) device provider, (b) network provider, (c)
platform provider, (d) application provider, and (e) application customer. As the
terms indicate, device providers are the stakeholders that provide devices for (I)IoT,
and network providers provide network systems, gateways, and connectivity for the
(I)IoT. Platform providers provide the unified, distributed IT platform with well-
defined interfaces, over which an application can be served end to end, while appli-
cation providers are the ones who provide the (I)IoT service over the platform,
networks, and devices provided by the corresponding providers. Apparently, the
application customer is the user of the (I)IoT application or service.
Based on these five business roles, ITU identifies five business models depend-
ing on the number of operators that are involved in an application and their specific
roles. Figure 5.4 shows these five models (Models 1–5), presenting the business
roles as stacked boxes – analogously to the vertical layer model – and indicating
operators with different fill patterns in the boxes; boxes (roles) with the same fill
pattern in a stack indicate that the same organization is the operator of these boxes.
In Model 1, for example, the same organization has the roles of device, network,
46 5 Industrial Internet of Things

Application Application Application Application Application

customer customer customer customer customer

Application Application Application Application Application

provider provider provider provider provider

Platform Platform Platform Platform Platform

provider provider provider provider provider

Network Network Network Network Network

provider provider provider provider provider

Device Device Device Device Device

provider provider provider provider provider

Model 1 Model 2 Model 3 Model 4 Model 5

Operator A Operator A Operator C

Fig. 5.4 IoT business models identified by ITU

platform, and application provider, while, in Model 2, one stakeholder has the roles
of device, network, and platform provider and another one has the role of the appli-
cation provider.
The Industrial Internet Consortium (IIC) focuses on similar concepts and devel-
ops a reference IIoT architecture that has several similarities with the ITU approach
and reference model. Clearly, the IIC approach to the architecture development
addresses the interests and concerns of all types of stakeholders in an integrated
way, originating from use cases and focusing on complete business models and
applications at all levels, from devices to IIoT services. IIC follows the approach
that different stakeholders who need to make different decisions have architectural
viewpoints that are at different levels of abstraction. These viewpoints enable stake-
holders to focus on the parameters of interest and develop appropriate architectures
that achieve their goals and address the problems they have identified. For this pur-
pose, IIC has identified four different viewpoints: (a) business, (b) usage, (c) func-
tional, and (d) implementation.
The business viewpoint addresses the concerns of business stakeholders, who
define and specify IIoT systems and services in their organizations or for customers.
These concerns, such as return on investment, cost of maintenance, and similar, are
addressed through a model that enables the definition of visions and values which
are translated to key objectives and then to high-level specifications of business
tasks, named fundamental capabilities. The stakeholders involved include business
developers as well as system engineers and product managers.
The usage viewpoint describes how the system is used, implementing the key
objectives and the capabilities that have been specified through the business view-
5.4 IIoT Architecture 47

point. The viewpoint is described with a model that identifies the system and its
activities, the involved parties – humans or machines – and their roles, and, finally,
tasks, i.e., actions that are executed by parties with a specific role. As tasks are the
actions in the system, they are precisely specified and described per role with, so
called, functional and implementation maps that specify the exact functions and
implementation subsystems that are necessary for a task’s complete execution. The
stakeholders involved in the usage view include not only the systems engineers and
the product managers of the related employed products but all stakeholders that are
involved in IIoT system and service specification, including the end users.
The functional viewpoint presents the functional architecture of the IIoT system,
describing its components, dependencies, and coordination, meeting the require-
ments and specifications that have been developed through the usage viewpoint. The
stakeholders involved in this viewpoint are system and subsystem developers, prod-
uct developers, and managers as well as system integrators.
Considering the focus of IIC on IIoT and the increasing adoption of industrial
control systems (ICS) within the industries of several sectors and in the operation
and management of critical infrastructure, the IIC reference model focuses on its
functional architecture of IIoT systems on the integration of ICS with classical
information technology (IT) systems in a unified, effective model that meets the
requirements of all stakeholders – as specified in the business and usage view-
points – and enables their effective decisions. The inclusion of ICS and IT in a uni-
fied model presents several challenges. Industrial control systems, the systems of
Operational Technology (OT), have been developed following a different evolution
path from typical IT systems, because of their goals and requirements that typically
include continuous operation, safety, and real-time constraints; OT systems have
been mostly developed and owned by control and operations engineers, they employ
different technologies for processing, communications, and interfaces because they
interface directly with the environment through sensors and actuators, and they are
managed by their owners independently, since they are typically part of demanding
systems and services in terms of dependability, continuous operation, real time, and
safety. As a result, their technologies, practices, and standards have evolved inde-
pendently from the ones for IT. However, the increasing capabilities offered by
advanced sensors and actuators, processors, and memories have enabled ICS to
execute highly complex operations that have been developed for complex IT sys-
tems, such as high-volume data collection and analysis, multivariable modeling and
optimization, etc. Importantly, at the same time, the increased capabilities and the
increasing complexity of ICS have led them to be more vulnerable to failures and
cyber-attacks, leading to additional functional requirements for their correct and
efficient operation.
In order to address the integration of IT and OT in a unified model, the IIC
approach to the reference architecture divides IIoT systems in five domains, each
one grouping the functionality required for a logically distinct high-level operation
of the system. These five domains are (a) control, (b) operations, (c) information,
(d) application, and (e) business. Figure 5.5, from [IIC17], illustrates the decompo-
sition of the functional representation of an IIoT system into the five domains and
48 5 Industrial Internet of Things

Business

Control flow
Data flow

Information

Application
Operations

Control
Sense Actuate

Physical system

Fig. 5.5 The IIC reference architecture functional domains

shows the data and control flow among the domains, as specified by IIC. The control
domain effectively represents the control loop realized by industrial control sys-
tems, i.e., it contains the sensors, the logic, and the actuation that constitute a plant
implemented by one or more industrial control systems. The operations domain
includes the functions that are required for the operation of the industrial control
systems in the control domain; the operation includes system monitoring and man-
agement as well as optimization for the efficient operation of the systems, especially
considering the requirements of several application domains for continuous opera-
tion, meeting real-time requirements, and achievement of low-power objectives.
The information domain is responsible for collecting data from all domains and
analyzing them to enable high-level decisions for the system, e.g., coordinating and
optimizing the end-to-end operation of several industrial control systems in the con-
trol domain. The application domain includes functionality that is application-
dependent and effectively includes the models and operation rules of the application
at hand; an important part of this domain is the set of APIs and user interfaces so that
other applications or human users can use the application effectively. Finally, the
business domain includes systems and functions that enable management and deci-
sion making at the business level, e.g., with enterprise resource planning systems
(ERP), manufacturing execution systems (MES), etc.
5.5 Basic Technologies 49

It is important to note that the IIC approach is centered around the concept of a
control plant, i.e., it addresses all viewpoints around a control loop that implements
a plant. Since control loops can be simple, with one system, or complex with mul-
tiple systems typically organized in a hierarchy, the IIC functional domain decom-
position can be applied at all levels of a hierarchy. Thus, the decomposition of an
IIoT system in the domains does not represent a layered approach as the ITU
approach, but rather a logical functional decomposition within a layer or across lay-
ers in a hierarchy. Because of this, the IIC reference architecture identifies “cross-
cutting functions” that are effectively hierarchical (or layered) IT infrastructure
functions necessary for the development of a complete IIoT application. These
functions include connectivity, distributed data management, analytics, intelligent
and resilient control, and any other application function that is necessary for the
specific application domain or use case. For example, connectivity has to be imple-
mented in a hierarchical fashion, following standards and practices, interconnecting
components within an industrial control system or across several such systems,
where each system can be viewed as a collection of functions from all five specified
domains. Observing the crosscutting functions mentioned, one can realize that they
effectively constitute a layered architecture analogous to the one by ITU. In that
respect, one can consider the IIC approach and the ITU approach as complementary,
with the IIC reference architecture being a generalization of the ITU one, since it
includes crosscutting functions analogous to the ITU layers, while it enables the
development of more detailed functional models per layer addressing complete con-
trol loops and providing support to all types of stakeholders – from device designers
to business developers – for effective decision making.
This analogy and complementarity becomes more apparent with the implemen-
tation viewpoint, which addresses the implementation details of the functional
viewpoint developed for an IIoT system. The implementation viewpoint includes all
the necessary technical and technological details that are necessary for the imple-
mentation of a complete IIoT system and its application, including system function-
ality, technological requirements, communication and network protocols, all types
of interfaces, and a mapping of the functional blocks that are specified in the func-
tional viewpoint onto typical implementation architectures, such as the three-tier
architecture (where the three tiers are the edge, platform, and enterprise) and the
layered databus architecture.

5.5 Basic Technologies

The basic technologies that enable the evolution of IIoT are the sensors, cyber-
physical systems, and the related communications and networking technologies that
enable their connectivity, among them or to other systems, including enterprise net-
works. As basic technologies, we designate the ones that are all common to all
application domains and use cases.
50 5 Industrial Internet of Things

A fundamental technology for IIoT, and IoT in general, is the technology of

RFID (radio-frequency identification) which enables the transmission of a micro-
chip’s identification information to a reader over wireless media. It is one of the first
technologies that enabled and supported the IoT concept, because RFID technology
enabled the automatic identification, monitoring, and operation execution related to
RFID-equipped tags. For this reason, RFID technology spread widely since the
1980s in the applications for logistics and supply chain management [Fuq15].
Wireless sensor networks (WSN) constitute another fundamental technology for
IIoT, considering their widespread employment in industrial automation and their
increasing deployment in critical infrastructures. The solutions for effective WSNs
need to address a large number of issues, ranging from communication reliability
and real-time requirements to low-power communication due to the deployment of
a large number of battery-operated sensors in the field. The significant advances in
the area have resulted to a large number of potential solutions and standards for reli-
able and efficient communication in various environments, e.g., WLAN, Zigbee
[Zig], Bluetooth [Blu], 6LoWPAN [Mon07, Hui11, She12], etc. Importantly, they
have led to the development of smart (intelligent) sensors, even ones that are auton-
omous and do not need recharging [Eno].
In addition to the low-level communication protocols that are necessary for con-
nectivity, additional, higher-level protocols are necessary to support distributed
computing operations and IIoT applications. Such protocols include service discov-
ery, e.g., multicast DNS (mDNS) [Che13], as well as application protocols that are
suitable for the various IIoT application domains such as Constrained Application
Protocol (CoAP) [She14], Message Queue Telemetry Transport (MQTT) [Mqt16],
and Advanced Message Queuing Protocol (AMQP) [Amq14].

5.6 Applications and Challenges

IIoT applications span a wide range of IoT application domains. Operational tech-
nology (OT) systems have become the basic computation platform for the operation
and management of most critical infrastructure. The high processing and storage
capacity of PLCs, their ability to manage real-time applications with high availabil-
ity, and their easy management by available SCADA systems have made them quite
popular as building blocks of large infrastructures beyond the manufacturing floor,
for which they were originally introduced. Today, a large portion of infrastructure is
based on industrial control systems (ICS) and makes this critical infrastructure a
potential provider of IIoT services and user of IIoT technology. The energy sector is
probably the most demanding one on the use of ICS, since the production and pro-
cessing of energy is part of a country’s heavy industry and thus, naturally, includes
large ICS platforms. In addition, ICS are used heavily in power distribution net-
works, such as the electricity network. Considering the emerging smart grids that
provide monitoring devices, i.e., PLC-like systems, to customers, it becomes appar-
ent that ICS are the main computing infrastructure in power systems end to end,
from production to consumption.
5.6 Applications and Challenges 51

A large number of distribution networks follows this ICS-based model of opera-

tion, including water distribution and management networks and water processing
sites. Importantly, oil and gas distribution networks use this technology managing
pipelines and storage tanks as well as the overall network’s operation. Transportation
also presents a significant area of application, where ICS and other cyber-physical
systems are used for traffic management, i.e., operation and management of traffic
lights, for toll payment, etc.
All these application areas of IIoT will require additional deployment and adop-
tion of components, especially cyber-physical and ICS in particular, in order to
provide the envisioned services at a large enough scale to improve the lives of citi-
zens significantly. The IIoT revolution is still at its beginning. In this evolution
process, the sectors that currently depend on ICS technology will be the forerunners
of IIoT technology and will provide the leadership in IIoT development. Currently,
the power sector and especially the electricity production, distribution, and con-
sumption processes are the most mature ones, having large bases of ICS at most
stages of the service provisioning infrastructure. Despite its maturity, the sector
presents quite challenging problems for its next generations. We present some of
these challenges here, as a sample illustration of the continuous challenges that need
to be resolved in the path to IIoT. Analogous problems exist in other IIoT applica-
tion areas as well, but the scope of our presentation is to illustrate directions and not
to enumerate problems in all application domains.
Stability and continuous operation of the power production and distribution sys-
tems constitutes a critical requirement for the development of modern economies.
Monitoring the state of the power grid system is a challenging process that is fea-
sible through advanced techniques for fault diagnosis and identification. In this
direction and considering the advances in smart grids, we need more advanced tech-
niques for fault detection and isolation in environments with distributed, intercon-
nected power generators. Detection methods based on χ2 distribution statistics
enable one to identify, with a high degree of confidence, whether the grid operates
well or if there is a fault; furthermore, fault localization and isolation can be achieved
by applying such techniques in segments of the grid. Conventional methods for
distributed fault diagnosis are limited though, because they do not address the non-
linear dynamics of the grid’s behavior, using either algebraic methods that do not
address the dynamics or sets of linear differential equations that do not address the
nonlinear characteristics. Currently, there is significant effort to develop methods
for distributed fault diagnosis taking into account the nonlinear dynamics, focusing
on nonlinear modeling, nonlinear state estimation, nonparametric state estimation,
development for statistical criteria for fault diagnosis and isolation as well as observ-
ability, and diagnosis with distributed sensor networks in the power grid [Rig11,
Rig13, Rig15, Rig17].
Power optimization of large consumers, such as large organizations or buildings
like hospitals, etc., is a significant challenge which can be addressed by IIoT. Data
collection and preparation for processing are critical to the implementation of
innovative power management and control. Actually, data preparation is emerging
as a critical, time-consuming process especially in heterogeneous environments,
requiring the adoption of new and innovative methods for data cleaning, accounting,
52 5 Industrial Internet of Things

grouping, and conversion, so that data are presented to processing in a homoge-

neous fashion. A promising direction to the optimization of power consumption is
the identification of patterns in consumption, based on the collected data. Pattern
recognition methods play an important role here in two directions, specifically rec-
ognition of patterns based on real consumer behavior and development of desired
patterns that lead to lower consumptions [Kok09, Hat11, Kou11].
Installation of IoT technologies at a large scale, as in the case of buildings, energy
networks, and production lines, requires appropriate processes, mechanisms, and
tools. The tools for deployment and configuration for the IoT, and especially IIoT,
subsystems constitute a challenge because of the high complexity and heterogeneity
of the cyber-physical systems used [Ant16]. The problem becomes more acute
when considering the limited resources of wireless embedded systems, the strict
requirements for initialization of secure wireless connections, and the requirements
for monitoring the parameters that are used for scheduling in real-time wireless
networks, such as IEEE 802.15.4e, IETF 6TiSCH6top, and ISA 100.11a [Kou16].
In contrast with small-scale deployments, e.g., in home environments, installation
processes at a large scale are error prone, despite their formalization, and lead to
installations that have significant costs for reinstallation or reconfiguration when
new devices are added or when changes are made, e.g., an office floor reconfigura-
tion. A characteristic example of a formalized, but error-prone, installation method
is the “outside-in” installation sequence, where sensors, actuators, and controllers
can be installed by technicians before the network, and IT infrastructure in the
building is installed. Clearly, it is necessary to develop effective tools for the man-
agement of IIoT resources such as wireless sensors and their networks.
IoT technologies, in general, are easily adopted in the industrial and enterprise
environments [Bi14], while the addition of wireless cards for the identification of
products and materials enables the management of their complete life cycle
[CEP10]. Thus, there is a need to integrate these smart and identifiable objects in the
industrial enterprise infrastructure and processes. Considering the heterogeneity
that characterizes industrial enterprise environments and its layered management,
from high-level ERP systems to low-level production management systems, the
integration of these devices achieving interoperability is a clear challenge. However,
when the challenge is met, the resulting system enables the flexibility of industrial
processes and their mapping and distribution on “things” of the IIoT, increasing
autonomy within the enterprise.

References

[Amq14] ISO/IEC 19464:2014 (2014). Information technology: Advanced message queuing pro-
tocol (AMQP) v1.0 specification.
[Ant16] Antonopoulos, C., et al. (2016). Integrated toolset for WSN application planning devel-
opment commissioning and maintenance: The WSN-DPCM ARTEMIS-JU Project. MDPI
Sensors.
References 53

[Bi14] Bi, Z., Xu, L. D., & Wang, C. (2014). Internet of Things for enterprise systems of mod-
ern manufacturing. IEEE Transactions on Industrial Informatics, 10(2), 1537–1546.
[Blu] Bluetooth specifications. https://www.bluetooth.com/
[CEP10] (2010, March). CERP-IoT vision and challenges for realising the Internet of Things.
CERP-IoT – Cluster of European research projects on the Internet of Things.
[Che13] Chesire, S., & Krochmal, M. (2013). Multicast DNS. IETF RFC, 6762.
[Eno] EnOcean Alliance. https://www.enocean-alliance.org
[FoF] Factories of the Future. European factories of the Future Research Association (EFFRA).
http://ec.europa.eu/research/industrial_technologies/factories-of-thefuture_en.html
[Fuq15] Al- Fuqaha, A., et al. (2015). Internet of things: A survey on enabling technologies
protocols and applications. IEEE Communications Surveys & Tutorials, 17(4), 2347–2376.
[GE17] GE Digital. Industrial Internet insights from GE Digital. https://www.ge.com/digital/
content/industrial-insights-from-ge-digital
[Hat11] Hatziargyriou, N. (2011, September 13–15). Network of the future. Presentation on
behalf of CIGRE TC at the panel session “The electric power system of the future: an interna-
tional overview”. CIGRE international symposium, “The electric power system of the future:
Integrating supergrids and microgrids”. Bologna, Italy.
[Hui11] Hui, J., & Thubert, P. (2011, September). Compression format for IPv6 datagrams over
IEEE 802.15.4-based networks. IETF RFC 6282.
[IEC16] OSI. Information technology – Security techniques – Information security management
systems – Overview and vocabulary. ISO/IEC 27000:2016. http://ww.iso.org
[IIC14] Industrial Internet Consortium. http://www.iiconsortium.org/
[IIC17] IIC. (2017). The industrial Internet of Things volume G1: Reference architecture.
IIC:PUB:G1:V1.80:20170131.
[Ind14] Germany Trade and Invest. (2014, July). Industrie 4.0 smart manufacturing for the
future.
[ISA16] ISA. (2016, December). The 62443 series of standards – Industrial automation and con-
trol systems security. ISA. http://www.isa99.isa.org/Public/Information/The-62443-Series-
Overview.pdf
[ITU12] ITU-T. Overview of the Internet of Things. ITU-T SERIES Y: Global information
infrastructure Internet protocol aspects and next-generation networks, recommendation
Y.20606/2012.
[Kok09] Kok, K., et al. (2009, June 8–11). Smart houses for a smart grid. 20th international
conference and exhibition on electricity distribution: Part 1, CIRED 2009, Prague.
[Kou11] Kourtis, G., Hadjipaschalis, I., & Poullikkas, A. (2011). An overview of load demand
and price forecasting methodologies. International Journal of Energy and Environment, 2,
123–150.
[Kou16] Koulamas, C., Giannoulis, S., Fournaris, A. (2016). IoT components for secure smart
building environments. Components and services for IoT platforms: Paving the way for IoT
standards. Springer.
[Man13] Manyika, J., et al. (2013, May). Disruptive technologies: Advances that will transform
life business and the global economy. McKinsey Global Institute www.mckinsey.com/mgi
[Mic17] Tanner, P. (2017, June 28). Micron benefits from memory mar-

ket’s faster growth rate. Market Realist. http://marketrealist.com/2017/06/
micron-benefits-from-memory-markets-faster-growth-rate/
[Mon07] Montenegro, G., Kushalnagar, N., Hui, J., & Culler, D. (2007, September). Transmission
of IPv6 packets over IEEE 802.15.4 networks. IETF RFC 4944.
[Mqt16] ISO/IEC 20922:2016 Information technology: Message queuing telemetry transport
(MQTT) v3.1.1
[Rig11] Rigatos, G. G. (2011). Modelling and control for intelligent industrial systems: Adaptive
algorithms in robotics and industrial engineering. Springer.
[Rig13] Rigatos, G. (2013). Advanced models of neural networks: Nonlinear dynamics and sto-
chasticity in biological neurons. Springer.
54 5 Industrial Internet of Things

[Rig15] Rigatos, G. (2015). Nonlinear control and filtering using differential flatness approaches:
Applications to electromechanical systems. Springer.
[Rig17] Rigatos, G. (2017). Intelligent renewable energy systems: Modelling and control.

Springer.
[She12] Shelby, Z., Chakrabarti, S., Nordmark, E., & Bormann, C. (2012, February). Neighbor
discovery optimization for IPv6 over low-power wireless personal area networks (6LoWPANs).
IETF RFC 6775.
[She14] Shelby, Z., Hartke, K., & Bormann, C. (2014, June). The constrained application proto-
col (CoAP). IETF RFC 7252.
[Zue10] Zuehlke, D. (2010). SmartFactory—Towards a factory-of-things. Annual Reviews in
Control, 34(1), 129–138.
[Zig] ZigBee specifications. http://www.zigbee.org/
Chapter 6
Security and Safety

6.1 Introduction

The Internet of Things (IoT), including the Industrial Internet (IIoT), refers not only
to the connectivity of systems and devices but to the related applications and ser-
vices that provide monitoring and control of complex systems and services. The
application domain spans a wide range of industries, from health to industrial con-
trol and from transportation to surveillance systems. Its expansion and growth
incorporate several technologies and disciplines, such as electronics, embedded net-
works, hybrid systems, and control. The inclusion of information technology (IT) as
well as operational technologies (OT) creates a challenge for the development of
systems and services that are technologically interdisciplinary. The resulting chal-
lenges to integrate these technologies in new design methodologies for robust and
effective IoT systems and services are significant. Currently, even the terminology
used by different stakeholders presents challenges and inconsistencies to the com-
mon understanding of properties and goals of IoT infrastructure and applications.
Considering the targeted applications and services of IoT, in this chapter, we
address security and safety of IoT systems and services with an approach that spans
from systems to applications (services or processes) in a unified way, using termi-
nology that originates from computing, networking, and control, since these disci-
plines constitute the main pillars of IoT technologies in all IoT application domains.
This approach is consistent with the reference architectural models of both ITU and
the Industrial Internet Consortium, as presented in Chap. 5. For convenience, we
address security in this chapter following the ITU model, which divides security
mechanisms in two parts, one for generic security and one that is application depen-
dent; we use the terms application dependent and process dependent
interchangeably.
IoT applications, in general, collect data through sensing devices, process this
data, and take actions that range from sending notification and raising alarms to tak-
ing actions through actuators on physical systems. A simple generic model for this

© Springer International Publishing AG 2018 55

D. Serpanos, M. Wolf, Internet-of-Things (IoT) Systems,
https://doi.org/10.1007/978-3-319-69715-4_6
56 6 Security and Safety

Control Center
(C)

Control commands Data acquisition

Control
actions Measurements
Device
Actuators Sensors
(D)

Fig. 6.1 Control loop

operation is the model of the control loop that is used across many application
domains and is depicted in Fig. 6.1. In this model, a device D is controlled by a
control center C. Measurements of the parameters of interest are collected from D
through sensors and delivered to C which makes the necessary calculations and
takes the necessary decisions and actions for the application; if the application
requires automatic actions, C sends the necessary commands to actuators that con-
trol D. The model is generic and covers application across domains ranging from
health to transportation and from aerospace to manufacturing. In a health applica-
tion, for example, sensors measure patient parameters, such as temperature and glu-
cose levels, and send them to a monitoring program – analogous to the control
center – and decisions are made depending on the application; a message may be
sent to attract a patient’s or a doctor’s attention, or an insulin pump may be opened
to administer more insulin. In a manufacturing floor, sensors may detect the arrival
of a component and send the data to a control center which, in turn, sends the appro-
priate commands to the machine that will process the component accordingly.
The control loop model shown in Fig. 6.1 is implemented on a computational
platform that has a different structure from the one indicated in the control model.
Figure 6.2 shows a typical hierarchical computational structure for industrial sys-
tems, an important class of IoT systems, showing how the computing systems, net-
works, sensors, and actuators are typically used to implement the operational
computing infrastructure of the control loop. Sensors and actuators are attached to
the controlled device (D in Fig. 6.1), programmable logic controllers (PLCs) imple-
ment simple controls – one per PLC typically – and the supervisory control and data
acquisition (SCADA) system implements the control loop for the complete process,
also denoted as plant. The PLCs in the structure are simple industrial computers,
and their number differs according to the application. In a smart grid, for example,
different PLCs may take actions locally per transformer, while SCADA controls the
6.1 Introduction 57

Fig. 6.2 Hierarchical
computing structure for
Supervisory Control
control loop and Data Acquisition
(SCADA)

Network

PLC PLC PLC

Network

Valve Pump

complete smart grid; in a water management system, a different PLC may control
each pump, while SCADA controls the water system of an industrial site.
In this environment, there are several properties we want to achieve. From the
control point of view, these properties are typically safety properties. For example,
we want to avoid overloading of a smart grid, to avoid the overflow of a fluid tank,
or to avoid overdose of a pharmaceutical substance that is automatically adminis-
tered to a patient. These properties can be violated because of several reasons. A
programmer may have introduced a bug in the program, the requirements of the
system may have missed a condition that should had be taken into consideration, the
middleware of the system may give the wrong priorities to control processes, or,
simply, a malicious party may attack the system and cause it to take the wrong
actions.
The safety requirements for applications are typically expressed as requirements
on the control loop which implements applications. These expressions are based on
assumptions about properties of the infrastructure on which the application is imple-
mented. For example, an HVAC control system assumes that the temperature mea-
surements that are input to the system are correct within some approximation. This
implies that the safety properties are based on assumptions for data integrity that
need to be satisfied by the infrastructure. In general, safety requirements include
infrastructure security ones, such as integrity, implicitly or explicitly. A typical
explicit security property is the protection of personal information in a health man-
agement system. Thus, it becomes clear that security is a requirement for safety as
well, since data integrity is necessary at least.
58 6 Security and Safety

IoT technologies involve several stakeholders, including vendors, service pro-

viders, regulators, and customers. Although the interests of the independent stake-
holders are different and, thus, the security requirements they place on IoT
technologies may differ, there is a set of core security requirements that, in general,
addresses the requirements of all different categories of the stakeholders. This set of
requirements includes (1) confidentiality, (2) integrity, (3) authentication, (4) access
control, (5) non-repudiation, (6) dependability, (7) safety, and (8) privacy [Ser13].
Confidentiality is the property that provides protection of data, stored or trans-
mitted, from being disclosed, while integrity enables the confirmation (verification)
of the correctness of the related data. Authentication enables the identification of
any party involved in a transaction, whether producing, processing, transmitting, or
receiving data. Access control ensures service provision to authorized users, while
non-repudiation disables participants to transactions to deny actions or their partici-
pation. Dependability requires provision of system and service functionality with
specific properties such as continuous service even in the presence of errors and
failures, meeting specific real-time requirements, etc. Safety is a service and process
requirement that warrants service provisioning so that there is no hazard to users.
Finally, privacy protects personal information from access by unauthorized actors.
Scientific and engineering methods and techniques to meet these requirements
are known, in general, because such requirements have been long addressed in sev-
eral IT systems in a wide range of application domains. However, meeting the
requirements in the IoT and IIoT context with OT characteristics requires new
approaches, because of several additional factors. These factors include the models
of component failures, the available resources for security provisioning, as well as
the profile of attackers, including their potential resources. These factors are strong
differentiators in the process of security provisioning in the IoT and IIoT context,
for several reasons. First, embedded and CPS systems have already been deployed
in significantly larger numbers than the non-embedded (typical IT) systems such as
servers, laptops, etc. Second, most of these systems are resource limited in terms of
computational, communications, and power resources, and their manufacturers
place strong low-cost requirements in order to penetrate large consumer markets. As
a result, these systems are deployed in various environments, including hostile ones
where malicious users get access to these systems for unspecified lengths of time
and with unspecified capabilities to tamper with them. A final reason is the strict
requirement for safety in several domains such as automotive, industrial, aeronau-
tics, etc.
These differentiating factors of embedded systems place significant demands on
their security, because their large deployment numbers and the diverse operating
environments, with many unknown or unanticipated characteristics, lead to a large
number of potential attackers with varying capabilities. In addition, many applica-
tion domains place security requirements that are relevant to safety, dependability,
and privacy, as in the case of transportation systems, medical systems, surveillance,
etc. The necessity to meet all these requirements on systems with limited resources
and low targeted cost leads to highly challenging problems and the need for low cost
technologies that achieve the required goals.
6.1 Introduction 59

Privacy Safety Layer 2: Properties

Security Dependability Layer 1: Mechanisms

Fig. 6.3 Security property layers

In order to identify the requirements and mechanisms that are required to provide
the necessary security properties in the IoT and IIoT context, we follow the layering
shown in Fig. 6.3, which has been introduced in [Ser13]. Figure 6.3 defines our view
of the relationship between application and process properties, such as safety and
privacy, and security and dependability mechanisms which are provided at the sys-
tem level and are used as primitives to provide the application and process
properties.
The depicted layering is based on our approach to differentiate system level
properties, such as secure storage, secure communication, tamper resistance, etc.,
from properties that are required and provided at the application level. In this
approach, we consider that (embedded) systems and their interconnections are built
to operate resiliently overcoming failures, accidental or malicious, that lead to infor-
mation loss, leakage, and availability. Dependability mechanisms focus more on the
aspects of reliability and availability considering accidental failures, using probabi-
listic models for the failures, while security mechanisms focus on the provision of
alternative properties, e.g., confidentiality, authentication, availability, etc., based
on defined malicious attack models. Although some dependability and security
properties, such as the availability of information, are common between the two
disciplines, others, such as confidentiality or continuous operation, are complemen-
tary. In general, dependability is complementary to security, because an attacker can
insert faults and failures – analogously to launching attacks on security mecha-
nisms – that the dependability mechanisms cannot recover from. Clearly, the com-
bination of dependability and security mechanisms at the system level provides
trusted platforms that are both secure and available under accidents and attacks.
Safety and privacy are often described as security requirements in many applica-
tion domains, although they are different from the typical security considerations in
many ways. Typically, privacy protection and safety are requirements for processes,
applications, and services, rather than for generic systems. In our approach, privacy
and safety are dependent on security, because they employ security mechanisms for
their implementation, such as data integrity and confidentiality. Interestingly, safety
and privacy are overlapping, because privacy is a safety issue in some contexts, such
as the financial transactions. It is important to note that, as Fig. 6.3 indicates, secu-
rity and dependability are requirements for privacy and safety. If security mechanisms
60 6 Security and Safety

are lacking, an attacker can violate privacy by easily collecting data or can alter
processes and applications, leading to unsafe conditions.
The threat model we consider for IoT systems is one that includes both compu-
tational attacks and data attacks. Computational attacks include all malicious actions
in a computing system that affect the correct execution of a program and/or lead to
information leakage. Data attacks constitute all attacks on input or communicated
data. We extend the concept of data attacks to include false data injection attacks,
which are malicious interventions that input inappropriate (illegal) data to a system.
False data injection (FDI) attacks are an emerging class of attacks to IoT systems,
which do not attack the IoT systems themselves but input wrong data to a control
system in order to lead it to a wrong decision. In that respect, they are mostly safety
attacks. For example, in an HVAC system, a false data injection attack would be to
input a higher temperature to the system, instead of the correct measure, in order to
lead it to lower the temperature further. Clearly, this type of attacks can lead to haz-
ardous conditions that may endanger processes and systems, even human life.

6.2 Systems Security

IoT systems are embedded computing systems that employ architectures analogous
to general-purpose ones. A typical structure of an IoT system is shown in Fig. 6.4,
where the system contains four main subsystems: (i) processing, (ii) memory, (iii)
input/output, and (iv) power. In general, a secure system requires protection as a
whole in addition to protection of all its components individually. The specific
requirements are placed depending on the operational environment and the expected
capabilities of attackers. In a surveillance system, for example, optical sensors
(cameras) need to be secured individually, but the whole network needs to operate
dynamically in case individual cameras are compromised or destroyed.
The security of stand-alone systems is achieved with several levels of protection
that include physical and hardware security as well as trusted computing platforms.
Anti-tampering techniques enable different levels of physical protection ranging

Fig. 6.4 Organization of a

Bus
typical IoT system Processor Memory

I/O

Power (Battery)
6.2 Systems Security 61

from tamper evidence to tamper response and tamper resistance and are employed
accordingly depending on the security requirements of the system and its opera-
tional environment. Techniques for tamper evidence simply indicate whether a
device has been tampered with. Tamper-response methods combine tamper detec-
tion with tamper reaction, where appropriate actions are taken after tamper detec-
tion; for example, they destroy stored sensitive data. Tamper resistance methods
prevent tampering with devices and protect any sensitive data in the device from
attacks.
Anti-tamper technologies have been developed to protect systems after their
deployment, so they need to address physical and hardware attacks of attackers with
variable capabilities in a wide range of hostile environments, especially for critical
applications such as surveillance. They need to combine physical as well as algo-
rithmic mechanisms. Traditional encryption of data, for example, is not a sufficient
solution to data protection nowadays, especially in limited-resource systems, where
encryption can be overcome with simple attacks. Side-channel attacks have changed
the attacks on cryptosystems exploiting physical parameters of the implementations
of cryptographic algorithms, such as timing and power consumption, rather than
attacking the algorithms themselves [Koc96, Koc99, Qui01] or introduce faults dur-
ing cryptographic computations [Bar06, Joy09].
Complex hardware systems such as processors and micro-controllers are suscep-
tible to physical and hardware attacks similarly to dedicated circuits, such as cryp-
tographic circuits [And96, Bly93]. Defenses against such attacks require dedicated
hardware, specialized design techniques, or even new architectural concepts. For
example, a sensitive program can be protected from attacks by storing it in a special
design of execute-only memory that allows instructions stored in memory to be
executed only and does not allow any other manipulation [Lie00]. Encrypted buses
protect data from leakage during data transfers between a processor and its memory
[Bes81, Kuh97]. Decay caches can protect from side-channel attacks avoiding
cache information leakage [Ker08].
Anti-tampering techniques protect against attacks after system deployment. New
business environments can drive embedded systems insecure by planting hardware
Trojans during the design and manufacture phase [Jin10].
Embedded and cyber-physical systems, in general, are widespread and have
attracted a large range of attacks [Rav04]. Defense against them requires a combina-
tion of software and hardware techniques, in order to cover all potential attacks.
This is especially important in emerging cyber-physical and IoT systems, which
include operating systems or specialized middleware. More complex programmable
systems require adoption of such methods as secure booting [Arb97] to establish
system integrity, process isolation, and process level attestation techniques [Mic11]
to protect running processes as well as techniques for context switching, exception
handling, inter-process communication, and memory management [Lie03, Gar03].
Overall, the increasing programmability of these systems requires appropriate soft-
ware security techniques. Software techniques also offer a cost advantage over
hardware techniques. Furthermore, the combination of software techniques with
62 6 Security and Safety

trusted computing modules [Pea02] enables the development of trusted computing

platforms for applications and services.

6.3 Network Security

Secure communication requires encryption and authorization mechanisms as well

as a secure routing method in a network. Traditional encryption schemes, such as
AES [AES01], RSA [RSA78], etc., provide a high level of security, as has been
proven in general-purpose computing systems, but they are quite demanding in
computational and memory resources. Clearly, they are becoming more viable can-
didates for adoption in environments where embedded systems obtain increased
computational resources. However, today, they are still too demanding computa-
tionally for most embedded applications and services. Elliptic curve cryptography
provides a promising solution to IoT environments, because it requires lower com-
putational resources than algebraic public key cryptography while providing a high
level of security [Miller1986]. Importantly, significant effort is spent to develop and
standardize appropriate algorithms for cryptographic primitives for IoT environ-
ments, taking into account their characteristics. The development of the Secure
Hash Algorithm-3 (SHA-3) by NIST is a significant step in this direction, providing
a family of hash functions and extendable output functions that are useful for pseu-
dorandom bit generation, key derivation, and digital signatures in IoT environments
[Mor15].
Sensor networks are an important class of IoT subnets that need special atten-
tion, because they usually form ad hoc sub-networks with large number of nodes
that have very limited computational resources. Thus, sensor network protocols
often need to satisfy stricter performance requirements than more complex embed-
ded systems. These limitations typically lead sensor networks to implement crypto-
graphic mechanisms at the link layer. In such limited environments, a good
encryption strategy is to use mechanisms with different complexity, depending on
the value of the communicated information [Zhu03].
Key management is a critical component of secure IoT communication because
keys are the base of cryptographic mechanisms. If key management has weak-
nesses, keys will be compromised (disclosed or leaked) leading to ineffectiveness of
any cryptosystem independently of its strength. The use of global communication
keys provides a solution, but such keys cannot be predefined in networked systems
usually, because the security of the network can be easily compromised. This leads
to the necessity to develop and adopt effective methods to generate and distribute
keys. There exist such effective methods mainly using temporary global keys and
random key distribution. One such method uses temporary global keys and a global
permanent key to establish a main key; then, it destroys the global key in order to
avoid key leakage, i.e., the main risk with global keys [Per04]. In an alternative
approach, one can use random key distribution. In this case, the system uses a large
number of keys and performs communication choosing random subsets of keys.
6.3 Network Security 63

When the key set sizes are chosen appropriately, all network end points of a network
can communicate successfully [Cha03].
Networked systems, especially through the Internet, need to ensure that data are
being communicated only among authorized users and processes and that these
exchanged data are “legal.” This is usually achieved through the use of firewalls,
which are typically implemented at the network and application layers, in end point
systems or in the network infrastructure [Bol95]. IoT systems typically have very
well-defined communication needs, and thus, firewalls can be easily configured to
allow strictly the limited type of legitimate communication. The decision about
where the firewall should be implemented, i.e., at the network or application layer,
at the end system, or in the network, depends on the end point system, the network
and their available resources, as well as on the network topology. For example, ad
hoc networks need protection at the node level, while more centralized systems can
rely more on network level protection [Sli02].
Denial-of-service (DoS) and distributed denial-of-service (DDoS) attacks are a
significant threat against IoT systems or exploiting IoT systems. DoS attacks over-
load resources, such as processor, memory, and network, of the targeted system, in
order to prevent it from performing its intended functionality or serving its users.
In general, there are two basic types of DoS attacks [Hus03]. The first type of
attack exploits vulnerabilities, hardware or software, by sending carefully con-
structed packets to the target system; the typical goal is to crash the target system.
Often, such vulnerabilities are exploited because systems are not patched. This
makes IoT systems especially vulnerable to these attacks, because many IoT sys-
tems are not configured to update their software automatically and a wide popula-
tion of users is not sufficiently aware of the risks and actions they need to take to
protect their systems accordingly
In the second type of attacks, the distributed denial-of-service (DDoS) ones, a
large population of compromised systems create vast amounts of network traffic
toward a victim system; this traffic is combined with legitimate traffic as well. The
overload of the aggregated arriving traffic at the target system overloads its resources
and renders it incapable to serve its legitimate users. The recent incident of the Mirai
botnet attack [New16] demonstrated clearly that IoT devices are vulnerable to mal-
ware injection and they can be effectively used to launch DDoS attacks; in the Mirai
case, they attacked an Internet directory service, causing significant and costly dis-
ruptions to Internet connectivity worldwide.
DDoS attacks are difficult to stop because they exploit shared network services
that are accessed by all systems connected to a network. The current version of the
Internet Protocol (IPv4) allows systems to send IP packets with arbitrary values in
the source IP address field, making it difficult to identify sources of offending IP
packets in many attacks [Wan07]. Current efforts to defend against DDoS attacks
are usually based on intrusion detection and traceback schemes for detection, filter-
ing, and tracing of an attack [Pen07]. Intrusion detection employs signature- and
anomaly-based detection techniques [Cab01, Wan02], while packet marking [Bel03,
Sav01] and packet logging [Sno02] are used for attack traceback.
64 6 Security and Safety

6.4 Generic Application Security

Interconnected IoT systems provide the infrastructure for distributed applications

and services. Currently, the vast majority of deployed and emerging applications
follows the client-server model, where remote devices (clients) are connected to
servers or the cloud, in general, to deliver information, such as collected data and
alarms. The servers typically collect data, monitor the operation and processes of
IoT connected devices, and send to the devices control programs or data, in order to
adapt their operation accordingly. For example, a medical device may collect infor-
mation about a monitored patient, deliver it to a centralized server, and receive from
the server tuning information to adapt its operation appropriately, e.g., to change the
frequency of collected data or to change an algorithm used in its local data process-
ing. A connected car may send reports and alarms related to engine operation and
receive a prompt to execute a more detailed test, in case of an alarm or suspicious
data degradation.
Considering the emerging application and service models, it becomes clear that
IoT systems are distributed systems that execute coordinated processes, where each
process is typically a control loop, i.e., a process that, in general, receives sensor
data and transmits actuator commands. When the client-server model is adopted, the
IoT devices execute simpler control loops, while servers execute hierarchically
higher level operations, when they are not simply collecting data. In the Industrial
IoT, this hierarchy is expressed through the programmable logic controllers (PLCs)
as the local, simpler, and lower level devices (clients) and the supervisory control
and data acquisition (SCADA) system as the centralized, higher level server system
that monitors and coordinates the complete supervised process.
Based on this hierarchical application model, we consider two levels of distrib-
uted application for security purposes. The first level, generic application security
support, is the one that provides generic services to the IoT environment, such as
system update and upgrade, while the second one, process, is the one that imple-
ments the specific process for the specific IoT system, e.g., health, car, industrial,
etc.
Generic application security support includes mechanisms to defend against
attacks to distributed denial-of-service, secure upgrading, etc. Distributed denial-of-
service solutions exploit mechanisms at the network layer, as described in the previ-
ous section, extending them where necessary to include specifics from the application
configuration at hand, such as the location of the servers.
Upgrading and patching IoT systems constitute another challenge that requires
inclusion of security mechanisms, because upgrading and patching open systems up
to security risks. The functionality for upgrading and patching is necessary for many
reasons; software bugs of deployed software need to be fixed, and new features may
need to be added to an IoT system’s functionality. However, the ability to transmit
code to an IoT system raises the risk that one may attack the system by inserting
malicious code instead of the legitimate, intended code. Thus, security mechanisms
need to be included in the upgrading services to warrant the secure and safe
6.5 Application Process Security and Safety 65

u pgrading of the IoT systems. There are several approaches to this challenge. One
can limit or prevent the ability to upgrade software components that manage critical
system resources in highly hostile application environments. Alternatively, in safer
environments, strict access control mechanisms can be used to enable upgrades of
different software components by different operators. Mobile code transmission
may be prevented, while wired code transmission may be allowed when connectiv-
ity is in a controlled environment. In general, remote management of systems, espe-
cially IoT systems with limited resources, requires a secure architecture that
addresses the operational environment as well as the profiles of the potential
attackers.

6.5 Application Process Security and Safety

Application processes, such as control processes in an industrial environment, are

programs that execute the necessary code to calculate the required outputs and
implement the process’s actions. For example, in an HVAC system, an application
process may take as input a request to increase the temperature of the controlled
environment, and, as a result, it will calculate the necessary increase for the extracted
hot air temperature and its volume and will control and adjust the related actuators
accordingly to achieve the result. In a more complex environment such as a smart
grid, an identified need or a request to add power to the network will lead to the
calculations for the necessary power, the identification of the appropriate generators
to activate and, finally, the control of the appropriate actuators that will add the
generators to the grid. Such application processes in the (I)IoT environment have
safety requirements, which are typically expressed as properties that need to be met;
for example, in the HVAC system, the temperature of the hot air needs to be within
a specified temperature range. Clearly, security of the involved computing and net-
work systems is a prerequisite for meeting the safety requirements; a compromise
of these subsystems can lead to wrong calculations and, thus, to wrong actions that
violate the safety properties that are required to be met.
Provision of security and safety in (I)IoT environments is one of the areas where
the interdisciplinary nature of IoT expresses itself: safety requirements are applica-
tion dependent and are set, in most cases, by the engineering of the controlled sys-
tems, while security – a prerequisite of safety – requires methods of computer and
network security, since the IoT systems themselves are distributed computing sys-
tems effectively. Bringing all safety and security requirements together is a chal-
lenge that has motivated a lot of research and development work recently and will
require significant effort in the future to lead to effective solutions that are easily
deployed in the field.
The most promising integrated approach to safety and security, from a computa-
tional perspective, is to view the problem as a verification and monitoring one.
Since application processes are implemented with programs and safety properties
are set by the application designers, e.g., control engineers, one can view the p rocess
66 6 Security and Safety

of developing the application programs as one where the application designers pro-
vide the specifications of the application, including the safety properties, and then
the software is developed accordingly to meet these specifications and be secure
from vulnerabilities overall. In this fashion, the safety and security problem becomes
a verification and monitoring problem: first is the verification of the produced appli-
cation software, i.e., that it meets the set requirements, and second the monitoring
of the execution of the verified program in order to ensure that it is not altered and
executes as expected, based on the specification.
This approach is a behavioral approach to safety and security, since it is based on
the specification of the application process. In this context, application behavior is
defined by the executable specification that is the starting point of the approach, and
this is the way the term is used in the remainder of this text.

6.6 Reliable-and-Secure-by-Design IoT Applications

The concept of secure-by-design applications is an extension to the principle of

correct-by-construction programs introduced half a century ago [Dij67]. The chal-
lenge posed by IoT applications is that IoT systems typically include a cyber-
physical subsystem that interacts with the environment. Thus, in contrast to the
original concepts developed for behavioral models of programs with discrete and
linear characteristics, the models for cyber-physical and IoT systems need to accom-
modate continuous and nonlinear characteristics. A model of the environment is
also necessary but challenging, because there exist uncertain environmental varia-
tions that affect the behavior of physical subsystems; furthermore, it is necessary to
model the environment at different levels of abstraction.
The development of reliable and secure-by-design applications has attracted the
attention of several efforts, which focus on the development of effective program-
ming language environments. Ur/Web [Chl96] is a language that enables develop-
ment of reliable and secure web applications by design. For security, Ur/Web
ensures that the produced application does not have vulnerabilities, such as for code
injection attacks and SQL injections, while for reliability it ensures that the applica-
tion will not crash during generation of web pages, it will not produce dead intra-
application links, etc. The language guarantees these reliability and security
properties through an enriched type system based on dependent. In this fashion, Ur/
Web achieves an important result: it provides a unified web model, where a pro-
grammer develops web applications in a single programming language that can be
compiled to other web standards. In another effort, the Jeeves language focuses on
run-time, enabling enforcement of security policies and guarantees that programs
do not violate security properties by design [Yan12]. Analogous efforts have been
made to apply these approaches in the cyber-physical systems application domain.
The ROSCoq framework [Ana15] employs the Coq proof assistant [Ber04] to model
cyber and physical resources of robots through an extended logic of events and then
to prove various properties of the model. VeriDrone [Mal16, Cha16], a reasoning
6.7 Run-Time Monitoring 67

framework also developed in Coq, ensures security of cyber-physical system mod-

els at different but independent levels, i.e., from high level models to C
implementations.

6.7 Run-Time Monitoring

Run-time monitoring systems for security can be classified based on two parame-
ters: (i) the method that describes the behavior, i.e., profile based or model based,
and (ii) the method that compares the behaviors, i.e., matching to bad behavior or
deviation from good behavior. This leads to a classification with four classes, as
shown in Fig. 6.5.
Profile-based approaches monitor parameters of the observed system and build a
profile of system operation. Class 1 monitoring systems that detect attacks by
matching with bad behavior (Class 1 in the figure) typically use statistical methods
and machine learning methods to build profiles of bad behavior and statistical pro-
files of attacks [Hod04, Val00]. They are more robust than model-based systems
(Class 2 systems) because machine learning typically generalizes from the collected
data, but they suffer from high false alarm rates, and they do not provide rich infor-
mation for diagnosis when an alarm is raised. Systems in Class 3, which detect
deviations from good behavior, usually build a statistical profile of good behavior
and detect deviations from that [Kim04, Lak05].These systems are actually more
robust than the ones in Class 1, because they do not depend on any past information
of attacks and, thus, they raise alarms when new attacks are launched, because all
deviations from good behavior are detected. However, not only do they provide
limited diagnosis information, i.e., only that something extraordinary has happened,
but they suffer from high false alarm rates, because the deviation may not be mali-
cious or accidental, but it can also be normal but just out of the statistically accepted
profile behavior.
Model-based monitoring systems, Class 2 and Class 4 systems, use a model of
the behavior of the monitored system. Such systems are popular in highly secure

Fig. 6.5 Classification of Behavioral description

run-time security
monitoring systems Profile based Model based

Bad behavior Class 1 Class 2

matching

Behavioral
comparison

Good behavior Class 3 Class 4

deviation
68 6 Security and Safety

environments, where successful attacks have high cost. Because they use a behav-
ioral model of the observed system, these monitors provide rich diagnostic informa-
tion when alarms are raised, in contrast to profile-based monitors. Despite this rich
information though, Class 2 monitors are limited because they can detect only
known attacks; this originates from their bad behavior models which are already
known by definition, i.e., the attacks exist [Pax99, Roe99]. Signature-based systems
are typical examples in this class. Class 4 monitors detect deviations from a good
behavior model [Wat07, Gol07] and thus provide even higher diagnostic informa-
tion, because there is adequate knowledge of the exact problem, e.g., the exact
instruction, that led to a detected deviation. However, the execution overhead of the
models of good behavior poses limitations to run-time system performance.

6.8 The ARMET Approach

A promising approach that addresses safety and security in a unified way in IoT
systems and cyber-physical systems is the ARMET approach [Kha17]. ARMET is
based on three basic concepts: (i) we can build secure-by-design systems, (ii) we
can monitor these systems at run-time for correct operation to detect attacks or fail-
ures, and (iii) when there is a failure or an attack, we can have plans to recover,
depending on the problem and how much information we have about it. ARMET
has been developed focusing on industrial control systems, but it is applicable to
other IoT systems as well, since their software complexity is comparable to that of
industrial control systems.
With the ARMET approach, an IoT application is developed starting from an
executable specification which is provably consistent with the safety properties set
for the application. From this executable specification, the application code is
derived. Given the executable application specification and the application code for
the target system, ARMET monitors the behavior of an application while it exe-
cutes, by comparing its observed behavior to the expected behavior based on the
application’s specification; to achieve this, a middleware executes the executable
specification in parallel with the application execution on the IoT system and calcu-
lates predictions of the application’s behavior. Figure 6.6 shows the structure of the
ARMET middleware system, which is composed of several components: (i) the
run-time security monitor, (ii) the diagnosis module, (iii) the recovery module, (iv)
the trust model, (v) the adaptive method selection module, and (vi) the backup mod-
ule. The run-time security monitor is a critical component of the middleware, which
takes as input the executable specification of the application and the state of the
system that executes the code of the application. The monitor observes the behavior
of the application execution, and, in parallel, it predicts the state of the application
execution by executing its specification; the specification execution defines the
expected “good behavior” of the application and, optionally, known “bad behavior”
of the application that includes known attacks. Comparing the predictions with the
observations, the monitor can detect deviations that indicate a failure of the
6.8 The ARMET Approach 69

ARMET

Recovery Adaptive method selection

Diagnosis Trust model Backup

Run-time security monitor

(RSM)

Application Application
specification code

Fig. 6.6 The ARMET system

a pplication or an attack. When such a detection is made, ARMET proceeds to a

stage of diagnosis, in order to identify the failure or attack based on a trust model
that it includes. After the diagnosis phase is concluded, all available information is
used by the recovery module. Based on the diagnostic information, the recovery
module chooses an appropriate adaptive method for recovery and enables the sys-
tem to recover, taking into consideration previous states, as stored by the backup
module. It is important to note that the system will operate under all scenarios of
failures and attacks, even unknown ones, i.e., failures and attacks that have not been
anticipated and are not included in the trust model. In a worst-case scenario, when
no useful information is provided by the diagnosis module, the system will recover
by returning to a previous clean state. Furthermore, the approach is based on one
assumption: the executable application specification is executed in a safe environ-
ment and cannot be attacked, i.e., its predictions are always correct; although this
may seem as a strong assumption, conventional trusted platforms enable the devel-
opment of a low-cost IoT platform that meets this requirement and makes the
assumption realistic, such as Intel SGX [Cos16] and ARM TrustZone [ARM05].
The process to develop executable specifications and prove its properties is a
typical program verification process that can be implemented with various existing
tools that enable automated or semiautomated proofs. In the case of ARMET
[Kha17], the process is based on Fiat [Del15] and employs deductive synthesis to
develop reliable-and-secure-by-design industrial control applications through inter-
active stepwise refinement of declarative specifications; cyber and physical resources
are included as first class models, and nonfunctional properties, such as security and
performance, are modeled integrated with functional properties. Apparently, the
70 6 Security and Safety

in_pump

h wh
out_pump

Fig. 6.7 Water tank

executable specification can produce automatically executable code for the target
system, IoT or industrial.
The ARMET run-time security monitor (RSM) successfully identifies inconsis-
tencies between predictions, produced by the execution of the specification, and
observations of the application code execution, because of its executable specifica-
tion language [Kha15]; importantly, the predictions are generated automatically.
The run-time security monitor (RSM) is the first one to be formally proven as sound
and complete [Kha15]; the proof means that the monitor is also free of false alarms
(detections), an important, desirable property in practical systems, where false
alarms lead to lost resources that are used to explore the false alarms. Importantly,
ARMET’s specification language allows the specification of faulty behaviors as
well as attack plans, which can be used by the monitoring system for threat
detection.
The ARMET approach is based on the concept that a system can be specified
with an executable specification. Based on an appropriate functional specification
for a system, one can express the safety and security properties that the system
should meet as conditions of the specification and include them in the specification
as well. As an example, let us consider the case of a water tank which has a height
h, as shown in Fig. 6.7, and two pumps that are controlled, one for filling the tank
with water, denoted in_pump, and one, out_pump, for draining the water out; each
of the two pumps has only two possible states, i.e., open or closed. Furthermore, we
assume that there is a sensor that measures the water height, denoted wh, in the tank.
We want to have a water management system, where a user issues commands to
pour water or drain water from the tank. For simplicity, we consider that a user can
perform three actions, FILL, DRAIN, or NOTHING, and that the system operates
in cycles, synchronously with a clock. So, during every cycle (clock tick), one
action can be performed. A FILL action implies that in_pump opens, out_pump
closes, and for this one time unit water is poured in the tank. A DRAIN action
means that in_pump closes, out_pump opens, and for this one time unit water drains
out of the tank. When the action is NOTHING, then both pumps are closed and the
state of the tank remains the same. In an environment like this, an obvious safety
property is that we do not want the tank to overflow under any conditions.
6.8 The ARMET Approach 71

«Enumeration»
Action
- FILL
- DRAIN
- NOTHING

«StereoType»
WaterTankSpec
- water_level : Integer :=0
- SENSOR_ACCURACY : Real := 0.01
- FILL_RATE : Integer := 1
- DRAIN_RATE : Integer := 1
- TANK_MAX : Integer := 10
+ readValue (reading : Integer) : void
+ doAction (water_level : Integer ) : Action

context WaterTankSpec ::readValue(reading : Integer)

+ pre: reading – self_water > SENSOR_ACCURACY
+ post: self_water = reading

context WaterTankSpec :: doAction(water_level : Integer) : Action

+ pre: forall( a : Action | (a = FILL implies water_level + FILL_RATE <= TANK_MAX) and
(a = DRAIN implies water_level - DRAIN_RATE >= 0))
+ post: result = FILL implies self.water_level= old(self.water_level)+ FILL_RATE and
result = DRAIN implies self.water_level= old(self.water_level) - DRAIN_RATE

Fig. 6.8 Water tank control executable specification

Figure 6.8 shows one executable specification, written in UML, which imple-
ments the three defined actions, assuming that each action FILL or DRAIN has as a
parameter an integer value for the variable water_level, which specifies the target
height of the water that the user wants to obtain; furthermore, the specification
ensures that the water tank never overflows. In the specification, the three actions
are defined in enumeration: SENSOR_ACCURACY defines the measurement
accuracy of the reading sensor for the water level in the tank, FILL_RATE is the
72 6 Security and Safety

incoming water rate through in_pump, and DRAIN_RATE is the rate of the outgo-
ing water when out_pump opens. TANK_MAX is the height h of the tank.
When an action is issued by the user, the system first takes a reading of the water
level with the sensor, as specified in readValue, and identifies whether the target
water height differs from the measured height within the sensor’s accuracy bounds.
If the target height is different, then the corresponding action is performed, pouring
water in or draining water out until the target height is achieved. The safety property
is enforced, because of the precondition that is expressed in doAction(), which
ensures that a FILL action is performed when its result leads to a water height that
is less or equal to TANK_MAX.
Since RSM is sound and complete, it is proved that it will detect all computa-
tional attacks on the application. This means that any attack that influences the
execution of the application and leads to wrong calculations will be detected. This
has been confirmed with several computational attacks [Kha17]. Importantly, RSM
captures a wide range of false data injection attacks as well. For example, if an
attacker wants to overflow the water tank of the example and alters the reading of
the sensor to a lower value – with the purpose to cause insertion of larger volumes
of water – RSM will identify the attack, because the execution of the specification
will calculate a different value for the water level than the one measured with the
sensor. The difference between the expected water level and the one read will lead
to a detection of the deviation; it will raise an alarm and, eventually, will cause the
action to be stopped. Although there exist complex false data injection attacks that
are not detected by RSM, its detection of common attacks combined with the proof
that it detects all computational attacks makes the ARMET behavioral approach a
powerful tool for the protection of processes and applications in the IoT space.

6.9 Privacy and Dependability

Privacy protection is one of the most significant challenges in IoT systems because
of the legal requirements in many application domains such as home environments,
smart grids, and health systems. There are increasing restrictions and constraints on
the collection, storage, and processing of personal information involved in all appli-
cations, including IoT. Privacy protection solutions may need to integrate a range of
methods and techniques, such as time-limited storage of sensitive information,
access control systems to enable access only for authorized personnel, accounting
systems to enable auditing, etc. The burden to comply with the required policies and
laws is further increased by the increasing amount of information considered as
personal or private, which leads to a need for adaptive and scalable solutions that
accommodate new policies as the relevant legal requirements emerge [Mul06]. The
ARMET approach provides a powerful solution to the problem of privacy protec-
tion, when privacy protection is viewed as a safety property. Privacy protection
originates from legal requirements that can be expressed as conditions in an infor-
mation system, i.e., they can be expressed as preconditions, postconditions, or
References 73

invariants in a program; for example, a function that is used by unclassified users

can be restricted from accessing specific variables that are available only to highly
classified ones. With this view, privacy requirements can be expressed as safety
requirements, refined into conditions, and enforced with a run-time monitor, like
RSM, which will detect all attempts to violate the defined conditions. Importantly,
the programmability of the conditions enables dynamic adjustment of the run-time
monitors as new conditions are established by emerging legal frameworks.
Interestingly, the behavioral approach to security and safety provides a promis-
ing solution to the challenge of combining dependability and security in the same
framework [Ser08]. Dependable systems have been developed for a long time with
well-understood methodologies, but they are based on fault models that consider
faults and errors accidental [Sie82]. In the case of security attacks though, malicious
attackers insert faults on purpose, and the models of these faults are fundamentally
different from the accidental ones. The behavioral approach to security considers
only the attack model, e.g., computational or false data injection, and is not influ-
enced by its origin – accident or on purpose. Thus, it detects accidental faults and
malicious attacks with the same method and in the same way. Attribution of the fault
is made in ARMET, for example, only after detection and based on the available
information and the trust model used. Independently of the attribution though, the
behavioral approach will detect a problem, providing a unified approach to security
and dependability.

References

[AES01] NIST. (2001). Advanced Encryption Standard. FIPS Publication 197, November 26,
2001.
[Ana15] Anand, A., & Knepper, R. (2015). ROSCoq: Robots powered by constructive reals.
In Proceedings of the 2015 International Conference on Interactive Theorem Proving (pp.
34–50). Springer LNCS-9236.
[And96] Anderson, R., & Kuhn, M. (1996). Tamper resistance: A cautionary note. In Proceedings
of the 2nd Workshop on Electronic Commerce, USENIX Association, Berkeley, CA, 1996,
pp. 1–11.
[Arb97] Arbaugh, W., Farber, D., & Smith, J. (1997). A secure and reliable bootstrap architec-
ture. In Proceedings of the IEEE Symposium on Security and Privacy, 1997, pp. 65–71.
[ARM05] ARM Security Technology. (2005). Building a Secure System using TrustZone

Technology. ARM white paper, Document PRD29-GENC-009492C, 2005. http://infocenter.
arm.com/help/topic/com.arm.doc.prd29-genc-009492c/PRD29-GENC-009492C_trustzone_
security_whitepaper.pdf
[Bar06] Bar-El, H., Choukri, H., Naccache, D., Tunstall, M., & Whelan, C. (2006). The sor-
cerer’s apprentice guide to fault attacks. Proceedings of the IEEE, 94(2), 370–382.
[Bel03] Belenky, A., & Ansari, N. (2003). IP traceback with deterministic packet marking. IEEE
Communications Letters, 7(40), 162–164.
[Ber04] Bertot, Y., & Castran, P. (2004). Interactive theorem proving and program development-
Coq’Art: The calculus of inductive constructions. Berlin Heidelberg: Springer.
[Bes81] Best, R. (1981). Crypto microprocessor for executing enciphered programs. US patent
4,278,837, July 1981.
74 6 Security and Safety

[Bly93] Blythe, S., Fraboni, B., Lall, S., Ahmed, H., & De Riu, U. (1993). Layout reconstruction
of complex silicon chips. IEEE Journal on Solid-State Circuits, 28(2), 138–145.
[Bol95] Bolding, D. (1995). Network security, filters and firewalls. Crossroads, 2(1), 8–10.
[Cab01] Cabrera, J., Lewis, L., Qin, X., Lee, W., Prasanth, R., Ravichandran, B., & Mehra, R.
(2001). Proactive detection of distributed denial of service attacks using MIB traffic vari-
ables—A feasibility study. In Proceedings of the IEEE/IFIP International Symposium on
Integrated Network Management, pp. 609–622.
[Cha03] Chan, H., Perrig, A., & Song, D. (2003). Random key predistribution schemes for sen-
sor networks. In Proceedings of the IEEE Symposium on Security and Privacy, pp. 197–213.
[Cha16] Chan, M., Ricketts, D., Lerner, S., & Malecha, G. (2016). Formal verification of stabil-
ity properties of cyber-physical systems. In CoqPL’16, Jan 2016.
[Chl96] Chlipala, A. (2016). Ur/web: A simple model for programming the web. Communications
of the ACM, 59(8).
[Cos16] Costan, V., & Devadas, S. (2016). Intel SGX explained. Cryptology ePrint Archive:
Report 2016/086, IACR.
[Del15] Delaware, B., Pit-Claudel, C., Gross, J., & Chlipala, A. (2015). Fiat: Deductive synthesis
of abstract data types in a proof assistant. In Proceedings of the 42nd Annual ACM SIGPLAN-
SIGACT Symposium on Principles of Programming Languages (POPL’15), Mumbai, India,
Jan. 15–17, 2015, pp. 689–700.
[Dij67] Dijkstra, E. W. (1967). A constructive approach to the problem of program correctness,
August 1967, circulated privately.
[Gar03] Garfinkel, T., Rosenblum, M., & Boneh, D. (2003). Flexible OS support and applica-
tions for trusted computing. In Proceedings of the 9th Conference on Hot Topics in Operating
Systems (Vol. 9, pp. 25–25).
[Gol07] Goldsby, H. J., Cheng, B. H. C., & Zhang, J. (2008). AMOEBA-RT: Run-Time

Verification of Adaptive Software. In Proceedings of Models in Software Engineering
(MODELS 2007), Nashville, TN, USA, September 30–October 5, 2007, LNCS-5002, Springer,
pp. 212–224.
[Hod04] Hodge, V., & Austin, J. (2004). A survey of outlier detection methodologies. Artificial
Intelligence Review, 22(2), 85–126.
[Hus03] Hussain, A., Heidemann, J., & Papadopoulos, C. (2003). A framework for classify-
ing denial of service attacks. In Proceedings of the conference on applications, technologies,
architectures, and protocols for computer communications (pp. 99–110). New York: ACM.
[Jin10] Jin, Y., & Makris, Y. (2010). Hardware Trojans in wireless cryptographic ICs. IEEE
Design and Test, 27(1), 26–35.
[Joy09] Joye, M. (2009). Protecting RSA against fault attacks: The embedding method. In
Proceedings of the Workshop on Fault Diagnosis and Tolerance in Cryptography (FDTC),
pp. 41–45.
[Ker08] Keramidas, G., Antonopoulos, A., Serpanos, D., & Kaxiras, S. (2008). Nondeterministic
caches: A simple and effective defense against side channel attacks. Design Automation of
Embedded Systems, 12(3), 221–230.
[Kha15] Khan, M. T., Serpanos, D., & Shrobe, H. (2015). On the formal semantics of the cogni-
tive middleware AWDRAT. Technical Report MIT-CSAIL-TR-2015-007, Computer Science
and Artificial Intelligence Laboratory, MIT, USA, March 2015.
[Kha17] Khan, M. T., Serpanos, D., & Shrobe, H. ARMET: Behavior-Based Secure and Resilient
Industrial Control Systems. In Proceedings of the IEEE, Preprint. URL: http://ieeexplore.ieee.
org/stamp/stamp.jsp?tp=&arnumber=8011473&isnumber=4357935
[Kim04] Kim, S. S., Reddy, A. L. N., & Vannucci, M. (2004). Detecting traffic anomalies through
aggregate analysis of packet header data. In Proceedings of 3rd International IFIP-TC6
Networking Conference (NETWORKING 2004), Athens, Greece, May 9–14, 2004, Springer
LNCS-3042, pp. 1047–1059.
[Koc96] Kocher, P. (1996). Timing attacks on implementations of Diffie-Hellman, RSA, DSS,
and other systems. In Advances in Cryptology – CRYPTO’96. Springer, pp. 104–113.
References 75

[Koc99] Kocher, P., Jaffe, J., & Jun, B. (1999). Differential power analysis. In Advances in
Cryptology-CRYPTO’99. Springer, pp. 789–789.
[Kuh97] Kuhn, M. (1997). The Trust No1 cryptoprocessor concept. http://www.cl.cam.ac.uk/
mgk25/.
[Lak05] Lakhina, A., Crovella, M., & Diot, C. (2005). Mining anomalies using traffic feature dis-
tributions. In Proceeding of the 2005 Conference on Applications, Technologies, Architectures
and Protocols for Computer Communications (SIGCOMM 2005), Philadelphia, PA, USA,
August 22–16, 2005, pp. 217–228.
[Lie03] Lie, D., Thekkath, C., & Horowitz, M. (2003). Implementing an untrusted operating
system on trusted hardware. ACM SIGOPS Operating Systems Review, 37(5), 178–192.
[Lie00] Lie, D., Thekkath, C., Mitchell, M., Lincoln, P., Boneh, D., Mitchell, J., & Horowitz, M.
(2000). Architectural support for copy and tamper resistant software. ACM SIGPLAN Notices,
35(11), 168–177.
[Mal16] Malecha, G., Ricketts, D., Alvarez, M. M., & Lerner, S. (2016). Towards foundational
verification of cyber-physical systems. In Proceedings of 2016 Science of Security for Cyber-
Physical Systems Workshop (SOSCYPS), April 2016, pp. 1–5.
[Mic11] MICROSOFT. (2011). Shared source initiative. http://www.microsoft.com/resources/
ngscb/default.mspx
[Mor15] Dworkin, M. J. (2015). SHA-3 Standard: Permutation-based hash and extendable-
output functions. Federal Information Processing Standards (NIST FIPS) – 202, August 04,
2015.
[Mul06] Muller, G. (2006). Special issue: Privacy and security in highly dynamic systems-
introduction. Communications of the ACM, 49(9), 28–31.
[New16] Newman, L. H. (2016). What we know about Friday’s massive east coast internet out-
age. WIRED, October 21, 2016.
[Pax99] Paxson, V. (1999). Bro: A system for detecting network intruders in real-time. Computer
Networks, 31(23–24), 2435–2463.
[Pea02] Pearson, S. (2002). Trusted computing platforms: TCPA technology in context. USA:
Prentice Hall.
[Pen07] Peng, T., Leckie, C., & Ramamohana-Rao, K. (2007). Survey of network-based defense
mechanisms countering the DoS and DDoS problems. ACM Computing Surveys, 39(1), Article 3.
[Per04] Perrig, A., Stankovic, J., & Wagner, D. (2004). Security in wireless sensor networks.
Communications of the ACM, 47(6), 53–57.
[Qui01] Quisquater, J. J., & Samyde, D. (2001). Electromagnetic analysis (EMA): Measures and
counter-measures for smart cards. In Proceedings of the International Conference on Research
in Smart Cards: Smart Card Programming and Security, Springer LNCS-2140, pp. 200–210.
[Rav04] Ravi, S., Raghunathan, A., Kocher, P., & Hattangady, S. (2004). Security in embed-
ded systems: Design challenges. ACM Transactions on Embedded Computing Systems, 3(3),
461–491.
[Roe99] Roesch, M. (1999). Snort – lightweight intrusion detection for networks. In Proceedings
of the 13th USENIX Conference on System Administration (LISA ‘99), pp. 229–238.
[RSA78] Rivest, R. L., Shamir, A., & Adleman, L. (Feb. 1978). A method for obtaining digital
signatures and public-key cryptosystems. Communications of the ACM, 21(2), 120–126.
[Sav01] Savage, S., Wetherall, D., Karlin, A., & Anderson, T. (2001). Network support for IP
traceback. IEEE/ACM Transactions on Networking, 9(3), 226–237.
[Ser08] Serpanos, D., & Henkel, J. (2008). Dependability and security will change embedded
computing. Computer, 41(1), 103–105.
[Ser13] Serpanos, D. N., & Voyiatzis, A. G. (2013). Security challenges in embedded systems.
ACM Transactions on Embedded Computing Systems, 12(1s), Article 66.
[Sie82] Siewiorek, D., & Swarz, R. (1982). The theory and practice of reliable system design.
Bedford: Digital Press.
76 6 Security and Safety

[Sli02] Slijepcevic, S., Potkonjak, M., Tsiatsis, V., Zimbeck, S., & Srivastava, M. (2002). On
communication security in wireless ad-hoc sensor networks. In Proceedings of the 11th IEEE
International Workshop on Enabling Technologies, pp. 139–144.
[Sno02] Snoeren, A., Partridge, C., Sanchez, L., Jones, C., Tchakountio, F., Schwartz, B., Kent,
S., & Strayer, W. (2002). Single-packet IP traceback. IEEE/ACM Transactions on Networking,
10(6), 721–734.
[Val00] Valdes, A., & Skinner, K. (2000). Adaptive, model-based monitoring for Cyber Attack
Detection. In Proceedings of the 3rd International Workshop on Recent Advances in Intrusion
Detection (RAID 2000), Toulouse, France, October 2–4, 2000, Springer, pp. 80–93.
[Wan07] Wang, H., Jin, C., & Shin, K. (2007). Defense against spoofed IP traffic using hop-count
filtering. IEEE/ACM Transactions on Networking, 15(1), 40–53.
[Wan02] Wang, H., Zhang, D., & Shin, K. (2002). Detecting SYN flooding attacks. In Proceedings
of the 21st Annual Joint Conference of the IEEE Computer and Communications Societies
(INFOCOM’02), pp. 1530–1539.
[Wat07] Watterson, C., & Heffernan, D. (2007). Runtime verification and monitoring of embed-
ded systems. Software, IET, 1(5), 172–179.
[Yan12] Yang, J., Yessenov, K., & Solar-Lezama, A. (2012). A language for automatically
enforcing privacy policies. In Proceedings of the 39th ACM Symposium on Principles of
Programming Languages (POPL 2012), Philadelphia, PA, USA, January 25–27, 2012,
pp. 85–96.
[Zhu03] Zhu, S., Setia, S., & Jajodia, S. (2003). LEAP: Efficient security mechanisms for large-
scale distributed sensor networks. In Proceedings of the 10th ACM Conference on Computer
and Communications Security, pp. 62–72.
Chapter 7
Security Testing IoT Systems

7.1 Introduction

Systems need to be evaluated for conformance to specifications and requirements,

including security, and IoT systems are no exception. Verification and validation
techniques are one option to ensure that systems are built according to specifications
and requirements, but their use is limited for two main reasons: (i) the complexity
of these processes is growing exponentially with the size of the checked system, and
(ii) the current business models that include long supply chains with different pro-
viders and developers of system components do not enable a unified description of
designs and implementations that can be checked as a whole. In the case of IoT
systems specifically, the size of most systems is not prohibitive for formal verifica-
tion and validation methods; however, the lack of a complete implementation with
the same tools and models leads to fragmented application of verification techniques
to components. Testing constitutes an important and necessary phase in system
development, which complements all other approaches and enables the evaluation
of integrated systems. Thus, testing is an integral part of the systems development
cycle with the purpose to evaluate system correctness, performance, and security at
least. Importantly, testing is a method used by customers and certification authori-
ties to evaluate conformance of systems to standards and to provide certifications at
the device, system, and product level.
The wide deployment of consumer electronics devices has brought significant
attention to testing and its methodologies not only for accepting devices by consum-
ers but also for security, since attackers exploit testing methodologies to identify
vulnerabilities and exploit them for their purposes. This is especially important to
IoT systems which typically have a cyberphysical component. Identification of vul-
nerabilities in IoT systems and their exploitation may compromise their safety prop-
erties and lead to significant operational problems that result to monetary losses,
operation disruption, and even loss of life.

D. Serpanos, M. Wolf, Internet-of-Things (IoT) Systems,
https://doi.org/10.1007/978-3-319-69715-4_7
78 7 Security Testing IoT Systems

Successful testing of IoT systems is critical considering that many of them have
strong requirements that are crucial to their operation, such as meeting real-time
constraints, satisfying specific safety properties, and continuing operation even
under strained conditions. Furthermore, IoT systems include a communication
component, which constitutes a testing challenge because the specifications of com-
munication protocols often have undefined parameters that lead to differing imple-
mentations by different vendors; this is the reason why interoperability in
communication systems is an important challenge. The criticality of IoT testing,
especially for security, becomes more apparent when considering industrial IoT
systems, which are extensively used in critical infrastructures nowadays, such as
energy networks, water management systems, etc. Successful testing not only con-
firms the expected operations but takes away from attackers the tools to cause mal-
functions and disruptions; in the emerging environment, even crashing an application
or an operation may be more catastrophic than hijacking them.
Hardware and software testing are technological areas with significant effort in
the market and in academia for decades. A large number of methodologies and tools
have been developed, but software testing has been a significantly harder problem
than hardware testing because of several differentiating characteristics software has,
such as evolution through added features and functionality, fault models and lack of
re-use. Considering that most IoT systems are built using off-the-shelf hardware
components and computing subsystems, we address software testing for security in
this chapter, and, more specifically, we focus on the testing of their communication
protocol implementations, since it is the point of entry to systems and a common
target of attackers. We present fuzz testing, the most common testing approach for
security, which requires no information about the internal structure of the system
that is tested. As industrial IoT systems constitute an attractive target for attackers
that exploit testing techniques, we use as an example the Modbus protocol and
describe fuzz testing techniques for its implementations, which give successful
results for existing protocol implementations in the field.

7.2 Fuzz Testing for Security

Vulnerabilities in network systems and applications are identified and disseminated

publicly [Nis, Sfo, Str]. The cost of fixing these vulnerabilities can be high, while
their exploitation may have quite costly consequences. As a result, there is strong
research and development effort to reduce such vulnerabilities.
Static analysis of source code is one approach that does not require program
execution but is limited because it does not detect vulnerabilities that are activated
by dynamic instruction sequences, during program execution, e.g., dependent on
subroutine calls [Che07, Vie00]. Also, these methods present a high false-positive
rate leading to significant overhead for the evaluation of the results. Alternatively,
dynamic analysis methods intervene in program execution. StackGuard, for exam-
ple, expands a C compiler and produces executable code that identifies potential
7.2 Fuzz Testing for Security 79

execution faults – examining addresses, for example – without changing the func-
tionality of the original programs [Cow98]. TaintCheck applies taint analysis for
automatic vulnerability analysis without need for the source code [Cla07, New04].
Dynamic analysis methods enable powerful mechanisms for vulnerability detection
at the cost of execution time overhead, because of the additional code that is inserted
in the application program. Simulation has also been proposed for vulnerability
testing, where a simulation environment is used to inject faults to a program and
check its behavior [Du02]. This is a systematic method, but it is limited to input
patterns that may cause errors.
Fuzz testing (fuzzing) provides an alternative, reliable approach with successful
results and advantages over the previous methods. Fuzzing is a testing method that
applies test inputs (vectors) to a system under test (SUT) and observes its outputs,
as shown in Fig. 7.1. The goal of the fuzzer is to identify faults in the SUT, e.g., to
detect inputs that lead to a system crash. The effectiveness of the fuzzer is based on
its ability to identify as many vulnerabilities as possible covering effectively the
input value space. If there is inability to identify whether a system or a program has
crashed during a test, the effectiveness of the fuzzer cannot be evaluated.
Fuzzing provides several advantages over static and dynamic analysis. First, it
can be applied to programs whose source code is not available. Second, it is inde-
pendent of the internal complexity of the tested software which limits in practice
other methods, such as static analysis. Because of this independence, the same fuzz-
ing tool can be used to test similar programs independently of the programming
language used for their coding. Finally, the identified faults and errors can be
directly associated to the user input and can be evaluated easier.
Fuzz testing has its limitations. The space of input values is vast, and thus, it is
impossible to test large systems for all their potential input values within reasonable
time frames. A fuzzer that produces random input values can discover faults and
vulnerabilities, but, in general, it will not detect easily many important vulnerabili-
ties unless it follows some specific strategic approach. Its effectiveness depends on
its ability to identify representative input values, which may originate from attacks
or common errors with invalid inputs, and detect vulnerabilities that are useful to
attackers.
Fuzz testing can be classified in three (3) categories, depending on the informa-
tion that is available for the system under test (SUT) [Tak08] [Sut07], as shown in
Fig. 7.1:

Fig. 7.1 Fuzz testing Fuzzed inputs

configuration

-Tester/ System-Under-Test
Fuzzer (SUT)

Outputs
80 7 Security Testing IoT Systems

• White-box testing: the source code or the specification of the SUT is known.
• Black-box testing: the internal structure of the SUT is unknown –testing is lim-
ited to observations of SUT inputs and outputs.
• Grey-box testing: partial information for the SUT internal structure is available,
e.g., through reverse engineering or static analysis results.

7.2.1 White-Box Fuzzing

Modern white-box fuzz testing tools exploit the information about the system’s
internal structure using symbolic execution techniques or taint analysis to identify
vulnerabilities. Symbolic execution replaces symbolic values in the source code or
the program flow, in order to evaluate code execution paths [Cad13]. These tech-
niques have been explored widely in efforts such as DART [God05], SAGE [God12],
EXE [Cad06], and KLEE [Cad08]. Tools like AEG [Avg11] and CRAX [Hua12]
combine symbolic execution with concrete execution, employing concolic testing
[Sen05] to identify vulnerabilities that lead to control flow hijacking. Such tools
have been very successful in fuzz testing of Windows and Linux applications
[God12, Cad06]. The techniques have the advantage that they can explore all pos-
sible modes of applications, since they use the source code, and identify dead code.
However, they cannot identify logic errors in programs and are unable to explore all
execution paths in large programs with complex structures. Tools that use taint anal-
ysis identify potential attack points in programs by tracing tainted values and then
fuzz the input values to these attack points [Sch10]. BuzzFuzz [Gan09] and
TaintScope [Wan10] are two representative tools that exploit taint analysis
techniques.

7.2.2 Black-Box Fuzzing

Black-box fuzzing techniques do not have any structural information about the sys-
tem under test. Since testing requires application of inputs to the system and obser-
vation of its outputs, one of the most popular targets of black-box fuzzing is the
implementation of communication protocols because they provide the first point of
entry to systems and they typically implement some standard; so, our description is
focused on protocols, although the techniques can be applied to application and
system software in general.
There are two main approaches to generate fuzz testing inputs to protocols: (i)
data generation and (ii) data mutation [Nal12, Tak08, Sut07]. Data generation tech-
niques create input packets to a protocol implementation either randomly or with a
systematic method that takes into account the specifications of the specific protocol.
The contents of these packets may be completely random, or they may take into
account the structure of the packets, i.e., their fields, and insert either random or
7.2 Fuzz Testing for Security 81

special values in the fields, depending on various parameters, such as the system
interface or a specific targeted operation. In this case, the specification of the proto-
col needs to be integrated in the fuzzer. Clearly, the effectiveness of the fuzzing
process depends on the successful integration of the protocol specification in the
fuzzer, since any problem in that integration may lead to limited or no coverage of
a wide range of tests.
Mutation fuzzing creates the test inputs based on legal protocol packets. It takes
as input the legal packets and changes (mutates) some of their data, e.g., specific
fields, in order to create the test packets that are input to the system. This approach
is especially useful in cases where the protocol is complex, because the fuzzer does
not construct packets from scratch but uses known legal packets and mutates them.
Thus, the fuzzer does not need to include the protocol specification, and the author
of the fuzzer does not need to delve into the details of the protocol, thus avoiding the
risk of misinterpretations and creation of inappropriate packets.
These two main approaches are coupled with techniques that choose the values
that are used in the generated or mutated packets. The most common techniques are:
1. Random: generates of random values without any consideration of packet struc-
ture, legal values, etc. The technique is fast, low cost, and quite successful
[Mil90, Mil95, Mil06] but limited because it is characterized by low test
coverage.
2. Block-based: manages data values in blocks, taking into account the specifica-
tions of protocols and creating meaningful blocks of values, in contrast to ran-
dom values. The technique has been used widely in frameworks and tools, such
as Spike [Ait02], SNOOZE [Ban06], Sulley [Ami14], Peach [Pea14], Autodafè
[Vua06], and AspFuzz [Kit10], and is especially useful in mutation fuzzing. The
success of the technique depends on the successful integration of protocol specs
in the fuzzers.
3. Grammar-based: embeds a grammar in the fuzzer, in order to cover part of the
specification of legal inputs to the system under test. Fuzzing inputs are created
with the consideration of the grammar. PROTOS [PRO] is a representative tool
using this technique.
4. Heuristic-based: generates new fuzzing inputs taking into account the effective-
ness of the inputs applied in the past. Processing of the outputs obtained from the
prior tests can be done with various methods such as with appropriate genetic
algorithms [Spa07] or statistical analysis [Zha11].
There exist also approaches that construct protocol descriptions or specifications by
observing real protocol traffic. With this information, related tools can make more
effective decisions about how to mutate observed packets, in order to increase the
effectiveness of mutation fuzzers. General Purpose Fuzzer (GPF) [Vda14] and
AutoFuzz [Gor10] are representative tools that employ this approach. Interestingly,
in mutation fuzzing there is also the approach of creating test cases based on exist-
ing attack traffic [Ant12, Tsa12].
82 7 Security Testing IoT Systems

7.3 Fuzzing Industrial Control Network Systems

Fuzz testing for industrial networks has attracted significant interest in the market
and in academia, considering the increasing adoption of industrial control systems
in critical infrastructures. Many commercial and open source fuzzing tools support
industrial protocols. Sulley [Dev07] provides fuzzing modules for ICCP, Modbus,
and DNP3 since 2007. ProFuzz [Koc], a fuzzing tool based on Scapy [Bio], sup-
ports fuzzing in PROFINET. Achilles test platform [Ach17] supports fuzzing for
SCADA protocols, like Modbus/TCP and DNP3.
There is also research work in fuzzing industrial protocols using various tech-
niques. Black-box mutation fuzzing, for example, has been explored for SCADA
networks without any knowledge about the networking protocol [Sha11] and using
the LZ-Fuzz tool [Bra08] to evaluate its effectiveness. OPC-MFuzzer [Wan13,
Qi14] is a mutation fuzzer (based on Peach [Pea14]) for OPC SCADA fuzzing.
Based on three different mechanisms to produce fuzzing inputs, the tool identified
and confirmed known vulnerabilities that had been included previously in the
National Vulnerability Database (NVD) [Nis].
Modbus fuzzing has attracted significant attention as well. BlackPeer [Byr06]
produces inputs and checks outputs using a grammar that is included in the tool;
although successful, it has limited flexibility as it cannot adjust easily to new tests.
Sulley [Dev07], a block-based framework, enables methodical and easy mutation
fuzzing through its Modbus module; however, its block-based approach is limited
for testing devices that deviate from the standard implementation and are custom-
ized by the users. A framework for fuzz testing Modbus for security has also been
proposed based on Scapy [Kob07].

7.4 Fuzzing Modbus

7.4.1 The Modbus Protocol

Modbus is an application protocol for industrial control system communication,

which has become a standard published by Modbus IDA [Mod, ModS]. Its specifi-
cation defines the protocol for direct communication over serial links as well as
communication over TCP connections. The popular Modbus protocol stacks are
shown in Fig. 7.2; it should be noted that, in correspondence with the ISO protocol
reference model, Modbus is an application layer protocol defined to interface
directly to layer 1 (serial) and layer 2 (HDLC) protocols –stacks (a) and (b) in the
figure – or to TCP through an adjusting sublayer that is denoted as Modbus messag-
ing (mapping) on TCP, as shown in stack (c).
The protocol implements client/server (alternatively, master/slave) communica-
tion through a request-response model between a control center and field devices,
such as a SCADA and PLCs. For example, a SCADA master unit (client) may
7.4 Fuzzing Modbus 83

Modbus Application
Protocol

Modbus Messaging
(Mapping) on TCP

TCP

Modbus Application Modbus Application

IP
Protocol Protocol

Serial Master/Slave HDLC

Ethernet Data Link and
Physical Protocols
Physical Protocol Physical Protocol
(RS-232/RS-485) (RS-485)

(a) (b) (c)

Fig. 7.2 Modbus protocol stacks

Fig. 7.3 Modbus
Function code
application packet Data
(FC)

request the reading of a sensor attached to a slave PLC (server), or it may request
the writing of a command to an actuator to turn a switch.
Modbus application packets are simple, composed of two fields, a function code
(FC) and data, as shown in Fig. 7.3. Requests from servers send the function code
that defines the operation to be performed and the related data, e.g., an address or
command. A response from a client includes the function code that was executed at
the client and the resulting related data. Since an operation may not be successfully
executed at the client, the protocol defines that the client will respond with the origi-
nal function code if the related operation is executed correctly, or it will send an
exception code indicating that the operation was not executed.
Modbus has three different classes of function codes: public codes, user-defined
codes and reserved ones. Public codes are defined by the standard and include num-
bering and operation definition. Reserved codes are also public, but they cannot be
used freely, since they have been defined and reserved for interoperability purposes
with legacy industrial control systems. User-defined codes are available to develop-
ers and users to implement specialized function codes at will. Since the function
code field is 8 bits, function codes can have 256 values, in the range 0–255. Public
codes are in the ranges 1–64, 73–99, and 111–127; these ranges include the reserved
codes. User-defined codes may have values in the ranges 65–72 and 100–110. The
codes 128–255 are used to indicate errors; each function code has its unique related
exception code, which differs from the function code at the most significant bit;
with the 8-bit format, all function codes have “0” as their most significant bit and all
exception codes have it as “1.”
84 7 Security Testing IoT Systems

Fig. 7.4 Encapsulated Modbus application packets

Most Modbus function codes perform read and write operations to device data.
For this purpose, Modbus considers that devices store data in tables. There are four
different table types, based on the data entry size (1 bit or 16 bits) and the access
operation allowed (read or read/write). The tables are denoted as (i) discrete input,
with 1-bit entries and only read operations allowed; (ii) coils, with 1-bit entries and
read/write operations allowed; (iii) input registers, with 16-bit entries and only read
operations allowed; and (iv) holding registers, with 16-bit entries and read/write
operations allowed. All four types of tables can have up to 64 K entries. Importantly,
these tables are actually virtual, meaning that they can be physically separate in the
device’s memory or they can overlay over the same physical memory cells. Modbus
can also access files, which are sequences of records (up to 10,000), and each record
has a length measured with 16-bit units.
Modbus application packets (protocol data units, or PDUs) are encapsulated in
lower layer protocol packets to be transmitted. When serial connections are used,
the application packets are encapsulated by the data link control (DLC) protocol and
produce DLC PDUs that are then transmitted by the serial protocol. DLC packets
add an address field for the slave next to the function code field and a checksum next
to the data field of the application protocol, as Fig. 7.4a shows. In the case of the
serial physical layer, there are two formats for the DLC packets, denoted as RDU
and ASCII. The main difference between the two is the size of the slave address and
the size of the function code field: in RTU format, they are both one byte, while in
the ASCII format, each one is 2 bytes long.
Modbus over TCP is performed by extending the Modbus application packet first
with an additional header, named MBAP (Modbus Application Protocol) header as
shown in Fig. 7.4b, and then encapsulating this extended packet by the TCP/IP pro-
tocol stack, which employs Ethernet at the data link control and physical protocol
layers, as shown in Fig. 7.2c.
Modbus does not include security mechanisms such as authentication, confiden-
tiality, or integrity. The lack of security renders its implementations vulnerable to a
wide range of attacks. The lack of confidentiality enables attackers to extract infor-
mation from captured packets, while the lack of integrity checks does not allow a
receiver of a packet to identify whether the packet has been altered. Replay attacks
are possible as well and the lack of non-repudiation mechanisms can lead to inabil-
ity to analyze and audit systems credibly.
7.4 Fuzzing Modbus 85

7.4.2 Modbus/TCP Fuzzer

There exist several Modbus fuzzers, as described in Sect. 7.2. In this subsection, we
present the approach and results of MTF (Modbus/TCP fuzzer) [Voy15] as a repre-
sentative example. The choice of MTF is based on its characteristics that show the
trends in fuzzing technology today: it is an automated tool, it provides good cover-
age of input tests, and it does not require physical access to the system under test,
operating remotely over the network. These characteristics make MTF an attractive
tool for testing security and compliance of Modbus connected devices.
MTF incorporates the specification of Modbus/TCP and supports fuzzing both
master and slave devices on the network. As an automated tool for fuzzing, MTF
operates in three main phases: (i) reconnaissance, (ii) attack, and (iii) failure detec-
tion. In the first phase, MTF identifies the operational characteristics and parameters
of the tested system. In the second phase, it applies tests to the system and collects
its responses, while in the third phase it evaluates the collected (observed) responses
to identify security problems and system failures.
Reconnaissance is an important operation in automated black-box or gray-box
fuzzers, because it identifies the operations performed by the system under test and
its important parameters. In the case of Modbus, in order to generate meaningful
tests, one needs to know the function codes used by the system as well as its mem-
ory model, i.e., the four memory types – discrete inputs, coils, input registers, and
holding registers – that are specified by the standard. MTF explores the function
codes through different methods, in order to accommodate different types of devices
that may be fully or partially conformant with the standard. A straightforward
method is to ask the device for identification information – the standard specifies
function code 43 for this operation – and then, based on this, to find information
off-line about the supported function codes, e.g., from a manual. Alternatively, it
sends legitimate requests and examines the responses, which indicate whether the
requests have been executed or not (as described in the standard specification), or it
monitors traffic from the device and extracts functional information from that.
In regard to the memory model of the tested system, MTF effectively identifies
the boundary memory addresses for each type of memory. This is done either
actively, sending packets with the appropriate function codes probing specific
address values, or passively, observing traffic which eventually indicates memory
bounds, although these bounds may be approximate.
Taking into account the list of function codes and the memory mapping for the
four memory types, the fuzzer can construct legitimate packets and fuzz them in
order to test the system. Since the supported function codes are known, MTF con-
structs a set of packet sequences for each supported function code, where each
sequence implements a potential attack to the system; such attacks include packet
removal, packet injection, and packet field manipulation.
Packet field manipulation is performed with field values that are boundary, ran-
dom, or illegal.
86 7 Security Testing IoT Systems

When tests are applied, the response, or its absence, is recorded. The tool records
the sequence of all tests and related responses and produces a list of errors which are
invalid responses (out of specification), valid but with incorrect parameters (values,
size, etc.), and delayed or incomplete (no response). Further processing of the
records, including both the valid request/response pairs and the errors, leads to
detection of security and dependability problems, i.e., malicious or accidental
failures.
The MTF approach is representative of the trends in fuzzing industrial protocols.
It provides a complete approach to fuzzing, starting with reconnaissance, continu-
ing with meaningful tests and, finally, analyzing the results for security and reliabil-
ity failures. Its practicality has been demonstrated through the prototype
implementation described in the original work [Voy15], which has been used to
evaluate several commercial and open source Modbus subsystems and for several
attacks. The attacks include packet dropping, packet injection, illegal field values,
altered function codes, and even flooding, leading to denial of service attacks.
Importantly, many of these attacks have been successful against commercial
Modbus implementations, as the reported original results demonstrate. Interestingly,
MTF succeeds in attacking these implementations much more efficiently than alter-
native tools, i.e., with a significantly smaller number of packets. Overall, the results
demonstrate that the approach of generation fuzzing is an effective and efficient
fuzzing method.

References

[Ach17] Wurldtech- GE Digital, Achilles Test Platform, 2017. https://www.ge.com/digital/sites/

default/files/achilles_test_platform.pdf
[Ait02] Aitel, D. (2002). An introduction to SPIKE, the Fuzzer Creation Kit. Presented at The
BlackHat USA Conference. www.blackhat.com/presentations/bh-usa-02/bh-us-02-aitel-spike.
ppt
[Ami14] Amini, P. (2014). Sulley: Pure Python fully automated and unattended fuzzing frame-
work. https://github.com/OpenRCE/sulley
[Ant12] Antunes, J., & Neves, N. (2012). Recycling test cases to detect security vulnerabili-
ties. In Proceedings of the 23rd International Symposium on Software Reliability Engineering,
Dallas, Texas, November 27–30, 2012, pp. 231–240.
[Avg11] Avgerinos, T., Cha, S. K., Hao, B. L. T., & Brumley, D. (2011). AEG: Automatic Exploit
Generation. In Proceedings of the Network and Distributed System Security Symposium
(NDSS’11), San Diego, California, February 6–9, 2011.
[Ban06] Banks, G., et al. (2006). SNOOZE: toward a Stateful NetwOrk prOtocol fuzZEr. In
Proceedings of the 9th Information Security Conference (ISC ‘06), pp. 343–358.
[Bio] Biondi, P. Scapy, python interactive packet manipulation framework. http:/www.secdev.
org/projects/scapy/
[Bra08] Bratus, S., Hansen, A., & Shubina, A.(2008). LZFuzz: A fast compression-based Fuzzer
for poorly documented protocols. Technical Report TR2008–634, Dept. of Computer Science,
Dartmouth College, New Hampshire.
[Byr06] Byres, E. J., Hoffman, D., & Kube, N. (2006). On shaky ground – A study of security
vulnerabilities in control protocols. In Proceedings of the 5th International Topical Meeting
References 87

on Nuclear Plant Instrumentation Controls, and Human Machine Interface Technology,

American Nuclear Society, Albuquerque, November 12–16, 2006.
[Cad06] Cadar, C., Ganesh, V., Pawlowski, P., Dill, D., & Engler, D. (2008). EXE: Automatically
generating inputs of Death. In: Proceedings of CCS’06, Oct–Nov 2006 (extended version
appeared in ACM TIS-SEC 12:2, 2008).
[Cad08] Cadar, C., Dunbar, D., & Engler, D. (2008). KLEE: Unassisted and automatic gen-
eration of high-coverage tests for complex systems programs. In Proceedings of OSDI’08,
December 2008.
[Cad13] Cadar, C., & Sen, K. (2013). Symbolic execution for software testing: Three decades
later. Communications of the ACM, 56(2), 82–90.
[Cla07] Clause, J., Li, W., & Orso, A. (2007). Dytan: A generic dynamic taint analysis frame-
work. In Proceedings of the 2007 International Symposium on Software Testing and Analysis
(ISSTA’07), London, UK, July 9–12, 2007, pp. 196–206.
[Che07] Chess, B., & West, J. (2007). Secure programming with static analysis. USA: Pearson
Education.
[Cow98] Cowan, C., et al. (1998). StackGuard: Automatic adaptive detection and prevention of
buffer-overflow attacks. In Proceedings of the 7th Usenix Security Symposium, San Antonio,
Texas, January 26–29, 1998.
[Dev07] Devarajan, G. (2007). Unraveling SCADA protocols: Using Sulley Fuzzer. Presented at
the DefCon’15 Hacking Conference, 2007.
[Du02] Du, W., & Mathur, A. P. (2002). Testing for software vulnerability using environment
perturbation. Quality and Reliability Engineering International, 18(3), 261–272.
[Gan09] Ganesh, V., Leek, T., & Rinard, M. (2009). Taint-based directed whitebox fuzzing.
In Proceedings of the 31st International Conference on Software Engineering (ICSE’09),
Vancouver, Canada, May 16–24, 2009, pp. 474–484.
[God05] Godefroid, P., Klarlund, N., & Sen, K. (2005). DART: Directed Automated Random
Testing. In Proceedings of the 2005 ACM SIGPLAN Conference on Programming language
design and implementation, Chicago, USA, June 12–15, 2005, pp. 213–223.
[God12] Godefroid, P., Levin, M. Y., & Molnar, D. (2012). SAGE: Whitebox fuzzing for security
testing. ACM Queue, 10(1).
[Gor10] Gorbunov, S., & Rosenbloom, A. (2010). Autofuzz: Automated network protocol fuzz-
ing framework. IJCSNS, 10(8), 239–245.
[Hua12] Huang, S. K., Huang, M. H., Huang, P. Y., Lai, C. W., Lu, H. L., Leong, W. M. (2012).
CRAX: Software crash analysis for automatic exploit generation by modeling attacks as sym-
bolic continuations. IEEE 6th International Conference on Software Security and Reliability,
June 20–22, 2012, pp. 78–87.
[Kit10] Kitagawa, T., Hanaoka, M., & Kono, K. (2010). AspFuzz: A state-aware protocol fuzzer
based on application-layer protocols. In Proceedings of the IEEE Cymposium on Computers
and Communications, Italy, 2010, pp. 202–208.
[Kob07] Kobayashi, T. H., Batista, A. B., Brito, A. M., & Motta Pires, P. S. (2007). Using a
packet manipulation tool for security analysis of industrial network protocols. In Proceedings
of 2007 IEEE Conference on Emerging Technologies and Factory Automation, Patras, 2007,
pp. 744–747.
[Koc] Koch, R. Profuzz. https://github.com/HSASec/ProFuzz
[Mil90] Miller, B. P., Fredriksen, L., & So, B. (1990). An empirical study of the reliability of
UNIX utilities. Communications of the ACM, 33(12), 32–44.
[Mil95] Miller, B. P., et al. (1995). Fuzz revisited: A re-examination of the reliability of UNIX
utilities and services. Technical report TR-1268, Department of Computer Sciences, University
of Wisconsin-Madison.
[Mil06] Miller, B. P., Cooksey, G., & Moore, F. (2006). An empirical study of the robustness of
MacOS applications using random testing. In Proceedings of the 1st International Workshop
on Random testing. Portland, Maine, July 20, 2006, pp. 46–54.
88 7 Security Testing IoT Systems

[Mod] ModBus Organization. ModBus Application Protocol Specification http://www.mod-

bus.org/docs/ModbusApplication/ProtocolV11b.pdf
[ModS] Modbus Serial Line Protocol and Implementation Guide V1.02 (Modbus_over_serial_
line_V1_02.pdf).
[Nal12] McNally, R., Yiu, K., Grove, D., & Gerhardy, D. Fuzzing: The State of the Art. Technical
Note DSTO-TN-1043, Defence Science and Technology Organization, Australia, 02–2012.
[New04] Newsome, J., & Song, D. Dynamic taint analysis for automatic detection, analysis, and
signature generation of exploits on commodity software. Technical report CMU-CS-04-140,
2004 (revised 2005).
[Nis] http://nvd.nist.gov
[Pea14] Peach Fuzzing Platform, http://www.peach.tech/products/peach-fuzzer/, 2017.
[PRO] PROTOS-Security Testing of Protocol Implementations. http//www.ee.oulu.fi/roles/

ouspg/Protos/
[Qi14] Qi, X., Yong, P., Dai, Z., Yi, S., & Wang, T. (2014). OPC-MFuzzer: A novel multi-layers
vulnerability detection tool for OPC protocol based on fuzzing technology. International
Journal of Computer and Communication Engineering, 3(4), 300–305.
[Sch10] Schwartz, E. J., Avgerinos, T., & Brumley, D. (2010). All you ever wanted to know
about dynamic taint analysis and forward symbolic execution (but might have been afraid to
ask). 2010 IEEE Symposium on Security and Privacy.
[Sen05] Sen, K., Marinov, D., & Agha, G. (2005). CUTE: A concolic unit testing engine for
C. In Proceedings of the 10th European Software Engineering Conference (held jointly with
13th ACM SIGSOFT International Symposium on the Foundations of Software Engineering),
September 5–9, 2005, pp. 263–272.
[Sfo] http://www.securityfocus.com
[Sha11] Shapiro, R., Bratus, S., Rogers, E., & Smith, S. (2011). Identifying vulnerabilities in
SCADA systems via fuzz-testing. Critical Infrastructure Protection V, IFIP AICT, 367, 57–72.
[Spa07] Sparks, S., Embleton, S., Cunningham, R., & Zou, C. (2007). Automated vulnerability
analysis: Leveraging control flow for evolutionary input crafting. In Proceedings of the 23rd
Annual IEEE Computer Security Applications Conference (ACSAC 2007), pp. 477–486.
[Str] http://www.securitytracker.com
[Sut07] Sutton, M., Greene, A., & Amini, P. (2007). Fuzzing: Brute force vulnerability discov-
ery. Addison-Wesley Professional.
[Tak08] Takanen, A., DeMott, J., & Miller, C. (2008). Fuzzing for software security testing and
quality assurance.
[Tsa12] Tsankov, P., Torabi Dashti, M., Basin, D. (2012). SECFUZZ: Fuzz-testing security pro-
tocols. In Proceedings of the 7th International Workshop on Automation of Software Test (AST
2012), June 2–3, 2012, Zurich, Switzerland.
[Vda14] VDA Labs, “General Purpose Fuzzer.” Rockford, Michigan, 2014, www.vdalabs.com/
tools/efs gpf.html
[Vie00] Viega, J., et al. (2000). ITS4: A static vulnerability scanner for C and C++ code. In
Proceedings of 16th Annual IEEE Conference Computer Security Applications (ACSAC'00),
New Orleans, Louisiana, 2000, pp. 257–267.
[Voy15] Voyiatzis, A. G., Katsigiannis, K., & Koubias, S. (2015). A Modbus/TCP Fuzzer for
testing internetworked industrial systems. In Proceedings of the 20th IEEE International
Conference on Emerging Technologies and Factory Automation (ETFA 2015). Luxembourg,
September 8–11, 2015, pp. 1–6.
[Vua06] Vuagnoux, M. (2006). Autodafe: An Act of Software Torture. Swiss Federal Institute of
Technology (EPFL), Cryptography and Security Laboratory (LASEC). http://autodafe.source-
forge.net
[Wan10] Wang, T., Wei, T., Gu, G., & Zou, W. (2010). TaintScope: A checksum-aware directed
fuzzing tool for automatic software vulnerability detection. 2010 IEEE Symposium on Security
and Privacy, Oakland, CA, USA, 2010, pp. 497–512.
References 89

[Wan13] Wang, T., et al. (2013). Design and implementation of fuzzing technology for OPC pro-
tocol. In Proceedings of 9th International Conference on Intelligent Information Hiding and
Multimedia Signal Processing, Beijing, China, 2013, pp. 424–428.
[Zha11] Zhao, J., Wen, Y., & Zhao, G. (2011). H-fuzzing: A new heuristic method for fuzz-
ing data generation. In Proceedings of Network and Parallel Computing, LNCS, Vol. 6985,
Springer, 2011, pp. 32–43.
Index

A security, 13
Advanced Message Queuing Protocol semantics-aware communication
(AMQP), 50 mechanism, 9
Alert system, 3 subscribers, 11
Amazon Simple Storage Service tree-based topology, 9
(Amazon S3), 12 XMPP protocol, 11
Amazon Web Services (AWS), 12 Zigbee () network, 10
Analysis system, 3 ARMET approach, 68–72
Application process security, 65, 66 AutoFuzz tool, 81
Architecture
AWS, 12
BLE, 10 B
CoAP, 11 Black-box fuzzing
databases communication protocols
Amazon Simple Storage Service implementation, 80
(Amazon S3), 12 data generation, 80
DTW, 13 data mutation, 81
Google Cloud Storage, 12 generated/mutated packets, 81
short-term and long-term storage, 12 GPF and AutoFuzz tools, 81
time-series data, 12 grammar-based, 81
unstructured databases, 12 heuristic-based, 81
DDS, 9 random and block-based techniques, 81
design goals, 9 Black-box testing, 80
event-style communication, 8 Block-based technique, 81
Google Cloud Pub/Sub, 11 Bluetooth low energy (BLE), 10
HTTP protocol, 8 Bluetooth stack, 11
LoRa, 10
Microsoft Azure system, 12
MQTT, 10 C
multi-hop, end-to-end communication, 7 Communication protocol, 7
Network Time Protocol (RFC1305), 13 Constrained Application Protocol (CoAP),
organization, 7, 8 11, 50
publish/subscribe model, 8 Continuous time Markov decision process
QoS parameters, 9 (CTMDP), 34
REST, 11 Control system, 3
RTPS, 9 Cost of ownership, 17, 18

D. Serpanos, M. Wolf, Internet-of-Things (IoT) Systems,
https://doi.org/10.1007/978-3-319-69715-4
92 Index

D transport and migration, 34–36

DARPA SHIELD chip, 20 workload curves, 26
Data Distribution Service (DDS), 9 Event latency, 3
Data link control (DLC), 84 Event populations, 30–32
Denial-of-service (DoS), 63 Event timing, 25
Dependable systems, 73
Devices, IoTs
chip cost, 20 F
communication power, 21 Factories of the future (FoF), 37
cost per transistor, 19, 20 False data injection (FDI), 60
design space, 17 Fog computing, 7
duty cycle model, 20, 21 Forwarders, 12
energy-scavenging technologies, 18 Fuzz testing
flash memory, 22 black-box, 80, 81
lifetime cost of ownership, 18 classification, 79
no-battery power management system, 18 configuration, 79
power consumption, 21 industrial control network systems, 82
RRAM devices, 22 limitations, 79 (see also Modbus)
Distributed control-oriented methods, 25 network vulnerabilities, 78
Distributed denial-of-service (DDoS), 63 simulation, 79
Duty cycle model, 20, 21 StackGuard, 78
Dynamic time warping (DTW), 13 static analysis, source code, 78
static and dynamic analysis, 79
white-box, 80
E
Embedded computing, 1
Energy-harvesting technologies, 18 G
Energy-scavenging technologies, 18 General Purpose Fuzzer (GPF), 81
Enterprise resource planning systems (ERP), Generic application security, 64, 65
37, 48 Google Cloud Pub/Sub protocol, 11
Environmental interaction modeling, 34 Google Cloud Storage databases, 12
Event-driven system analysis, 5 Grammar-based technique, 81
description, 25 Grey-box testing, 80
deterministic and stochastic event
timing, 25
devices, 28 H
distributed control networks, 30 Heuristic-based technique, 81
environmental interaction modeling, 34
event populations, 30–32
events and resources, 26 I
hubs, 28 Industrial control network systems, 82
internal timewheel, 27 Industrial control systems (ICS), 47, 50
leaf and non-leaf nodes, 25 Industrial Internet Consortium (IIC), 41, 42, 46
long-lived events, 27 Industrial Internet of Things (IIoT)
multi-hub networks, 29 appropriate architectures, 38
network models, 30 basic technologies, 49, 50
nodes and links, 28 business models, 45, 46
periodic event stream, 26 communication-centered functionality, 44
physical networks, 30 communication methods, 43
single-hub networks, 29 domains, 47
stochastic event populations, 32, 33 efficient and effective operations, 42
system architecture, 27 ERP and MES, 37
Index 93

functional viewpoint, 47 event latency, 3

General Electric company, 41 event loss rate and buffer capacity, 3
generic functionality, 45 event throughput, 3
human activity, 37 general-access devices, 1
ICS, 47 MEMS sensors, 2
IIC14, 41 news item and marketing trend, 1
IIC reference architecture functional possibilities, 1
domains, 46, 48, 49 power consumption, 2
implementation viewpoint, 49 privacy, 5
industrial process efficiency, 41 (see also reactive system, 3
Industrie 4.0) reliability and availability, 3
IoT and Industrie 4.1 relationship, 42 security and safety, 2, 5
ISO/IEC 27000 and ISA/IEC 62443 sensor network research, 1, 3
families, 38 service latency and throughput, 3
ITU, 42 service lifetime, 3
management vertical, crosscutting layer, 45 smart refrigerators, 1
network layer, 44 VLSI digital and analog electronics, 2
OT systems, 38 VLSI technology and Moore’s law, 4
“outside-in” installation sequence, 52 wireless networks, 4
PLC and SCADA, 37
power optimization, consumers, 51
reference architecture, 42, 44 K
service and application support layer, 45 Key-value pair, 25, 28
smart factories, 37
usage viewpoint, 46
Industrie 4.0 L
computational and communication LoRa protocol, 10
resources, 39
customers, 41
cyber-physical systems, 39 M
deployed semiconductor memory, 39–40 Manufacturing execution systems (MES), 38, 48
electrical energy, 39 Markov decision process (MDP), 34
embedded systems and communication Message Queue Telemetry Transport
networks, 40 (MQTT), 50
industrial information technology, 39 Microelectromechanical (MEMS) sensors, 2
manufacturing and production sectors, 39 Microsoft Azure system, 12
smart factory concept, 40 Modbus Application Protocol (MBAP), 84
transportation management authorities., 40 Modbus protocol
International Telecommunication Union application packet, 83
(ITU), 42, 43 authentication, confidentiality, integrity, 84
Internet-enabled physical devices, 1 client/server (alternatively, master/slave)
Internet of Things (IoT) communication, 82
alert system, 3 encapsulated application packets, 84
analysis system, 3 industrial control system communication, 82
applications, 2, 3 MBAP, 84
Bluetooth stack, 11 PDUs and DLC, 84
characteristics, 2 public codes, 83
conceptual devices, 1 reserved codes, 83
control system, 3 SCADA and PLCs, 82
embedded computing, 1 stacks, 83
event-driven/aperiodic sampling, 4 table types, 84
event-driven systems, 5 user-defined codes, 83
94 Index

Modbus/TCP fuzzer (MTF) S

attack, 85 Security, IoTs
boundary memory, 85 authentication, 58
description, 85 complex systems and services, 55
failure detection, 85 computational and data attacks, 60
industrial protocols, 86 confidentiality, 58
master and slave devices, 85 control loop, 56
memory types, 85 dependability, 58
reconnaissance, 85 depicted layering, 59
valid request/response pairs, 86 embedded and CPS systems, 58
MQTT protocol, 10 FDI attacks, 60
Multicast DNS (mDNS), 50 hierarchical computing structure, 57
Multi-hub networks, 29 HVAC control system, 57
pharmaceutical substance overdose, 57
privacy protection and safety, 59
N property layers, 59
National Vulnerability Database (NVD), 82 SCADA and PLCs, 56
Network security, 62, 63 scientific and engineering methods, 58
Network steering, 10 stakeholders, requirements, 58
Network Time Protocol (RFC1305), 13 technologies and disciplines, 55
noSQL, unstructured databases, 12 Security testing systems
consumer electronics devices, 77
customers and certification authorities, 77
O (see also Fuzz testing)
OpenFlow software-defined networking hardware and software testing, 78
protocol, 9 infrastructures, 78
integrated systems, 77
requirements, 78
P verification and validation techniques, 77
Power production and distribution Sensor networks, 1, 3
systems, 51 Single-hub networks, 29
Privacy protection, 72, 73 Soft real-time sensor networks, 1
Programmable logic controllers Stochastic event populations, 32, 33
(PLCs), 37, 56 Supervisory control and data acquisition
Protocol data units (PDUc), 84 (SCADA) system, 37, 56
Public codes, 83 Systems security
Publish/subscribe model, 8, 11 anti-tamper technologies, 61
complex hardware systems, 61
embedded and cyber-physical systems, 61
R organization, typical IoT system, 60
Radio-frequency identification (RFID), 50 side-channel attacks, 61
Random technique, 81 software techniques, 61
Reactive system, 3 stand-alone systems, 60
Real-Time Publish/Subscribe Protocol tamper-response methods, 61
(RTPS), 9
Reconnaissance, 85
Reliable-and-secure-by-design IoT T
applications, 66, 67 Timewheels
Reserved codes, 83 cloud analytics system, 29
Resistive RAM (RRAM) devices, 22 hubs, 25
REST protocol, 11 simulator event activity, 27
Run-time monitoring, 67, 68 single-hub network, 29
Run-time security monitor (RSM), 70 time-sorted queue, 27
Index 95

U White-box testing, 80
User-defined codes, 83 Wireless networks, 4
Wireless sensor networks (WSNs), 50

W
White-box fuzzing X
code execution paths, 80 XMPP protocol, 11
symbolic execution, 80
taint analysis techniques, 80
techniques, 80 Z
Windows and Linux applications, 80 Zigbee () network, 10