PROTOCOLS DEVELOPMENT FOR SECURITY AND PRIVACY OF RADIO
FREQUENCY IDENTIFICATION SYSTEMS
By
Fatin Sabbagha

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Packaging - Doctor of Philosophy
2014

ABSTRACT
PROTOCOLS DEVELOPMENT FOR SECURITY AND PRIVACY OF
RADIO FREQUENCY IDENTIFICATION SYSTEMS
By
Fatin Sabbagha
Radio frequency identification (RFID) is a technology that uses communication via radio
waves to exchange data between an RFID reader and an electronic tag attached to an object
for the purpose of identification and tracking. Although there are many benefits of adopting
this technology, there are many methods of attack that can be used to compromise the
system. The intent of this research is to determine how that may happen and what possible
solutions can keep that from happening. With this solution, protocols were developed to
implement better security. In addition, new topologies were developed to handle the problems
of the key management. Most of the previously proposed protocols focus on providing
mutual authentication and privacy between RFID readers and tags. However, these proposed
protocols are still vulnerable to be attacked. As a result, these protocols were analyzed and
the drawbacks shown for each one.
The previous works, from the beginning, assumed that the channels between readers
and the servers are secured. In the newly proposed protocols, a compromised reader is
considered along with how to prevent the tags from being read by that reader. The new
protocols provide mutual authentication between readers and tags and, at the same time,
remove the compromised reader from the system in efficient way.
Three protocols are proposed; in each one, a new idea is achieved. In the first protocol, a
mutual authentication is achieved and a compromised reader is not allowed in the network.
In the second protocol, the concept is similar to the first protocol, but the number of times

a reader contacts the server is reduced. Therefore, a new protocol is defined by a sharing
key between the server and the tags. In this case, key leakage is prevented with the compromised reader. In the third protocol, there is another concept to provide authentication and
privacy between tags and readers using a trusted third party. A session key is generated and
used to encrypt and decrypt messages between the tag and the reader using symmetric key
cryptography. Here also, key leakage is prevented with the compromised reader.
The developed topology is implemented using python language. The goal from this implementation is to realize and simulate work and to check the efficiency regarding the processing
time. Also, the three protocols are implemented by writing codes in C language and then
compiling them in MSP430. Its function the same as RFID tag. IAR Embedded workbench
is used, which is an integrated development environment with the C/C++ compiler to generate a faster code and to debug the microcontroller. In summary, the goal of this research
is to find solutions for the problems on previous proposed protocols, handle a compromised
reader, and to solve key management problems.

Copyright by
FATIN SABBAGHA
2014

To my husband Bashar Shaffo and my kids (Reeta and Matti) May God bless you with
His peace

v

ACKNOWLEDGMENTS

The research and writing of this dissertation has been one my greatest challenges, and there
is no doubt that it could not have been completed without the help and encouragement of
the following people, whom I have the privilege to acknowledge here. I would like to thank
Professor Robert Clarke for being my advisor, despite his many other professional commitments. I have been inspired by his wisdom, breadth of knowledge, and his commitment to
the highest academic standards.
Of course, my deepest appreciation goes to my family. My loving husband Bashar has
demonstrated unending devotion and sacrifice to make these pages possible. I also must
thank my dear children Reeta and Matti for their endurance on this journey. I am pained
by the time that we spent apart, and I am enjoying spending more time together with you
now. Your consolation is that never again will you have to hear me say that I need to go
work on my dissertation.

vi

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

x

Chapter 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 Auto ID Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.1 Barcode Systems . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.2 Biometric Identification . . . . . . . . . . . . . . . . . . . . . .
1.1.3 Machine Vision Technology . . . . . . . . . . . . . . . . . . . .
1.1.4 Card Technologies . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.5 Radio Frequency Identification Systems . . . . . . . . . . . . . .
1.1.5.1 Components of RFID devices . . . . . . . . . . . . . .
1.1.5.2 Read/write capacity . . . . . . . . . . . . . . . . . . .
1.1.5.3 RFID tag types . . . . . . . . . . . . . . . . . . . . . .
1.1.5.4 Radio frequency classifications . . . . . . . . . . . . . .
1.1.5.5 Key differences between barcode and RFID technology
1.1.5.6 RFID applications . . . . . . . . . . . . . . . . . . . .
1.1.5.7 Standards . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.6 Architecture Framework . . . . . . . . . . . . . . . . . . . . . .
1.2 The contribution and thesis organization . . . . . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

1
1
1
5
6
7
10
12
14
15
17
21
22
24
26
33

Chapter 2 Security and Privacy . .
2.1 Security properties . . . . . . .
2.2 Security in RFID systems . . .
2.2.1 Issues . . . . . . . . . .
2.2.2 Some proposed solutions:
2.3 Key management . . . . . . . .
2.4 Other solutions . . . . . . . . .
2.4.1 Proposed protocols . . .

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

34
34
37
37
46
54
57
58

Chapter 3 Methodology . . . . . . . . . . . . . . . . .
3.1 The Application . . . . . . . . . . . . . . . . . . .
3.1.1 An attack scenario . . . . . . . . . . . . .
3.2 System modeling . . . . . . . . . . . . . . . . . .
3.2.1 The Tree Key Graph . . . . . . . . . . . .
3.2.2 The Developed System . . . . . . . . . . .
3.3 The Communication Between Readers and Servers

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.

69
69
72
74
75
80
82

vii

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

3.4

The Developed Protocols
3.4.1 First Protocol . .
3.4.2 Second Protocol .
3.4.3 Third Protocol .

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

Chapter 4 Security and Time Analysis with Conclusion and Future works
4.1 Security analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.1 Analyzing the first protocol . . . . . . . . . . . . . . . . . . . . . . .
4.1.2 Analyzing the second protocol . . . . . . . . . . . . . . . . . . . . . .
4.1.3 Analyzing the third protocol . . . . . . . . . . . . . . . . . . . . . . .
4.2 Time analysis of the topology . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3 Operations in an RFID tag . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4 Time analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4.1 The time calculated in the first protocol . . . . . . . . . . . . . . . .
4.4.2 The time calculated in the second protocol . . . . . . . . . . . . . . .
4.4.3 The time calculated in the third protocol . . . . . . . . . . . . . . . .
4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6 Future Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
REFERENCES

83
83
88
92
96
96
97
98
100
103
104
106
106
110
111
116
120

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

viii

LIST OF TABLES

Table 1.1

Outline of the evolution of the Smart Card [26] . . . . . . . . . . . .

8

Table 1.2

Type of frequency and associated bandwidth [33] . . . . . . . . . . .

17

Table 1.3

UHF operating frequencies throughout the world [33, 44] . . . . . .

17

Table 1.4

Comparison between RFID frequencies according to [44, 52] . . . . .

19

Table 1.5

Comparison between RFID and Barcode [1, 5, 12] . . . . . . . . . .

21

Table 1.6

RFID Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

25

Table 3.1

The Steps For The Authenticated Reader in the first protocol . . . .

86

Table 3.2

The Steps For The Compromised Reader in the first protocol . . . .

88

Table 3.3

The Steps For The Authenticated Reader in the second protocol . .

91

Table 3.4

The Steps For The Compromised Reader in the second protocol . .

91

Table 3.5

The Steps For The Authenticated Reader in the third protocol . . .

94

Table 4.1

The goals of the three developed protocols that have been achieved

Table 4.2

The time duration in microseconds of the second proposed protocol . 115

Table 4.3

The time duration in microseconds of the first and third proposed
protocols when the number of digit in a reader increases . . . . . . . 115

Table 4.4

The time duration in microseconds of the first and third proposed
protocols when the number of digits in a tag increased . . . . . . . . 116

Table 4.5

The time duration in microseconds of the first and third proposed
protocols when the random number of a tag and a reader were equal 116

Table 4.6

The differences between the three developed protocols . . . . . . . . 118

ix

102

LIST OF FIGURES

Figure 1.1

Universal Product Code type A [19, 20, 47] . . . . . . . . . . . . . .

2

Figure 1.2

Finder Pattern and the data [35] . . . . . . . . . . . . . . . . . . . .

4

Figure 1.3

RFID System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

12

Figure 1.4

Eavesdropping range (Forward vs. backward channels) [43, 56] . . .

13

Figure 1.5

The EPCglobal Architecture Framework (Roles and Interfaces) [17, 32] 29

Figure 2.1

Security Objectives Dependencies adapted from[57, 58] . . . . . . .

36

Figure 2.2

RFID Malware test platform [60] . . . . . . . . . . . . . . . . . . . .

38

Figure 2.3

Tracking System . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41

Figure 2.4

Replay Attack [63] . . . . . . . . . . . . . . . . . . . . . . . . . . . .

42

Figure 2.5

Jamming Radio Frequency . . . . . . . . . . . . . . . . . . . . . . .

43

Figure 2.6

Man-in-the-middle Attack . . . . . . . . . . . . . . . . . . . . . . . .

45

Figure 2.7

Minimalist Cryptography Technique . . . . . . . . . . . . . . . . . .

48

Figure 2.8

Anti-counterfeiting phases [79] . . . . . . . . . . . . . . . . . . . . .

52

Figure 2.9

Key’s life-cycle [84] . . . . . . . . . . . . . . . . . . . . . . . . . . .

54

Figure 2.10

Hash-Locking: A reader unlocks a hash-locked tag [87] . . . . . . . .

60

Figure 2.11

Randomized Hash-Locking: A reader unlocks a tag whose ID is k in
the randomized hash-lock scheme [87] . . . . . . . . . . . . . . . . .

61

Figure 2.12

The checkout protocol in PAP [14] . . . . . . . . . . . . . . . . . . .

62

Figure 2.13

Comparison between PAP and IPAP in-store protocol [88] . . . . . .

63

x

Figure 2.14

The checkout protocol in IPAP [88] . . . . . . . . . . . . . . . . . .

64

Figure 2.15

The checkout protocol in RFIDGuard [90] . . . . . . . . . . . . . . .

65

Figure 3.1

An Attack Scenario in Port A . . . . . . . . . . . . . . . . . . . . .

72

Figure 3.2

A tree key graph [96] . . . . . . . . . . . . . . . . . . . . . . . . . .

76

Figure 3.3

Developed RFID System Model . . . . . . . . . . . . . . . . . . . .

79

Figure 3.4

LLRP endpoint [99] . . . . . . . . . . . . . . . . . . . . . . . . . . .

81

Figure 3.5

First Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

84

Figure 3.6

Second Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

89

Figure 3.7

Third Protocol

93

Figure 4.1

The execution time for deleting a compromised reader form the system103

Figure 4.2

The execution time for deleting a compromised reader form the system randomly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

Figure 4.3

Times, while increasing the number of digits in RFID tags only in
the first protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

Figure 4.4

Times, while increasing the number of digits in RFID readers only in
the first protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Figure 4.5

Times, while increasing the number of digits in both RFID tags and
readers equally in the first protocol . . . . . . . . . . . . . . . . . . 109

Figure 4.6

Times,with random digits in both RFID tags and readers in the first
protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Figure 4.7

Times, while increasing the number of digits of the random number
in RFID readers in the second protocol . . . . . . . . . . . . . . . . 110

Figure 4.8

Times, while generating random numbers with random digits in RFID
readers in the second protocol . . . . . . . . . . . . . . . . . . . . . 111

Figure 4.9

Times, while increasing the number of digits in RFID tags only in
the third protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

Figure 4.10

Times, while increasing the number of digits in RFID readers only in
the third protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

Figure 4.11

Times, while increasing the number of digits in both RFID tags and
readers equally in the third protocol . . . . . . . . . . . . . . . . . . 114

Figure 4.12

Times, with random digits in both RFID tags and readers in the third
protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

Figure 4.13

Customs using RFID system with secured protocols . . . . . . . . . 117

xii

Chapter 1
Introduction

1.1

Auto ID Technologies

Automatic identification technologies (Auto ID) are very popular in many services systems.
To increase the efficiency and productivity in many organizations, Auto ID can be used
because it runs procedures faster, is more accurate, and it eliminates error associated with
identification and data collection. Auto ID technologies can be used to provide information
about numerous things, such as people, animals, goods, and so on. It can also reduce
many problems related to life such as missed delivery dates, customer dissatisfaction, lower
productivity, etc. There are different types of Auto ID systems based around different
strategies and functions. The following sections are a brief overview of these systems.

1.1.1

Barcode Systems

This is the most common ID technology for encoding information. Some applications are
inventory control, Customs control points, work in progress, security, point of sale, account
receivables, marketing and business replies, time and attendance, etc. In general, there are
one dimensional barcodes (1D) and two dimensional barcodes (2D). 1D is made up of parallel
bars and spaces, which could be wide and narrow. Each character represented by a barcode
is a series of bar and space. All barcodes contain parts; these are a quiet zone, start code,

1

data, stop code, and a trailing quiet zone. Also, some 1D barcodes have check digits for
providing integrity. In addition, the density of the barcode can vary from low, to medium
to high density. The density can be determined by the X dimension, which determines the
width of the narrowest bar or space in the barcode.

Figure 1.1: Universal Product Code type A [19, 20, 47]

Barcodes can be read by different types of scanning readers. Some are contact readers
and others are non-contact, such as wands, charge coupled device (CCD), fixed focus optics
(FFO), and laser. Contact scanners needs physical contact, but non-contact scanners can
read items up to several inches away. Barcodes can be represented by different symbologies,
such as Code 39, Code 128, Universal Product Code (UPC), European Article Number/

2

Japanese Article Number (EAN/JAN), Code 93, Interleaved 2 of 5 (ITF), MSI Plessey,
Telepen, PosiCode A and B, Channel Code, and Coda Bar. [53] Each of these symbologies
has different characteristics. Some have fixed length while others have variable length. For
example, the most common barcode for retail product labeling is the UPC, which is a fixed
length 12 digit numbers only code. Figure 1.1 illustrates the syntax of this code.
2D is made up of individual dots or squares and it can be printed as a square, circular
or a rectangular symbol. It can be represented into two primary types: matrix codes and
stacked-bar codes. A stacked-bar code is a stack of linear barcodes, which can be read by
laser scanner, CCD or camera. A matrix code can be square, circular or hexagonal shape.
This type can be read by a camera or CCD reader. The difference between these two types
is related to the code density. Matrix codes have a higher density than stacked-bar codes
with identical information. [19, 20, 35]
2D comes in different forms such as the Quick Response (QR) code that was the earliest
2D barcode, Datamatrix, and 2D color barcodes that are the newest edition of 2D barcodes.
[34] The characteristics of 2D barcodes are high information density, high data capacity,
and wide coding range, for example it can encode image, sound, word, signature and finger
mark. [36] 2D consists of two broad areas: data areas and guide areas. The structure of
Datamatrix has two parts: the finder pattern and the data. The finder pattern defines the
shape of the 2D code, the size, the minimum size of the dots, and the number of rows and
columns in the shape. [35]

3

Figure 1.2: Finder Pattern and the data [35]

Figure 1.2 illustrates the shape of 2D Datamatrix. There are two parts in the finder pattern. The L finder pattern (the solid dark sides) defines the size, orientation, and distortion
of the symbol. The Clock Track (the light and dark elements) defines the basic structure
and also can determine the size and distortion. The maximum amount of encoded data in
a square form is 2,335 alphanumeric characters and 3,116 numbers. However, the maximum
amount of encoded data in a rectangular form is 72 alphanumeric characters and 98 numbers.
[35]
The differences between 1D and 2D barcodes is that 1D is used as an index to a database
to obtain the required information, while 2D can be a database unto itself with no need to

4

access as external database. 2D is used to encode more information (thousand characters) in
a smaller area than would be required of 1D codes. This can make procedures move quickly
and more reliably. [19, 20, 35]

1.1.2

Biometric Identification

Biometrics is a technology used to identify a person by using one of the uniquely human
characteristic features and comparing it with previously stored data. It is used for providing
authentication to get more sensitive information and resources. [22] There are other techniques used to provide authentication, but some of them such as passwords or documents
can be stolen or lost. Biometric technology, on the other hand, cannot be forgotten, lost or
stolen. [21] There are several types of biometric technologies, such as fingerprinting, hand
geometry, and retinal scans. In addition, voice recognition technology is a part of biometric
verification. [20] Each of these procedures can identify an individual correctly. Biometric
technologies can be used in many applications, for example in time and attendance, retail,
security and controlled access, computer access, couriers, banking, etc.
There can be devices installed at the secured area or entrance for getting information from
an individual. Although biometric is the best technology for identifying individuals, it still
has problems. It is costly compared with other technologies like card systems. That will be
discussed in the next section. Some other issues that the technology faces are the degradation
of biometric features over time, variance in recorded and actual biometric characteristics,
and threshold values for authentication and privacy. [19, 20]. Several studies and surveys
have been proposed to see how much individuals are aware of the technology and see the
benefits of it. [22] Because of increasing demand of high-performance security methods,
there are numerous studies about improving the image of biometric techniques. For example,
5

Maeva [37] introduces a newer development of the ultrasonic fingerprint imaging. Liang [38]
proposed a novel video based biometric identification model based on eye tracking technique.
Several visual attention characteristic are extracted from eye gaze data. Then, these features
are used to identify persons.

1.1.3

Machine Vision Technology

Machine vision is defined as the automatic extraction of information from digital images.
The purpose of extracting information could be for pattern recognition, part inspection, or
part positioning and orientation. It is a system that uses many types of advanced hardware and software components to perform the same tasks of inspectors who work to inspect
the quality of parts. However, this system provides high speed operations with greater
precision. Machine vision systems improve automated technical identification, inspection,
measurement, and guidance capability. [48] It can be used in many applications such as:
pharmaceutical packaging, food and beverage packaging, electronics manufacturing, the Department of Defense (DOD) supply chain, and so on. It is an essential part of a manipulative
task involving industrial operator (robots). The main four functions of machine vision are:
measurement, counting, location, and decoding. For measurement, it is used to check the
automated measurement by the machine to specified tolerances. Counting is another function of machine vision to look for a number of features on parts to see if there are missing
parts. Also, it checks the absence and presence of products in one package. Machine vision
is used to specify and report the position and the orientation of a part to specify tolerances.
The decoding function is to interpret, for example, 1D and 2D symbologies to extract some
information used later for tracking products. This information can also be used to verify
the correctness of data. These four functions play a role in saving cost, reducing defects,
6

tracking and tracing, and complying with regulations. [48] There are four components of a
machine vision system: lighting, frame grabber, image processing and analysis software, and
CCD cameras. [49] Companies can rely on the machine vision to improve the quality of their
products. For example, petroleum and petrochemical manufacturer Sinopec Group use this
technology to maximize production output, reducing costs, minimizing waste, and saving
power. [50] They used a smart camera to collect accurate data from 2D codes and text
characters on the bottle labels and 1D from the secondary packaging. Another case study
is when Wayne E. Bailey Company has experienced the benefits of using machine vision in
packing operation. Microscan QX830 laser barcode scanner is used to confirm the product
type (sweet potatoes), ensuring that each box receives the correct label. When the scanner
detects the wrong code, the product is removed from the line. The benefits from using the
technology are to increase labeling accuracy and traceability. [51]
Vision technology is used in packaging industries to inspect products for any defect, for
example in caps, seals, labels and so on. Companies get benefits from using the system
because it reduces cost and provides date and lot verification, color detection, robotic guidance, test tube cap and color inspection, error proofing, package integrity inspection, and
dimensional gauging. [48]

1.1.4

Card Technologies

Card technologies demonstrate an outstanding opportunity for improving comfort and security. There are three types: magnetic stripe cards, smart cards, and optical memory cards.
[39, 53] Magnetic stripe cards use a magnetic medium to encode data in binary format with
the polarity of the particles determining the 0 and 1 bit. A reader then translates the code
into alphanumeric characters. Magnetic stripe cards provide minimal security because data
7

can be read from and written to the card. In this case, the information can be stolen and
duplicated. With respect to capacity, it limits the amount of data to less than 2 Kbytes.
In addition, it needs contact read equipment. Because of these limitations, Magnetic stripe
cards are often replaced with smart cards. [40, 53] A smart card is a plastic card that uses a
microprocessor chip to store data, which is stored in binary format; the chip performs some
computing functions such as encryption and decryption functions using the data. In 1968,
smart cards were introduced by two German inventors, J¨
urgen Dethloff and Helmut J¨
urgen.
[25] In Table 1.1, a brief outline of the evolution of the smart card is presented.
Table 1.1: Outline of the evolution of the Smart Card [26]

Year

Event

1968

2 German inventors patent combining plastic cards with micro
chips

1970

Computations arise from the mathematical of Boolean logic(
AND, OR..etc.)Arimura invents and patents in Japan

1974

Roland Moreno invents and patents in France

1976

French DGT initiative, Bull (France) first licenses

1977

Motorola produces first smart card microchip

1979

Motorola develops first single chip microcontroller for bank in
France

1980

First trials in 3 French cities

1982

First U.S. trials in North Dakota and New Jersey

1991

AT&T declared its contact less smart card

1992

Germany uses smart card for health care

1996

First university campus deployment of chip cards in Poland

The components of this system are a smart card, reader, computer, software, and an
electronic meter. The principle behind the card technology is to provide authentication and
validation of the holder. There are many other advantages of using a smart card, some of
8

which are capacity, convenience, durability, and security. [24]
The stored information in the card can be protected from undesired access. There is a
password given to the holder to use whenever the card is connected to a computer. [24] Many
types of cards have been developed; these are contact based smart cards, contact less smart
cards, and combinations between these two types. A contact-based card contacts with the
reader to carry out functions. However, contact-less cards requires no contact with readers
because it has a chip and an antenna to receive power for communication and processing the
transaction. [27, 53]
Smart cards can be classified into two categories: a memory card or a processing -enabled
card. A memory card can be called asynchronous because the flow of data is in one direction;
data moves to the reader or to the computer system. The memory ranges from 8K- 128K
bit. [41]
In addition, this type of card provides a limited capacity to secure the stored data. [26]
The processor-enabled smart card or synchronous card is a more sophisticated card. It is also
called a “true” smart card because it is based on semiconductor technology. [28] It contains
a chip with a few hundred bytes of RAM. This type of card can perform cryptographic
operations so that the information will be protected by the encryption schemes with biometric
identification. The data flow is bi-directional because data can be read from and written to
the card. [26] This technology can be used widely in many applications such as bank cards,
ID cards, garage cards, health care cards, pre-paid cards, and hotel key cards and so on.
[19, 23, 53]
Smart cards have significant advantages over magnetic stripe cards especially in healthcare applications because it provides very high security and privacy; both contact and contact
less smart cards provide high level of security. They protect sensitive information for trans9

actions. In addition, smart cards have higher capacity than magnetic stripe cards. [40, 41]
Optical memory cards are another technology for securing identifications. It employs a
laser beam to write data to a reflective stripe of laser recording medium, composed of silver
particles suspended in a matrix. A laser reader reads tiny holes that are burned into the
medium; the absence and the presence of a hole represents 0 and 1 bit respectively. Optical
memory cards are a write once read many (WORM) type media and the data is non-volatile,
which means it is not lost when the power is removed. The storage capacity of this card is in
between 4Mbytes to 6.6Mbytes of data which is higher than the other card technologies. [53]
Because of the large capacity of this card, multiple features such as photographs, fingerprints,
and signatures can be recorded. This card can be used in some applications that need very
secure personal identification such as the US Permanent Resident Card (green card). [42]

1.1.5

Radio Frequency Identification Systems

In this section, the background of radio frequency identification systems (RFID) will be
introduced. RFID is one of those newer technologies that tend to make life easier; it is a
wireless device. RFID tags are small electronic devices that are used to identify objects to
which they are attached. Some other references define the RFID tag as a small computer
device because it has a storage unit and microcontroller. In operation, tags automatically
emit their unique serial numbers when they get queries for readers. [53]
In general, the benefits of RFID systems are:
1. Improving the efficiency of production processes
2. Improving product quality and service safety or authenticity
3. Improving product track-and-trace capabilities
10

4. Improving asset management
5. Reducing labor costs
6. Increasing supply chain efficiencies
7. Complying with customer mandates
8. Supporting new customer-facing strategies
9. Improving supply chain visibility
10. Facilitating collaboration with business partners
11. Improving inventory management efficiencies
12. Optimizing merchandise management and reducing out-of-stocks [54]
RFID systems can be used in many applications. This technology can provide suppliers,
manufacturers, distributors and retailers accurate information about their products. However, it may bring up some privacy threats. There are a set of security problems that an
RFID system suffers from, and one of the most critical one is cloning. Imagine what could
happen if someone successfully clones the contents of an RFID attached to a passport!

11

Reader
Insecure Channel
Secure Channel

Tag

Insecure Channel
Tag

Insecure Channel

Insecure Channel

Reader

Tag

Back-end data base

Figure 1.3: RFID System
1.1.5.1

Components of RFID devices

In general, the elements of an RFID system are a tag (transponder), a reader (transceiver),
computer with software, and a back-end database. See Figure 1.3. An RFID reader is usually
equipped with an RF transmitter and receiver, a control unit, and a memory unit. It is the
connection between tags and the back-end database.
The role of the reader is to interrogate a tag by sending a signal to read the stored
information on the tag. The tag would be then be authenticated based on the stored
database information. Tags obtain, read and transmit energy from the electromagnetic
field of transceivers. [16] Information is wirelessly exchanged, via radio wave, between the

12

reader and the tag, both of which are equipped with an antenna. In addition, a tag typically
has a microchip and encapsulation or protective layers. If tag is active, it will also have an
internal power supply (a battery), which boosts read capabilities. [15]
There are two types of readers. The first type is fixed (or stationary) readers, which
have a fixed location in a plant and are used when tags are passing within the range of the
reader. The second type is mobile readers, which can be moved around and are often used
for inventory control purposes. [15, 16]
The back-end database is divided into two parts: middleware and applications. These
two parts are run on computers within the network. Middleware is used to filter data and
provide the interface towards the applications. [16]
Forward Channel

Tag

Reader

Eavesdropper

Backward Channel

Backward range (Eavesdroppers within the range can be seen)
Forward range (Eavesdroppers can monitor the signal without detection)

Figure 1.4: Eavesdropping range (Forward vs. backward channels) [43, 56]

Generally, it is assumed that the channel between a reader and back-end database is
secure, but the channel between a reader and tag is not [1]. There are two channels between
13

a reader and tag, forward and backward. The forward channel is for interrogating a tag that
in turn responds to a reader through the backward channel. Because a tag gets its power
from the electromagnetic reader signal, the backward channel signal is much weaker than the
one in the forward channel. Because of this, any passive party can eavesdrop on a message
in the forward channel easier than one in the backward channel. See Figure 1.4.

1.1.5.2

Read/write capacity

RFID tags can come in many different formats. They can be read only (data cannot be
changed; it is programmed once by the tag manufacture), read/write (data can be modified or
rewritten over existing information), or a combination (data is stored permanently and some
part of the memory will be used for other later updating). [16, 18] Some tags unwritten and
then sent to packaging houses who encode the tags with some information about products.
These kinds of tags can be considered as WORM, which stands for Write Once, Read Many.
Others can be read only and WORM where manufacturers prewrite their information on
tags and then send them to be filled in at the packaging houses. [16, 18]
The supply chain applications and the technologies can be controlled by using read/write
tags. If not, new tags should be purchased in cases where the requirements are changed.
Read/ write tags can have greater stored information capabilities that lead to an increase
in the processing time. These tags may not be required to contact the back-end database
because they can have all relevant data embedded therein. Furthermore, they can be used in
different applications by erasing old information and rewriting them with new information.
This can lead to cost savings and provides flexibility to adapt to future changes. [16, 18]
The reading ability typically decreases when the distance between a tag and its reader
increases. Moreover, it will decrease due to differing factors. Some factors are the operating
14

frequency employed, type of tags, size of antennae (in both readers and tags), alignment of
antennas, the environment (with local materials and conditions), and so on. Based on these
factors, the reading range can be affected as well as the writing range. In this scenario, the
exchange of protocols between tags and readers will be affected too. [16, 18]
The reading speed will also be affected by different factors, some of which are:
1. Number of messages to be exchanged
2. Available data rate (frequency)
3. Error correction strategy and desired reliability
4. Anti-collision algorithm and number of tags within communication range
5. Amount of data to be transmitted
6. Required time for performing calculations especially for the authentication algorithms
[16]
Using low frequency (LF) and high frequency (HF) systems, 10-30 tags per second can
be read. When ultra-high frequency (UHF) is used, 100-500 tags per second can be read.
[1, 14, 16]

1.1.5.3

RFID tag types

Tags come in many different sizes and shapes. Some are very tiny and some have much larger
dimensions. The RFID tags can be classified into three kinds based on how tags are fed with
power. The first kind is active RFID tag, where a power source (battery) is located on the
RFID tag and has extra memory storage so that it has more energy and a more powerful

15

computational ability [1]. The second kind is passive RFID tag, which implies that the tag
has no internal power source (hence, it comes without a battery). There is no doubt that this
type of tag needs a source of power to operate and this would be derived from the received
signal energy of the reader. A passive tag has limited computational storage that can be
applied. Finally, there are semi-passive tags that use some battery power to maintain their
internal volatile memory. However, it gets its power from electromagnetic signals that are
sent by the reader, just like passive tags. [17] This tag may also be referred to as semi-active.
It can provide a longer read range than a passive tag because of the battery. In addition,
a semi-passive tag can accommodate environmental sensors on-board. The environmental
experience of any object will be recorded to the attached tag memory during the lifecycle of
that object. This can happen because of the battery. Although the batteries in semi-passive
tags have advantages, they also create a few problems such as: extra weight, larger size,
higher cost, shorter life, and so on. [33]
Because of the longer range of active and semi-passive tags, they are often used for
tracking high-value goods that need to be scanned over long range, such as cargo containers
or railway cars on a track. However, their price is too expensive to put on low-cost items.
[16, 17]
The advantages of the passive tags over active tags are low costs, smaller sizes, and lower
weights. Finally, because a passive tag does not have a battery, it has an extended lifetime.
In contrast, the advantage of active tags over passive is the communication will occur over
longer distances (signal strength is strong and always on), greater memory capacity, and
the ability to be rewritten. However, active tags have a limited life because of batteries,
though a lifespan of several years is common. Both passive and active tags can be used in
all environments. [16, 17]
16

1.1.5.4

Radio frequency classifications

Diverse frequencies ranges, which may have different characteristics, can be useful for various
RFID systems applications. There are four classes of commonly used tags, based on the
frequency: low frequency / LF , high frequency / HF , ultra-high frequency / UHF , and
microwave /SHF. [1, 14, 16] All of these ranges are part of the frequency bands called
Industrial, Scientific, and Medical (ISM) radio bands. [33, 53] Tables 1.2 and 1.3 illustrate
the range of each frequency band.
Table 1.2: Type of frequency and associated bandwidth [33]

Range

LF

HF

UHF

Microwave

Frequency
available

30-300 KHz

3-30 MHz

300-1000 MHz

1-6 GHz

Used for RFID

125—134
KHz

13.56 MHz

433 & 860-960
MHz

2.4 & 5.8 GHz

Table 1.3: UHF operating frequencies throughout the world [33, 44]

Country
United
Canada

Frequency Band (MHz)
State

and

902-928

Australia

918-928

Europe

865-868

Hong Kong

865-868 & 920-925

Japan

952-954

Korea

908.5-914

New Zealand

864-929

Singapore

866-869 & 923-925

The range for LF tags is short (a few inches at best) and the data transfer rate is the
17

lowest. LF tags are passive tags and store a small amount of data. They can be easily
read when they are attached to objects made of or holding water, wood, metal and other
liquids. They are used in many applications such as animal identification, access control,
asset tracking, automotive control, healthcare, and so on. LF tags have limited anti-collision
capability. Thus, it is very difficult to read multiple tags simultaneously. [33] An anticollision technique is required to ensure that tags and readers can communicate using the
same shared medium without disturbing each other. There are two algorithms for this
technique, probabilistic and deterministic anti-collision algorithms. Probabilistic algorithms
use the ALOHA protocol, which works as follows: when the communication medium is free,
a device starts to send information. If the collision occurs (several devices send data at the
same time), the medium will released all the stations and then each station waits for a random
time when they will start again. For the second type of algorithm, which is deterministic,
the binary tree walking algorithm is used between readers and tags. This algorithm works
as follows: each tag has a unique identifier; the reader controls the process by requesting all
tags whose address start with 0 bit; if the collision occurs (many tags answer), the reader
asks all tags whose address begins with 00. This process will continue until only one tag
answers. [16, 52]
HF tags (13.56 MHz) have a short reading range (maximum range about 3 feet) and the
data rate is higher than LF, but lower than UHF. HF tags can be passive or active tags. In
addition, HF tags may have anti-collision capabilities, which lead to facilitate the reading of
multiple tags. The cost of HF tags is lower than LF because it has simpler antenna design
than LF tags. They also can be read when attached to objects made of, or holding water,
wood, metal and other liquids. However, they can be affected by metal objects in the close
vicinity. Because of magnetic flux, which is used to power and communicate with the tags,
18

HF tags can be used in many applications such as a smart shelf because it covers the entire
area. HF tags are used in such different application such as credit cards, smart cards, library
books, airline baggage, and asset tracking. LF and HF tags both use near-field inductive
coupling to obtain power and communicate. [33, 44, 52]
UHF 433 MHz is used mostly for active tags, while UHF 860-960 MHz is used for passive,
active, and semi-passive tags. The range of passive UHF, for example, is up to 135 feet. [55]
UHF tags also have an anti-collision capability. They cannot be easily read when attached to
objects that hold or are around water and animal tissues because water absorbs UHF waves
and detuned the tag. Also, when the tags are around or attached to metal objects, they
become detuned. The reading capability will be impossible when a conductive material is
place between the interrogator antenna and the tags. UHF tags (passive and semi-passive)
typically use far-field backscatter coupling, though newer UHF tags are available that use
inductive coupling. [33, 44, 52]

Table 1.4: Comparison between RFID frequencies according to [44, 52]
Band
Frequency

LF
30-300 KHz

Typical
RFID
frequencies
Approximate
read
range
[52]

125-134 KHz

Typical data
transfer rate

Less than 1
Kbit/s

Less than 0.5
meter

HF
3-30 MHz

UHF
300 MHz- 3
GHz
13.56 MHz
433 MHz or
865-956 MHz
2.45 GHz
Up to 1.5 me- 433 MHz =
ters
up to 100 meter 865-956
MHz =0.5 to
5 meter
Approximately 433-956 =30
25 Kbit/s
Kbit/s 2.45
=100 Kbit/s

Continued on next page

19

Microwave
2-30 GHz
2.45 GHz

Up to 10 m

Up to
Kbit/s

100

Table 1.4 (cont’d)
Band
LF
Characteristics Short range,
low
data
transfer rate,
read
easily
when
they
are attached
to
objects
made
of,
or
holding,
water, wood,
metal
and
other liquids
Typical use

Animal
ID
Car Immobilizer

HF
Higher
ranges,
reasonable
data
rate
(similar
to
GSM phone),
read
when
attached to
objects made
of, or holding
water, wood,
metal
and
other liquids
Smart labels
Contact-less
Travel cards
Access
&
security

UHF
Long ranges,
high
data
transfer rate,
concurrent
read of < 100
items, cannot
penetrate water or metal,
but can with
non-water
based liquids
like oils

Microwave
Long range,
high
data
transfer rate,
read
easily
with metallic
objects

Animal
Tracking
Logistics

Moving vehicle toll

Most microwave tags use 2.54 GHz. Microwave tags can be passive, active or semi-passive
types. Usually, passive microwave tags are smaller than passive UHF tags, but the read range
for both is the same. [45] The read range of semi-passive microwave tags is about 100 feet
and 350 feet for active microwave tags. Microwave tags (passive and semi-passive tags) use
backscatter coupling to communicate, but active tags use their own transmitter. Microwaves
tags can work easily with metallic objects because of their shorter wavelength. [33, 52]
The read ranges for tags can be attained by different aspects. Some of these aspects are
frequency, the power supply, the size and form of antennas, the alignment of antennas, the
environment, and so on. [16] Table 1.4 shows the comparison between the various frequencies
for RFID systems. [33, 52]

20

Table 1.5: Comparison between RFID and Barcode [1, 5, 12]

RFID

Barcode

RF technology

Optical technology

May not need line-of-sight

Need line-of-sight

Proceed in bulks

Proceed one by one

Identify objects at item level

Identify objects at type level

High-up to $50 per tag, averaging $0.10$5.00 /tag

Less expensive

It can be extremely durable (internally attached)

Can be damaged (dirt, torn, etc.)

May be difficult to read when RF is passing through metal or liquid

Most can be read when it is damaged but
only a very few cannot be read when they
damaged

Can be read, rewritten

Can only be read

May contain high level of security (data
can be encrypted, password, etc.)

Can be duplicated and counterfeited;
have lower levels of security

1.1.5.5

Key differences between barcode and RFID technology

The function of an RFID tag is the same as the barcode but RFID usually has a globally
unique identifier for each item on which it is attached. RFID has some significant advantages
over barcodes; it may not need physical or visual contact between the reader and the RFID
tag i.e., it may not need line of sight, unlike barcodes. Also, large items in crates could be
scanned at the same time if the items are RFID labeled, unlike the use of barcodes where
each package needs to be scanned individually. In many cases, the supply chain system
benefits from this advantage of RFID tags.
In summary, Table 1.5 has the comparisons between RFID and barcode systems. [1, 5, 12]

21

1.1.5.6

RFID applications

As it has been mentioned above, RFID can be used in many applications in industries. The
RFID components are used to identify and track objects. This system is used in many
applications such as supply chain management, theft prevention, military, animal tracking,
medical system, automatic payment and E-passport and so on. With a unique ID, the
transportation of any cargo can be monitored in the entire value chain, from manufacture
to retailer. Also, because RFID has many benefits, for example, no human intervention
or being recyclable, the cost of the supply chain could be reduced and safety/security can
be achieved. [1, 2] The use of this technology could result in a 90% decrease in location
errors. [46] The data collection from RFID system also helps preventing errors in picking
and shipping. The use of standardized assessment tools, technologies and processes are
essential for improving supply chain security strategies. A track and trace system for all
products throughout the domestic and foreign supply chain is a significant step for ensuring
transparency and accountability of product authenticity and distribution. This could be
done by using RFID, which provides a secure identification and can track a product at every
point in the supply chain quickly. [7] Anti-theft security RFID labels play a key role in
deterring counterfeit activity as well as detecting and monitoring the actions of would-be
criminals and the movement of goods into and out of a retail establishment. This technology
also has many benefits in establishing an electronic pedigree. The FDA in food defense is
focusing on moving from a reactive approach to proactive approach to address the intentional
and unintentional food contamination. There are four goals of this food defense. The first
one is to prevent an outbreak/attack through awareness, preparedness and capacity building.
The second goal is intervention through targeted inspection and sampling. The third goal is

22

to respond rapidly and efficiently if needed. The last goal is to recover rapidly and restore
consumer confidence in the food supply. The FDA Combating Counterfeit Drugs identified
several key elements for combating counterfeit crimes and securing the nation’s drugs supply
by providing technology and increasing criminal penalties, whereas the FDA food initiative
is focusing on prevention and detection of the fraud. They both intersect in using technology
(RFID and barcode) for tracking products. Also, they both can identify what products are
likely to be counterfeited and entered into the legitimate supply chain, even if the prevention
of that is impossible. [8, 9] In medical health, RFID can be used to obtain and track a
patient’s information such as ID, name, emergency contacts and so on.[1, 5] Also, RFID
can be used to help surgeries in tracking surgery sponges or any tools on which RFID tags
are attached. There have been cases where doctors accidently left sponges and tools inside
patients, which cause damage to the patients. Therefore, a wand reader will be used on the
patient after surgery to determine the location of the sponges, if any are missed in closing.
[3, 4] In addition, RFID can be adopted to identify counterfeit medical drug packaging.
RFID can also be used to improve the security of credit cards because of counterfeiting. So,
before payment is made, a reader should authenticate the tag and card. Another example of
automatic payments in which RFID can be used is the public transportation ticket system.
This reduces the costs from selling tickets manually. Another example about using RFID is
automatic toll readers. When cars drive along a road, tags in the cars will be read and charged
a fee. This benefit increases throughput efficiency, and accuracy, and it is more convenient
than collecting fees from people. [1, 2] In an E-passport, RFID tag can be embedded to
record more information about the holder. This information could be related to fingerprint
and more sensitive details. Information will be recorded for people who are entering and
leaving the U.S. However, security should be provided to eliminate the counterfeiting, such
23

as having the information be encrypted. [1, 2, 6] RFID also has many benefits in animal
scientific research or pet tracking. In pet tracking, this technology is used to locate and
monitor the location of animals such as cows, pigs, cats, dogs or even fish, when they are
lost. In some other research, it can be used to monitor animal diseases [1, 2].
The DOD uses passive and active tags within its supply chain and with all contracts. All
cases and pallets that are shipped to the overseas DOD locations should be tagged. RFID
can also support soldiers in their fields. [10] RFID has three main benefits for the DOD.
These benefits are related to the supply chain management, asset tracking, and security. It
can help to improve visibility and the efficiency of the vast DOD supply chain. The military
can identify, track and manage their belongings in time. In addition, RFID systems secure
suppliers, equipment, and locations. However, there are also issues related to security. These
will be discussed more in Chapter Two. [11]

1.1.5.7

Standards

To deliver accurate data and information within supply chain, standards are provided by the
GS1. These standards are used by businesses to operate organizations efficiently by sharing
their information (location, movement and status) about their products and to speak the
same language as their trading partners.
There are many standards that have been developed in RFID system for different frequencies and applications. Some of these standards are related to the communication between
tags and readers and some are related to the formatting of the data or conformance between products and the standards, or applications. [29] In addition, the security of tags on
products can be included in some standards. Wal-Mart and the DOD are the two largest
USA drivers for using RFID technology today and to accomplish the same thing. The In24

Table 1.6: RFID Standards
ISO 14443
ISO 15693
ISO 18000 part 1
ISO 18000 part 2
ISO 18000 part 3
ISO 18000 part 4
ISO 18000 part 5
ISO 18000 part 6
ISO 18000 part 7
ISO
ISO
ISO
ISO
ISO
ISO
ISO
ISO
ISO
ISO
ISO

11748/ 11785
11785
17358
17363
17364
17365
17366
17367
10374.2
11784
18047

ISO 18046

For Proximity Cards (A few inches)
For vicinity (Out to 28 inches (from a single antenna) or 4 feet (multiple antenna))
For RFID Air interface communication
(Globally accepted frequencies)
For RFID Air interface communication
(Low frequency below 135KHz)
For RFID Air interface communication
(High frequency at 13.56 MHz)
For RFID Air interface communication
(High frequency at 2.45 GHz)
For RFID Air interface communication
(5.8 GHz)
For RFID Air interface communication
(860-930 MHz)
For RFID Air interface communication
(433.92 MHz)
For animal identification
Define the air interface
Application requirements
Freight containers
Returnable transport items
Transport unit
Product packaging
Product tagging (DOD)
RFID freight container identification
Define the structure of data on tag
For testing the conformance of RFID tags
and readers
For testing the performance of RFID tag
and readers

ternational Organization for Standardization (ISO) has established standards in many areas
such as in tracking cattle or goods in open supply chains. Also, the EPC Global Center
has introduced the EPC (Electronic Product Code) standard to identify and track consumer
goods through the international supply chain by creating their own air interface protocol.

25

Electronic Product Code Information Services (EPCIS) is a standard for sharing EPCrelated information between trading partners. EPC is used to provide a unique identity for
any item; it is a universal identifier. [17] Both Wal-Mart and DOD use the EPC standard.
However, DOD uses the ISO standards for air interface. The EPC Auto-ID Center has been
trying to develop protocols that can be used with different classes of tags. EPC tags are
divided into five classes. Class 0 is read only and has 64 bit. Class 1 is many readable and
one time writeable, and has 96 bit as a minimum. Class 2 generation 2 is read/ writes and
has 96 bit as a minimum. Class 3 is read/ writes and has battery power. The last one is class
4 which is for read/ write active transmitter. The first three tags are passive and the fourth
one is semi-active. For communication, class 0 and class 1 use different air interface. There
are four parts of the EPC number (header which identifies the version number, EPC Manager
or the manufacturer, object class, which identifies the product’s type, and the serial number
uniquely defines the product). [32] A second generation was developed because a global
standard was needed to be more closely aligned with ISO standards. Table 1.6 summarizes
the all ISO standards. [29, 30]

1.1.6

Architecture Framework

EPCglobal architecture framework is a group of standards for hardware, software, and data
interfaces. Many goals are considered in the EPCglobal architecture framework to provide
benefits to end users. Some of these benefits are: facilitating the exchange of data and goods
between partners, developing standards for global use, the interfaces between the components
of the architecture are identified in open standards, providing scalability and extensibility,
all parties in end users can capture data, keep it, and share them with whom they choose to
achieve their goals, and providing security and privacy to protect the information. [17, 32]
26

In general, there are three parts in the architecture framework: EPC physical object
exchange standards, EPC infrastructure standards for data capture, and EPC data exchange
standards. In the first part, all physical objects should be identified with Electronic Product
Code (EPC) in end users who are the parties in the supply chain for shipping, receiving
and so on. The standards are designed to ensure that all parties can access the information
properly. The following standards represent all within this part of EPCglobal Architecture
Framework: [17, 32]
• UHF Class 0 Gen 1 Tag Air Interface
• UHF Class 0 Gen 1 Tag Air Interface
• HF Class 1 Gen 1 Tag Air Interface
• UHF Class 1 Gen 2 Tag Air Interface v1.1.0
• UHF Class 1 Gen 2 Tag Air Interface v1.2.0
• HF Class 1 Tag Air Interface
• EPC Tag Data Standard [17, 32]
In the second part of architecture framework “EPC infrastructure standards for data
capture”, end users can build their internal system based on defining interface standards for
gathering and recording EPC data. The following standards represents all within this part
of EPCglobal Architecture Framework: [17, 32]
• EPC Tag Data Standard
• Low Level Reader Protocol
27

• Reader Management
• Discovery, Configuration and Initialization (DCI) for Reader Operation
• Tag Data Translation
• Application Level Events (ALE)
• EPCIS Capture Interface
In the last part of architecture framework “EPC data exchange standards”, data about
EPCs are shared and exchanged within the parties of end users. The standards are defined to
increase the visibility of the movement of objects in any place, and to facilitate the exchanges
by accessing the other EPC Network services. The following standards represent all within
this part of EPCglobal Architecture Framework: [17, 32]
• Core Business Vocabulary
• EPCIS Query Interface
• Pedigree Standard
• EPCglobal Certificate Profile
• ONS
• Discovery Services [17, 32]

28

EPC Network Service
Discovery
Services
(In Development)

ONS
Root

In Development

ONS

Manager
Number
Assignment

Tag Data
Translation
Definition
File

(offline service)

End User

ONS Interface

EPCIS Accessing
Application

Partner End User
EPCIS Accessing
Application

Local
ONS

“Pull” or “Push” mode

EPCIS Query Interface
Optional
Bypass For
Real-time
“push”

“Pull” or “Push” mode

EPCIS
Repository

EPC Core
Business
Vocabulary

Legend
EPCIS Capture Interface

EPCIS Capturing
Application

= HW/SW Role

From non-RFID
data carries

= Interface or Data
Specification (EPCglobal Standard)

Filtering & Collection (ALE) Interface

ALE Tag Memory,
Logical Reader, and
Access Control APIs

Filtering and Collection

Reader Interface
Reader Management
Interface

Reader
Management

RFID Reader
From nonGen2 Tag
EPC Tag Data Specification
Tag Air Interface (UHF Class 1 Gen 2, et al)

RFID Tag

Figure 1.5: The EPCglobal Architecture Framework (Roles and Interfaces) [17, 32]

29

In Figure 1.5 (the EPCglobal Architecture Framework), there are three parts connected
with each other: End User, Partner End User, and EPC Network Service. EPC codes on
RFID tags are observed by the End user and then the data will be shared with the other
End User (Partner End User). So, there will be interactions between End Users.
A EPCglobal architecture framework has interfaces that are administered by EPCglobal
standards (the plain green bars), and roles that are played by hardware and software components (the shadowed boxes). The following are the definitions for the items in Figure
1.5.
RFID Tag: There are four classes of a tag. These are Class1 (identity tag), Class2
(higher-functionality tag), Class3 (battery-assisted passive tag), and Class4 (active tag).
[17]
EPC Tag Data Specification: Defines the structure and the coding scheme of the
EPC; the meaning of data. Also, it defines the mechanism to federate the different coding
schemes, and binary representations. [31]
Reader Management Interface: Provides means to control and demand the configuration and observer the operational status of an RFID Reader. Also, if there is operational
problem, this interface will notify management stations about it.
Tag Air Interface: Defines the communication of data between RFID tags and readers.
In addition, it provides means to eliminate the interference between tags and readers.
RFID Reader: Gets the information from tags and sends this information to a host
application via the Reader Interface. Also, it has some other features such as filtering of
EPCs, aggregation reads, killing or locking and so on.
Reader Interface: Provides means to control the operation of RFID Reader. To access
the all capabilities of the UHF Class 1 Gen 2 Tag Air Interface, the EPCglobal Low Level
30

Reader Protocol (LLRP) standard was designed. This will include reading, writing, executing
other commands such as kill and lock commands, and sending reports if there are errors and
handling these errors.
Reader Management Interface: Provides means to control and demand the configuration and observer the operational status of an RFID Reader. Also, if there is operational
problem, this interface will notify management stations about it.
Reader Management: Controls the operational status and manages the configuration
of an RFID Readers. If there are issues, Reader Management will alert management stations
about them by providing mechanisms.
Filtering and Collection: If there is one or more than RFID Readers in the local
area that can cause the possibility of radio frequency interference, this role manages the
operations of them. Also, there are some other features such as decoding and encoding raw
tag data that are read from tags.
Filtering and Collection (ALE) Interface: Provides standards to filter and collect
data from Filtering and Collection role to the EPCIS Capturing Application role so that
there are no duplicates in EPCs in the list. There are other means that can be provided
in this interface such as to manage applications to secure client access to the ALE interface
and to share data from readers for multiple client applications.
EPCIS Capturing Application: This role manages many sources of data such as
filtered, collected EPC data that is received from the Filtering and Collection Interface. It
controls some actions like writing RFID tags and controls other devices.
EPCIS Capture Interface: EPCIS Capturing Application generates events. This
interface will provide a path for communicating these events to other roles that need them.
EPCIS Repository: When EPCIS Capturing Application generates events, these will
31

be recorded by this role and be made available when the EPCIS Accessing Application
queries them later.
EPCIS Query Interface: When the EPCIS Accessing Application requires data from
the EPCIS Repository, this interface provides means for that. Also, for mutual authentication between two parties, this interface allows this happen.
EPCIS Accessing Application: It executes some processes such as warehouse management, shipping and receiving etc. Core Business Vocabulary or Data Specification: For
denoting business steps, dispositions, and business transaction types, this section provides
standardized identifiers for use in EPCIS data.
Object Name Service (ONS) Interface: ONS is a research service to produce the
address of an EPCIS service by using an EPC. This interface makes means available for
locating the EPCIS service of the End User.
Local ONS: Provides the pointer to the EPCIS service.
ONS Root (EPC Network Service): This role identifies the local ONS service of the
EPC Manager organization for that EPC. If there is no local ONS, ONS Root will accomplish
the requests.
Manager Number Assignment (EPC Network Service): It issues unique EPC
Manager Numbers to each organization. Tag Data Translation: It provides machine-readable
files. This file is for translation between EPC encodings. This file is used by End Users to
be aware of a new EPC format. [17, 32]
Although some researches proposed different methods to overcome these threats, still
there are many vulnerable points in the system can be attached. For example, Divayan
Konidala [32] analyzed the security threats in the components of the EPCglobal Architecture
Framework and suggested some secured solutions. In this research, more details are going
32

to be concerned for providing security by proposing new protocols in the following chapters.

1.2

The contribution and thesis organization

The main contribution of this thesis is four chapters. Because the RFID system is one of the
most widely used technologies due to its advantages, it is opened to various attacks that try to
degrade the performance of the system. Therefore, a general introduction on security issues
is going to be introduced in Chapter 2. Also, different types of attacks against RFID system
and some proposed solutions would be reviewed in the same chapter. In addition, some
suggested solutions and protocols are going to be introduced in this chapter too. Although
some researches proposed different methods to overcome some threats, still there are many
vulnerable points in the system can be attached. In Chapter 3, an application is addressed
and new developed protocols (three protocols) are presented with different level of security
to prevent some attacks and to meet both the privacy and authentication concerns. In
addition, a new developed system model is going to be introduced in this chapter to solve key
management problems. After explaining the details of the approaches to meet the security
goals, discussion and analyzing the security and the time cost of the proposed protocols are
discussed in Chapter 4.

33

Chapter 2
Security and Privacy
In general, the term “security” refers to the “secure” condition of a system or services. When
a system has an ability to behave normally, especially when there are some efforts to make
this system misbehave, this situation expresses the security of the system. As a result, the
system will not suffer harm from some types of threats. [16]

2.1

Security properties

There are five goals or objectives with which any deployed system should be concerned.
• Confidentiality: data and resources should be accessed and shared by authorized
entities. All systems should protect their information from being leaked. This can be
done by protecting the communication channels within the system from unauthorized
entities.
• Integrity: protecting data or systems from being modified, deleted or duplicated. In
other words, protecting the accuracy of information and processing methods. Here,
the data in an RFID systems should be protected from unauthorized modification
(unauthorized readers and tags).
• Availability: keeping services or information available to authorized entities when
required. For RFID systems, all the system components and data are available at any
34

time to authorized entities. The system should not be corrupted in the presence of
multiple RFID readers and tags.
• Accountability: end users should access information in an appropriate way. Without
this principle, it is impossible to ascertain who is responsible and what has happened
within the system. This can be provided by an audit trail that provides more details
about the events and the actions that happened within the system.
• Non-repudiation: proving that a particular set of data is signed by the holder; the
holder cannot deny data modification and transmission. This leads to trust of one
system partners and also trust of the in integrity of data. In RFID systems, this goal
can be reached by providing mutual authentication between RFID readers and tags
and between readers and back-end databases. [16, 17, 57]

35

Non-repudiation

Availability

Accountability

Integrity

Confidentiality

Assurance

Figure 2.1: Security Objectives Dependencies adapted from[57, 58]

These five objectives are dependent on one another. It is impossible to achieve one
objective without considering of the others. For example, without integrity of the system,
confidentiality will not be valid. Further, all these objectives are related to another target
objective which is “Assurance”. It is the basis for confidence to protect the system and the
information it processes; it is established as a target when designing a system. By defining
all the functionality requirements of the above objectives, Assurance can be achieved and
undesired actions are not going to occur. Figure 2.1 illustrates the idea of these dependent
objectives. [58]

36

2.2

Security in RFID systems

The use of RFID is becoming more popular nowadays. So, the security of an RFID system
has to be a concern. In the next sections of this chapter, the RFID security threats and
some of the proposed solutions will be reviewed.

2.2.1

Issues

Opening up a system to the world or locking it down are not solutions for keeping the
system safe. Security techniques are available to prevent unauthorized access of system
resources. However, some issues should be taken into account first and then solutions should
be proposed to eliminate some issues.
With respect to RFID security and privacy, a number of threats, such as sniffing, tracking, spoofing, replay attacks and denial of service, are classified as a high level misuse of
properly formatted RFID data. The following is a brief description of some these threats
with examples:
• Malware attack: As defined, the standard architecture of an RFID tag consists of
an antenna, a microcontroller, and a storage unit. So, the RFID tag can be vulnerable
to all kinds of viruses and worms that may threaten a regular computer. [61]
RFID exploits is an example of a malicious RFID tag data corruption system. It
is the same as those on the Internet, for example Structure Query Language (SQL)
injection. SQL is a type of insertion attack executing SQL codes in the database to
retrieve authorized data, make modifications and deletions, etc. [60, 63] Back-end
RFID middleware can be exploited directly by RFID tags. One of the reasons that
facilitate this kind of attack is codes should be reduced to a small number of bits to
37

be appropriate for tags. One example of this injection is including the following SQL
command with the RFID read tag ID:
; drop table < tablename >

HTTP

SQL
Server
Database
Gateway

Management Interface

HTTP

RFID Reader
Interface

Figure 2.2: RFID Malware test platform [60]

The RFID information is usually sent to the back-end database for checking, and this
is where the problem occurs. It will delete one of the database tables. This leads to
corruption of the whole system. Another example is when an attacker might inject a
malicious code into an application using any script language to get some information.
RFID tags, which are used for tracking baggage in the airport, have a plethora of

38

information such as airport destination in its data field. When the tag is read, a query
is sent to the back-end database system to get the information about the destination
and destroy the system. Malware can infect any station within the system that is
shown in Figure 2.2. [60]
Viruses and worms are another kind of threat with respect to RFID tags. On the
Internet, viruses and worms can propagate into personal computers when relying on
a network connection that is compromised. The same thing can happen when relying
on an RFID tag and one therefore neglects the scanning mechanism of RFID.
RFID worms, which are considered a kind of malware, are downloaded from remote
sites. Then upon the execution, they try to change the functionality of the RFID
systems. In the case of RFID viruses, an Internet connection may not be needed as
they have the capability of self-replicating. The virus can basically be located on an
RFID tag, and then it propagates upon reading of the infected tag to corrupt the
back-end database or any other part of the RFID system.
To make it clear about the problems faced when a hacked RFID is used, the following
real-life scenario is cited. An attacker went to a supermarket and purchased some
items (supermarket uses RFID tags rather than barcodes). Then, outside of the store,
the attacker was able to remove the original RFID tag from items and attach fake
RFID tags that carried a malicious code. After a while, the attacker went back to the
supermarket to return those items with (now) fake RFID tags. This, of course, involved
scanning of the infected tags by the supermarket RFID readers and the embedded
malicious virus can then cause corruption of the back-end database of the supermarket.
[59, 60]

39

Some designers have proposed techniques for making a new kind of tags that are ”read
only” tags with the hope that no one will be able to modify them. However, this
technique still has a limited capability to solve the problem because an attacker can
fake blank tags to carry his/her malicious code. Another proposed solution to resist
this kind of attack is to limit the number of bits on an RFID tag. This is still not
a sufficient solution because there are many SQL commands that require only a few
numbers of bits such as a “shutdown” instruction that contains nine characters, or 63
bits. [59, 60]If this code is executed, the system will crash.
• Sniffing or eavesdropping: refers to intentionally listening to a private conversation
between a reader and tag such as an ID. There are two types of sniffing attack: passive
(any signal is not emitted by an attacker and not modify the message stream), and
active sniffing (an attacker modifies, deletes, injects, or replaces and replays messages).
[1, 6, 83] These will compromise the integrity and the confidentiality of the RFID
system. Personal identities can be revealed to anybody who is equipped with a tool
to get the information. Competition can find what you are shipping/ receiving. The
possible distances at which an attacker can listen to the messages exchanged between
a tag and a reader are categorized. [62, 63] See Figure 1.4 in Chapter 1.
For example, when a reader sends signals to a passive tag for data exchanging, it
powers the tag. In this case, the connection between them will be eavesdropped by
another reader that in turn, can monitor the resulting emissions. [2, 5] This attack
can be eliminated by limiting the distance between tags and readers. This can also be
done by using a metal screen to shield the communication between tags and readers,
and by encrypting the data. [6]

40

Figure 2.3: Tracking System

• Tracking and Tracing: when an attacker sniffs an ID of a tag, this ID can be used to
track the holder of that tag and determine the location of the item. Also, this attack
can be achieved when the response of the tag stays the same (the same identifier). See
Figure 2.3. This attack compromises the privacy of people or assets. Imagine your
glasses have an RFID tag that stores a 96 bits identifier. With this, an attacker can
follow your movement and know the places that you visit. Also, the attacker can place
several readers in your favorite store to learn your favorite items. Several techniques
are used to combat this type of attack. Mutual authentication between readers and
tags, shielding tags, killing tags after using them, and scrambling the ID of the tag are
some of the ways to combat this attack. [6, 63, 65]
• Spoofing: attackers can imitate real RFID tags after collecting data from an eavesdropping attack. An attacker can also retag items; for example, a tag that corresponds
to lower price can be replaced with one having a higher price before returning the
41

item on which the tag is attached. This simple switch can fool an RFID system and
compromise the integrity of the system. Some researchers have proposed a solution to
this problem. [1, 66]

Figure 2.4: Replay Attack [63]

• Replay attacks: an adversary can impersonate a real reader or tag to capture some
packets exchanged between two hosts, and then retransmits them at a later time.
This kind of attack compromises the integrity and confidentiality of the system, for
example, with garage door openers. In Figure 2.4, there are two attackers in the system
(represented by “Ghost” and “Leech”). The ghost places the tool near from the door’s
reader and the leech places the tool near a person who has the tag. By getting the
42

information from the tag and making the communication between the ghost and the
leech very fast, the ghost can open the door. Cryptography techniques, which are the
ability to exchange information between entities in a secret way using keys to prevent
an unauthorized entity from reading it [83], are not appropriate solutions to overcome
this type of attack. However, some authentication techniques are required to combat
this attack by using a password. [63, 64]

Jamming signals

Figure 2.5: Jamming Radio Frequency

• Denial of service (DOS): it denies service to valid users. This attack is easy to
accomplish with a jamming frequency. See Figure 2.5. In one technique, a shoplifter
carries a blocker tag, which is a passive RFID device that uses a sophisticated algorithm
to send jamming signals to disrupt reader communication. When a reader tries to send
a query to a tag, it cannot read the tag. In second technique, an attacker can shield
the tag from being read with a Faraday Cage, which is a metal enclosure lined with
43

aluminum foil that prevent the reader from reading the tag. In third technique, an
attacker with a powerful reader jams the reader by creating a more powerful return
signal than the signal returned from the active tags. It is very easy to provoke RFID
tags to communicate by an attack. In this case, the battery of the active tag will be
discharged so, it can no longer communicate. Therefore, when authorized readers try to
communicate with the tag, the tag is not going to respond; it will be in sleep mode. In
fourth technique, attackers can remove tags from items and block RFID readers query
signals. In fifth technique, they can flood the system with more information than it can
handle or remove a tag from an item and put it on other item causing wrong operation
in the system. These attacks compromise the availability and usability of the RFID
system. [1, 66]
• Cloning: this is another type of spoofing. It can be tag cloning or reader cloning. In
tag cloning, a tag can be duplicated, which may be similar in size or much larger than
the original one, to abuse private data or to get a restricted area. In reader cloning,
an attacker can duplicate an authorized reader. Authentication techniques such as
fingerprint or authentication techniques are used to combat this type of attack. [1, 6]
• Buffer overflow: this type of attack occurres when the data exceeds the length of
the memory. Some programming languages are not memory safe for example, the C
and C++ languages. When a program is executed, the length of inputs is not checked.
Because tags have a limited storage capacity, the buffer overflow may lead the program
to execute an arbitrary code. [63]
• Tampering with data: this occurs when an attacker modifies, adds, deletes or reorders data. For example, an adversary can modify the tag in a passport to contain the
44

serial number associated with a criminal. Also, an adversary can modify a passport
tag to appear to be a citizen in good standing. Another example of modification is
when an adversary changes a high-priced item to a lower priced item.

Reader
Insecure Channel

Tag

Tag

Insecure Channel

Insecure Channel

Tag
Back-end data base

Figure 2.6: Man-in-the-middle Attack

• Man in the middle: attackers are able to place their devices in the middle of two
hosts that are communicating and exchanging some information (i.e. a legitimate RFID
reader and tag). See Figure 2.6. An attacker records and replays the information that
could be changed (inserted, deleted, or modified). The main goal of this attack is
to make an RFID reader think that a certain tag is in place when it is not. Some
approaches have been proposed to deal with this attack. For example, a reader could
45

add a disturbing signal to the tone making it very difficult for the attacker to recognize
the signal. [1, 6]
For adding, additional tags can be added in a shipment to make it appears to have
more items than it does. For deleting, an attacker can remove or physically destroy
tags attached to objects to avoid tracking. For reordering, an attacker exchanges a
high-priced item’s tag with a lower-priced item’s tag. [63]
An RFID reader can be also compromised. There are two types of revocation: explicit
and implicit. Explicit revocation is when an RFID reader is lost, stolen, or compromised. An attacker who has the compromised reader can use it to identify and track
tags. Implicit revocation is when a reader certificate expires naturally. Usually, the
certificate expiration of readers is checked. If the reader is compromised before the
certificate is expired, the reader should be revoked from the system. So, explicit revocation is revocation before expiration and implicit revocation is certificate expiration.
[67]

2.2.2

Some proposed solutions:

Many researchers have proposed providing some security, authentication and privacy between
tags and readers because RFID tags are easily target for malicious attacks. This section
details solutions to prevent the above attacks on RFID system.
EPCglobal [17] defined some protection methods such as Application Level Event (ALE),
Reader Protocol (RP), Low Level Reader Protocol (LLRP), Reader Management (RM), and
EPC Information Services (EPCIS). [17]
To prevent a sniffing attack, tags that are not used anymore should be put in a shielded

46

enclosure. However, this solution is not effective if the tags should always be available for
legal queries. Another proposed method to enforce privacy and security on tags is the use of
“kill” or “sleep” commands. 32-bit kill passwords should be supplied by an interrogator to
kill a tag. However, these commands are unsatisfactory because consumers may want RFID
tags to remain responsive. [17, 31, 82]
One of the drawbacks of using cryptography to prevent threats is the extra cost; it is more
expensive and it needs more computational capabilities and it uses more power consumption,
especially for passive tags because they don’t have an internal energy source. Some designers
have proposed techniques for making a new kind of tags that are ”read only” tags with the
hope that no one can modify them. However, this is still a limited solution because an
attacker can fake blank tags to carry his/her malicious code. Another proposed solution is
to limit the number of bits on an RFID tag. This is still not a sufficient solution because
there are many SQL commands that require only a few numbers of bits such as a “shutdown”
instruction which contains nine characters or 63 bits. [59, 60]

47

“MY9A1”

“74B7A”
Tag 1

Tag 1

RFID Reader

RFID Reader

Figure 2.7: Minimalist Cryptography Technique

Juels designed a minimalist cryptography approach using a small portion of rewritable
memory and very limited resources. See Figure 2.7. He proposed that each tag have a small
number of pseudonyms. Then, for each query from a reader, a tag will release one of its
pseudonyms and then rotate them. This makes it hard for an adversary to associate any
tag with a particular object. The reader would store all the pseudonyms read from the tag
for identification purposes. So, if an unauthorized reader does not have all the pseudonyms
of the tag, there will be no matching between different pseudonyms of the same tag. [5, 68]
Juels further proposed a method to prevent unauthorized readers from harvesting all tag
pseudonyms by slowing tag’s response when it is queried quickly. Minimalist cryptography

48

is not perfect for real world deployments, but it can improve a tag’s ability and provide
privacy and mutual authentication between the reader and low-cost tags. [69, 70] This design
still has vulnerability compared to other approaches such as maximalist cryptography and
Shoehorning security.
Minimalist cryptography approaches use limited resources so that some resources are left
unused. In addition, minimum cryptographic operations are used. However, maximalist
cryptography approach proves that symmetric cryptography is sufficient on an RFID tag,
so security is maximized using low power microcontrollers. RC5-32/12/16, which is a symmetric block cipher used to encrypt and decrypt data, is applied on Wireless Identification
and Sensing Platform (WISP) UHF RFID tag because it is simple, needs low memory and
provides high security and a balance between security and performance for sensor networks.
[71, 107] In addition, in this method used a T1 MSP430F1232 microcontroller-based RFID
tag because of its low cost, low power consumption and it is wirelessly powered. Also, it
is used for RF or battery power application. This approach calculates how much computation can be used. Smith presents in detail WISP units that consume low power. [72] This
technique demonstrates that strong encryption can be performed on a tag and maximum
security can be achieved.
Shoehorning is another technique for providing security. It focuses on providing an authentication protocol rather than privacy using memory as input/output for cryptographic
model in a tag. [73] This method is like what Juels proposed. [69] He enhanced the using of
PIN-controls for killing and read/write to provide an authentication in Class-1 Gen-2 EPC
tags. However, his method can be attacked by eavesdropping attacks.
Shoehorning techniques use the same idea, but in a different way to provide a stronger
protocol within the EPC standard. It uses a challenge-response protocol to prevent the
49

cloning of a password because when a reader sends a command, which requests the tag’s
password, this password can be attacked. A reader and a tag should share the same secret
key. In addition, instead of sending a challenge from a reader to a tag, the tag can choose
the time of day as a challenge if the tag has a real time clock. ISO Standard 7816-4 offers
many commands used for authentication and accessing stored data. So, shoehorning uses an
Internal Authentication command from ISO 7816-4. Also some means are used to reduce the
cost by compressing ISO 7816-4. The idea of this is, if a challenge CR is implicit, that means
there are some bytes that are not used in the command, so that they can be eliminated.
The data field is about 512-bits; 32-bits are used for the password, so it still has space for
stronger authentication. According to Bailey [73], readers can respond to a challenge for
getting stronger authentication.
Most approaches are used depending on logical layer authentications. Cryptography is
better than just using a password because an attacker can eavesdrop it. However, cryptography is not a suitable solution against attacks because of the problem of key management.
Moreover, public key cryptography is a good solution, but it does not eliminate the fact
that a tag is very cheap for applying some algorithms. Using a Challenge-Response protocol
approach can eliminate an eavesdropping attack. However, the biggest problem in RFID is
providing database authentication and privacy because some RFID tags can be cloned. So, a
good tool needs to be provided to protect the tag. A fingerprinting approach is used for that
reason because the duplication becomes very hard. However, it still has one drawback that
it depends on direct optical contact, which denied the benefits of X-ray vision that RFID
grants. [74]
Although many researchers have proposed protocols to provide security and privacy between RFID readers and tags, RFID tags are still exposed to cloning. One reason for this
50

is the limited computational capabilities of the RFID tag. For example, the contents of a
tag’s memory can be copied and the same data can be injected into a new (counterfeit)
tag. To prevent the cloning of tags, protection techniques can be used on a tag to prevent
their memory contents from being read. However, this involves extra tag cost. [76] The
researchers at University of Arkansas have developed a robust method to prevent cloning of
passive tags by using the fingerprint technique based on some physical attributes of a tag,
then applying some applications of multi-class learning to classify the tags. An individual
tag is unique because of RF and manufacturing. [75] An algorithm was used to measure the
minimum power response of each tag by a sending radio frequency emitted by the reader to
the tag starting with the lowest frequency and increasing it step-by-step until the tag will
send a response. At a result, each tag has a unique power response for activation. This
unique physical characteristic can be used in the fingerprint technique without increasing
the cost of tags. RFID fingerprint data is stored in back-end database indexed by the ID of
a tag or sometime stored in the tag itself. [77]
In general, anti-counterfeiting used to prevent cloning has two phases. See Figure 2.8.
The first is an enrollment phase. Some trusted authority performs this phase by deriving
several fingerprints and recording the response. Then, these data are signed using a secret
key. All these data are printed on the product. The next phase to prevent cloning is the
verification phase. The devices read all data and perform the fingerprint technique after
measuring the responses to check if data has been printed by a legitimate authority, or not.
[76]

51

Enrollment
phase
Enrollment
Phase

Verification
phase
Verification
phase

Query Device
for ID

Query Device
for ID

Measure
response signal

Store device ID
and fingerprint

Match
enrolled?
?

Measure
response signal

ADC

Create
fingerprint

Create
fingerprint

ADC

Extract portion
of signal

Transform
features

Transform
features

Extract portion
of signal

Measure
components

Device
Features

Device
Features

Measure
components

Figure 2.8: Anti-counterfeiting phases [79]

52

Every tag has physical-layer features which are considered as a unique identification so
that they can be taken as fingerprints of tags. It is very hard to make a tag with the
same fingerprint as the original one because the variations are very difficult to reproduce.
Extraction techniques of the fingerprint are not expensive and it can be performed using
a low-cost purpose-build reader. Moreover, there are some classification systems that are
used to associate unknown RFID tag fingerprint to previously defined classes. One of these
systems is neural network approaches, which is one of multi-class learning techniques. As
mentioned, the fingerprints of tags are stored in either back-end database or in the chip. [78]
Some features can be used as a fingerprint: minimum power response at multiple frequencies, timing, phase or transients. An individual characteristic becomes like a secret key.
[79] Some have used the transmission time of EPC, PC, and CRC to distinguish individual
UHF passive RFID tags. [80] Then from the captured signals, the information was extracted
using MATLAB script. The data is classified using a K Nearest Neighbor algorithm (KNN).
Also, fingerprint RFID tags are based on the minimum power response which is the power
required to activate tags. [77] The results prove that every tag has a unique minimum power
response, which is a feature of the tags fingerprint.
There are some methods for extracting the feature of a tag. Danev’research [78] was
the first study about extraction of RFID physical layer fingerprints. According to their
results, the authors proved that a tag can be accurately identified; detecting of cloned tag
is easy; and the extraction of the fingerprints is inexpensive and could be applied with a
low-cost, purpose-built reader. They based this on the extraction of the modulation shape
and spectral features of a tag’s signal. A hardware setup is used to capture the signals from
tags and then record them to oscilloscope. For feature extraction and matching, Principal
Component Analysis (PCA) is used. With respect to the “Timing Feature”, which is defined
53

in [78] as “the time interval within which the tag responds to a wake-up command (WUQ)
and the duration of that response”, it is also used to identify RFID tags.

Planning
Storage
Distribution

Use
Suspension

Termination

Purging

Figure 2.9: Key’s life-cycle [84]

2.3

Key management

Key management is the basis of cryptography and it is an authenticated service. It is very
important to secure set of operations of any cryptosystem and to sustain encrypted data
with the keys during the key life-cycle. There are eight stages in the key life-cycle which
are: planning, key generation, key storage, key distribution, key use, key suspension, key
termination, and key purging. See Figure 2.9. [84]
54

The first two stages for planning and generating key are when an entity involves in an
organization and becomes an authorized member. The use of the key management is to facilitate the operational availability of keying material for standard cryptographic purposes.
The storage stage depends on the type and protection requirements. There are three types
of storage, which are operational, backup, and archive storage. For example, operational
storage is used when keying material are stored for normal cryptographic operation during
the cryptoperiod of the key. The backup storage provides a source for key recovery. The key
archive storage is a repository having key management of historical interest. In key termination stage, before removing a key from any organization, a reason for the termination should
be determined. This includes key compromise, removal of an entity from an organization,
etc. [85]
The most important part in key management is Key Management Audit because it
records all the operations of the key management associated with keys. It tracks all the
actions of key management, audits the conditions of using keys, and audits the changes of
key life cycle state. [84] However, there are some difficulties to manage and track keys of
systems. These difficulties are related to distributing keys, verifying the shared key, key
storage and validity checking. So, the Key Management Service helps with the scaling
problem of managing an increasing number of keys. The security of any system depends of
the strength of the keys, the techniques, and the protocols that are used to protect the keys
from modification by unauthorized disclosure. [81]
Distributing keys through a network should take place on secure channels. Some researches have been proposed to manage the distribution of keys between parties. A secret
key is used to lock a part of the tag’s memory. [17] This keeps the tag safe from tampering
with the content of the memory, but does not keep the content safe from unauthorized read55

ers. Also, it can be used for killing the tags. However, it is very challenging to initialize the
keys and propagate them to point-of-sale devices. [82]
There are three classes of cryptographic algorithms. These are hash functions, symmetric
key algorithms and asymmetric key algorithms. Hash functions generate small hash values
as an output from a large input value. This ability of hash functions is called “Compression”.
Adversaries cannot get the original input from the output. This function does not require
keys. It has many benefits such as: generating a random number, deriving a key, compressing
messages and the like. [16, 81, 83]
Symmetric-key cryptography is suitable for use in a closed system for encryption and
decryption. It can provide security for RFID systems. It transforms any input data using a
secret key that should be applied on the output to get the original input. A secret key should
not be revealed to unauthorized entities. This kind of cryptography provides authentication,
integrity and confidentiality, and generates random numbers. [16, 81]
Asymmetric key algorithm or Public-key cryptography is more suitable for open systems,
since both RFID readers and tags can use their public keys to protect the security of RFID
systems. It uses two keys (key pair: a private and public key). Anyone can know the public
key, but the private one should be kept secure by the entity that owns that key pair. This
algorithm can compute a digital signature, generate random numbers, and so on. [16, 81]
Key management should be designed in an appropriate way to eliminate the effect of a
compromised a single key onto the other keys. All compromised keys should be changed and
all affected keys should be replaced. [81, 83]

56

2.4

Other solutions

In general, attackers are becoming smarter and computers are becoming more powerful. As
described in Chapter Two and with respect to the RFID security and privacy, a number
of threats, such as sniffing, tracking, spoofing, replay attacks and denial of service, are
classified as a high level misuse of properly formatted RFID data. In the supply chain, the
transportation of cargo can be monitored in the whole chain from manufacture to retailer.
Also, because RFID has many benefits as described in Chapter One, the cost of the supply
chain could be reduced and safety can be achieved. [1, 2] For improving the accuracy and
reducing the time of processing, RFID can be used for store level inventory down to the item
level. Also, it can be used to define if a product is in the wrong location.
The use of standardized assessment tools, technologies and process is essential for improving the supply chain security strategies. A track and trace system for all products throughout
the domestic and foreign supply chain are a significant step in ensuring transparency and
accountability of product manufacturing and distribution. This can be done by using RFID,
which provides a secure identification and can quickly track a product at every point in the
supply chain. [7]
Anti-theft security devices, such as RFID labels, also play a key role (as described in
Chapter One) in deterring counterfeiting activity as well as detecting and monitoring the
actions of would-be criminals along with the movement of goods into and out of a retail
establishment. Chapter One lists many benefits for this technology in establishing an electronic pedigree. The FDA, for food defense, is focusing on moving from a reactive approach
to a proactive approach to address intentional and unintentional food contamination. The
FDA Combating Counterfeit Drugs identified several key elements for combating counterfeit

57

crimes and securing the nation’s supply of drugs by providing technology and increasing
criminal penalties, whereas the FDA food side is focusing on prevention and detection of the
fraud. They both intersect in using technology, which are RFID and barcode, for tracking
products. Also, they both can identify what products are likely to be counterfeited and
entered into the legitimate supply chain, even if the prevention of that is impossible. [8, 9]
All above reasons lead to the necessity of proposing methods or protocols to provide security
in RFID system.
In the next section, a brief introduction of proposed protocols is going to be introduced.
These will explain how RFID can be trusted for some applications, like the DOD, which has
many sensitive products that are shipped.

2.4.1

Proposed protocols

Research has been proposed to provide some security, authentication and privacy between
tags and readers because RFID tags are easily targeted for malicious attacks and because a
tag has both limited computational capability and limited storage. Many crypto algorithms
are broken, for example, DES was broken in 1998 and SHA-1 was broken in 2005. [84] In
this chapter, an overview of the related work in the field of RFID is going to be introduced
for providing authentication and privacy that can be used to increase the efficiency within
the supply chain.
The first approach to provide privacy on tags is the use of “killing” or “sleeping” commands. 32-bit kill passwords should be supplied by an interrogator to kill a tag. In this case,
the tag cannot be read by malicious readers. However, these commands are unsatisfactory
because it will eliminate all benefits of this system and consumers may want RFID tags to
remain active. For example, some smart machine “microwave” needs these tags to get the
58

time-to-cook information. Another example is a refrigerator that informs the owner that
some normally kept items are absent. Sleeping commands are used instead of killing to keep
the benefit from the tag by sending the readers PIN to activate and wake the tag again.
However, managing this technique in practice is difficult. [5, 14, 17, 86]
Another approach that provides privacy is shielding a tagged item. This approach uses
a radio wave-blocking material or scrambling of any outgoing signals from RFID tags. For
example, any container made of metal can be used to block the signals. However, this
approach has limited use because some items have very large shapes that cannot fit in some
containers. [5] An active jamming approach is another technique for providing privacy. This
can be accomplished by a device that transmits a signal to jam the RFID readers. However,
this approach is not useful because it can cause interference with nearby legitimate readers
if it transmits too high of a power signal. [5, 86]
Juels proposed another approach to deal with the privacy. This approach is using a
“privacy bit” in a tag. After authentication using the PIN for a tag, a reader changes the bit
in the tag from 0 to 1 or 1 to 0. 0 means that the tag is inside the store and 1 means that
it is about to enter places with limited access. However, there will be another tag “blocker
tag” that should interact with the tag to scramble the bits of all tags within the range. This
approach provides privacy only with the presence of the blocker tag. [86]
There are other approaches that provide authentication and some provide authentication
and privacy together. They focus on providing mutual authentication between tags and
readers. However, they presume that the channels between readers and back-end database
are secured. In addition, these approaches are still vulnerable to attack. Weis provides the
authentication scheme between tags and readers using a hash function. [87] Each tag has a
key and its hashed function, which is known as metaID. When a reader sends a query to a
59

(1) Query

(2) metaID

Database
da

(3) metaID

Reader

Tag

da

da

(4) (key,ID)

(5) key
(6) ID

Figure 2.10: Hash-Locking: A reader unlocks a hash-locked tag [87]
tag for getting the tag’s information, a tag sends metaID, which in turn will be forwarded
to the database. In the database, metaID will be matched against the stored data and if
there is match, the corresponding key and the ID of the tag will be sent to the reader. See
Figure 2.10. The key is then sent to the tag that in turn, will apply the hash function on
the received key and compare it with the stored one. If it matches, the tag will unlock itself.
This scheme has three drawbacks: the metaID can be eavesdropped and copied into a fake
tag. This can cause the reader to authenticate the fake tag.
The second drawback is the tag can be tracked using metaID because it is always the
same. This can affect to the privacy of the holder. Third, the key and the ID are sent in a
clear message that an attacker can easily obtain.
To prevent tracking the tag in the previous scheme, which is hash-locking, the same
authors [87] improved the scheme by generating a random number (r) every time the tag
received a query from the reader. Then, instead of sending the metaID, (IDk , r) will be
sent. So, this message will be different in every query. Then, the reader computes the hash
60

function for all IDs that are sent from the database with the random number r. If there
is one match, there will be authentication between tags and readers. The problem in this
scheme is r can still be eavesdropped with (IDk , r) and used to make a fake tag to hold this
information. Also, this scheme is inefficient because sending all IDs from the database to
the reader will consume more time for calculations especially if the number of IDs is large;
the computational cost for the reader will be linear. See Figure 2.11.

(1) Query
(2) r,h(ID k ||r)

Database
da

(3)Get all IDs
( 4)

Reader

Tag

da

da

ID 1 ,ID 2 ,…..ID n

(5) ID k

Figure 2.11: Randomized Hash-Locking: A reader unlocks a tag whose ID is k in the randomized hash-lock scheme [87]

A Privacy and Authentication Protocol for Passive RFID Tags (PAP) is another protocol
that deals with a privacy and authentication for passive tags. [14] This protocol provides
computation that can be applied in passive tags. The authors present four protocols used
in four locations: inside a store, at a checkout counter, at a return counter, and outside
the store. They used the random number and privacy bit concepts. Also, they assumed
that there will be no malicious readers inside a store and an attacker cannot connect to the
database to get the information of the tag.
61

The in-store and out-store protocols are very simple. The other protocols (checkout and
return) have more steps to be applied between readers and tags. Each tag has a secret key
(k), an ID, a privacy bit (0 means inside a store, and 1 means outside a store), and a generic
name. In the checkout protocol, see Figure 2.12, the reader gets the shared secret key of
the tag from the database by using tag’s ID. Then, the reader and the tag will authenticate
each other by computing the hash function between the random number and the key. If it
matches with what they already have, the mutual authentication will be satisfied and the
privacy bit will be changed from 0 to 1.

(1) Query
(2) ID,n t

(3) ID

Database

Reader

(4) key

Tag
(6) H 1 ,n r

da

da

da

(8) H 2
(5) H 1 =hash( n t , k)

( 9) veri f y H 2

( 7) veri f y H 1
If successf ul
H 2 =hash( n r , k)

Figure 2.12: The checkout protocol in PAP [14]

The idea of the return protocol is the same. But, the privacy bit will be changed from 1
to 0 after satisfying the mutual authentication between tags and readers.
Still, these protocols are vulnerable to attack. Kim [88] proposed Improved Privacy
and Authentication Protocol for Passive RFID Tags (IPAP) method for improving the PAP
protocols because he found that each protocol in the PAP scheme is vulnerable. For example,
for the in-store protocol, there is a possibility for an attacker to sniff the information. An

62

(1)Q u ery

(1)Query

(2)n ame,n t

Reader

Tag
(2)ID,n t

da

DB

(3)n ame

Reader

da

Tag
da

(4)p rod u ct
typ e

PAP in-store

da

IPAP in-store

da

da

Figure 2.13: Comparison between PAP and IPAP in-store protocol [88]
attacker can get the information of the tag without connecting with the back-end database
because a tag sends its ID when it receives a query. In this case, the attacker keeps the ID
and compares it with the scanned one during the checking out. If they match, the attacker
can obtain the information that is related to the purchased product of the consumer. So, the
IPAP scheme for the in-store protocol does not transmit any information about the product
in order to keep the privacy of the product. The name, which is not a unique identifier, will
be sent instead of the ID. See Figure 2.13. However, the name, which represents the numeric
representation of the product type, can be eavesdropped.
In addition, the return protocol in PAP has one significant issue, which is the tag sends
its name to the reader that, in turn, will communicate with the back-end database to get
the secret key of that tag. The problem is the name (or product type), which is the same
for all identical products, and it is not the serial number. In this case, the reader performs
a brute-force search to get the secret key of the tag.
IPAP provides a different concept. In this protocol, the first two steps are the same as

63

(1) Query
(2) name,n t
( 3 ) H 1 = h a s h ( n t , M k)

Tag

Reader

Database
(7) ID

da

(4) H 1 ,n r

(6) ID,H 2

da

da

(8)k t

(9) verify H2

(5) verify H1
I f s u c c e s sf u l
Unlock & send ID
H 2 = h a s h ( n r , M k)

Figure 2.14: The checkout protocol in IPAP [88]
the IPAP in-store protocol. The difference is in the next steps. See Figure 2.14. A reader
has the secret shared key of that tag and it computes the hash function and sends it to the
tag that, in turn, verifies the results. After the verification step, the tag now sends it’s ID.
However, this is also vulnerable because an attacker can eavesdrop product IDs. All the
previous methods to provide privacy and authentication have weak points and the channel
between back-end database and readers are assumed to be secured.
A man-in-the-middle attack can be applied on this protocol to obtain the ID between
tags and readers, and between readers and the backend database.
There are other solutions for providing security to an RFID system, but they are not
suitable for passive tags because of the tags limited computational capability. To avoid that,
one proposed solution uses external device to be carried during the day to provide security,
but this is impractical. [89]
Another approach for eliminating the drawbacks of PAP protocol is RFIDGuard. [90] A
kind of pseudonym challenge-response protocol is proposed in this approach. This research

64

(1)Query
( 2 ) p o i n t e r = p o i n t er ’

(3) ID, name, pointer ’
Tag

Reader

(4)

( 4 ) p o i n t e r = p o i n t er ’ + 1

da

Retrieve
p s e u d o n y m[ p o i n t e r ’ + 1 ]
m= r a n d o m n u mb e r

da

(5) m, pseudonym[pointer ’+1]
( 6 ) V e r i f y p s e u d o n y m[ p o i n t e r ’ ]
If succeeds,
Privacy bit =1

(7) pseudonym[m]

(8) Verify pseudonym[m]

Figure 2.15: The checkout protocol in RFIDGuard [90]
uses the same idea of PAP, which has four protocols, but with procedures that provide more
security while eliminating some drawbacks of PAP. Almost the same assumptions are used
in this approach. One of these assumptions is an attacker cannot hide himself easily between
the authorized readers and tags. Also, a privacy bit, a same list of pseudonyms for both a
tag and reader, and a pointer for this list are used in this approach. When a tag receives a
query from a reader, the tag sends the pointer of the list and ID of the tag. [90] See Figure
2.15.
Although this approach requires the passive tag to do a very small amount of computation
and deals with privacy and authentication at the same time, still it has drawbacks. First, a
malicious reader may exist within the range of interaction. Second, how many pseudonyms
can a passive tag carry? Also, there are no details about the list of pseudonyms for example:
if the pointer is at the end of the list, what does the tag do? Third, replay and man-in-themiddle attacks can be applied in the checkout and return protocols. An attacker can get the

65

information from the tag and forward it to the reader, and vice versa, until the attacker gets
the ID of the tag in the last step of these protocols. In spite of sending the pseudonyms after
XORing (XOR is a logic operation) with the random number that is established and shared
between readers and tags, the IDs are not going to be XORed with the random number.
They are sent in clear text. Fourth, in the checkout protocol, the tag sends its ID to the
reader. In this case, an attacker eavesdrops and keeps the ID and compares it with the
scanned one during the checking out. If they match, the attacker can obtain the information
that is related to the purchased product of the consumer. This problem is the same as the
problem in PAP. Fifth, the tag can be cloned by getting the list of pseudonyms. Although
there will be difficulties in aligning the pseudonyms with their pointer, this approach can be
vulnerable.
Hakeem proposed a novel key management protocol, which is Hacker Proof Authentication Protocol (HPAP), for providing mutual authentication between readers and tags and
securing tags ‘information. [95] He used time stamp, random number, hash function, encryption key, and Linear Feedback Shift Register (LFSR), which is a shift register its input
bits is linear function of its previous state with using XOR. He simulated his protocol using
C#.NET. However, this protocol has drawbacks, which are changing a key of a tag every
time a reader sends a query to the tag, needs more computation in the tag and in the server,
and require a number of time consuming arithmetic operations in LFSR. In addition, the
key of the tags and the timestamps are updated using LFSR. Because LFSR has a finite
number of possible states, it must eventually enter a repeating cycle.
Most of the proposed approaches focus on providing privacy and authentication between
readers and tags. Also, they focus on how to avoid the counterfeiting of a tag. A major
question is what to be done about the readers whether they are safe and whether they can be
66

compromised. Most of the above approaches assume from the beginning that the channels
between readers and the back-end database are safe. RFID readers can be lost, stolen,
compromised by an attacker, or decommissioned. In all of these cases, a reader should be
revoked correctly and effectively from the system before an attacker gets the compromised
reader so that he can identify and track tags.
As described in Chapter Two, there are two types of revocation: explicit and implicit
revocation. The explicit revocation is when RFID reader is lost, stolen, or compromised.
An attacker who has the compromised reader can use it to identify and track tags. The
implicit revocation is when a reader certificate expires naturally. So, the explicit revocation
is revocation before expiration and the implicit revocation is certificate expiration. [43]
Very few solutions are proposed to find a compromised reader, reader revocation, and
expiration checking. [91] Some solutions for reader revocation techniques are: date register
and time stamp, internal clock, on-line revocation, and so on. [93] The most important
challenge is the way for revocation and expiration checking of the RFID reader certificate to
be handled. Some approaches use Certificate Revocation Lists (CRLs), but these approaches
are weak in public key-enabled RFID systems. [92] One of the reasons that make this
approach weak is a passive tag has no clock. In this case, a tag cannot make sure that a
reader’s public key certificate (PKC) is expired or revoked. In addition, if a passive tag
has an internal clock, it also cannot be used because the clock needs uninterrupted power
to be sustained. Even for an active tag, it leads to problems such as: battery cost, clock
synchronization, and battery replacement. So, it is difficult to manage the system without
error. [93]
Public key cryptography infrastructure is used [94] to provide authentication between
readers and tags. It uses also a digital signature that tags have for validating a product.
67

However, this approach is not suitable for an RFID system because it is complicated and
needs high run-time complexity. Also, the disadvantages for using a public key cryptography infrastructure are the speed and it may vulnerable to impersonation after successfully
attacking on a certification authority. In this case, using public key cryptography infrastructure is not necessary and using secret key cryptography can be sufficient. This can be
done by monitoring and managing all keys in closed system by an appropriate authority.
Threats on products are increasingly becoming prevalent because of huge financial motives.
At any stage of a supply chain management from manufacturers to retailers, threats can be
introduced. Using an RFID system is a suitable solution for eliminating some of the threats
in different means. In next the chapter, a new approach is developed to provide privacy
and authentication in the RFID system by addressing an application that is based on an
electronic lock. This can be used to provide security for freight passing through customs
and in international airports. In addition, this approach focuses on handling a compromised
reader, not just providing security between readers and tags. Also, it handles key management and manages distributing the keys, and tries to eliminate some problems that can be
found in distributing the keys. This developed model aims to provide security (privacy and
mutual authentication between readers and tag) in the closed system with low computational
requirements. Also, this model aims to detect and revoke a compromised reader from the
closed system by an effective model.

68

Chapter 3
Methodology

3.1

The Application

Radio frequency identification (RFID) systems can be used in many applications such as
supply chain management, theft and counterfeiting prevention, Department of Defense measures, passport control, Customs, and so on. Every entry in the supply chain is vulnerable to
be attacked. It is necessary to develop solutions against possible attacks, which are discussed
in Chapter three, in international ports and customs control points for incoming containers
from land, sea or air.
Customs inspections at borders can lead to delays and additional logistics costs. The
processes of container transportation are vulnerable to security threats such as counterfeiting. In this case, an RFID system can provide physical and contents security by using an
electronic seal (eSeal). [104] It is used to guarantee the authenticity and the integrity of
freight containers. Also, it is used to not just provide physical security but also to store
information and different data for supply chain management. This information could be the
seal ID number, the container ID number, an alarm function to inform in real-time, sensors to indicate the environmental status of the container content, etc. When any container
moves from one port to another, there is no need to have a physical inspection where eSeals
are used. This leads to increased efficiency of the whole process including reducing logistics
costs, decreasing inventory stocks and increasing customer service levels. [101]
69

One type of eSeal, developed by Confidex, is used by Container Centralen (CC) on its
fleet of 3.5 million CC containers to improve the quality of the CC Container pool, their
capacity to detect counterfeit products, and to provide transparency in the supply chain.
[102, 103]
For international cargo transportation, the cargo is loaded into some type of container
and transported to a border customs control point like Hong Kong custom. At the border,
the cargo will be inspected and cleared to enter the country. After inspecting the cargo, a
decision will be made if the cargo is going to be rejected or accepted. [97] The above process
needs to be accomplished quickly and accurately.
The application that is addressed by this research is based on an electronic seal that can
be used, as mentioned above, to provide security for freight passing through customs. This
can help to ensure that the cargo is kept safe (tamper free) and need not be opened for
inspection at numerous border control points. In this scenario, either active or passive tags
can be used.
RFID technology can improve the security and visibility of the customs process and also
reduce the amount of time required for clearing each container through customs. In addition,
an agency can record the path of the containers and the place and time of each inspection.
This type of system is currently being used in the Hong Kong customs process and it can
be expanded to cover other border control points because of the benefits that can be gained
from the eSeal system. [97]
At each country border crossing, this application allows a number of readers to inspect
the cargo seals. This will also be the case for international airports. The tagged goods are
moved through different customs control points where an eSeal will be interrogated by the
customs agency. The eSeal consists of a physical lock activated by a built-in RFID tag,
70

which allows the lock to be opened or kept closed by means of receiving a signal from a
reader. The eSeal will be attached to the cargo container at the time when the container
is filled and sealed. Then, ID numbers of the eSeal will be read from the tags via readers.
The data are saved in the custom’s database. This data includes a unique ID, key and
other information. The system stores the eSeal ID number linked to the vehicle ID. Before
the container can be unlocked, the eSeal must be unlocked via an electronic key, which is
received from the back-end database. In addition, truck drivers cannot unlock the container
because they don’t have access to the electronic keys.
In 2011 the Hong Kong Customs and Excise Department (C & ED) used this technology
to increase the speed of all processes of checking on cargo in their customs. In 2012, C &
ED proved that using RFID system is an effective solution to provide security in freight
passing through customs. They estimate that the RFID system reduced the time required
to clear the containers in customs from 2-3 hours to 5 minutes because the containers have
not been opened between their inspection at the border control points and their arrival at
the airport. Also, the information about where the cargos have been is recorded during the
transportation, as is and the time for the inspection. The custom officials know that the
cargo containers are not opened during the transportation and the inspection at the border
control points. To facilitate this custom inspection process, RFID readers can be installed
at border control points and also international airports. A total of 38 readers are installed
in the Hong Kong C&ED. [97]
ESeal provide a high level of tamper resistance. However, the security of the eSeal itself
must be ensured. There are many possible attacks against the integrity and authenticity of
eSeals. For example, a malicious party can generate a fake alarm to deceive the interrogator,
sniff information about an eSeal by intercepting radio signals between eSeals and readers,
71

clone an eSeal and creating a device that cannot be discerned from a legitimate eSeal, and
disrupt the communication channel by jamming or shielding. Currently, RFID eSeals do not
provide any robust solution to these problems. [104, 105]

3.1.1

An attack scenario

There was a particularly extensive vulnerability assessment for eSeals in early 2005. Spoofing
and cloning attacks were identified as potential data integrity threats to eSeals. [104]
Imagine two port terminals (A and B), each of which has many interrogators to read tags
attached to cargo containers. See Figure 3.1.

Attacker B
sniffs the ID of
that eSeal and
its key

Attacker A sniffs
the ID of that eSeal
and its key that is
sent from the
server

Fixed reader

Server
eSeal on trailer

Figure 3.1: An Attack Scenario in Port A

72

When the cargo reached port A, the ID will be read from the tags via readers and then
the ID will be sent to the database by the reader to get the key of that tag to unlock the
tag. An eavesdropper with an antenna and some basic receiving equipment can gather the
same RFID tag ID that is compiled by the reader at the port because the ID is sent in a
clear message. After sniffing the ID, the key of that tag can also be sniffed because it is sent
in a clear message. In this case, the cargo now will be tracked to know its location and also
can be opened using the sniffed key. An adversary can steal large numbers of RFID data
and then place the real data onto counterfeit RFID tags. This way, both the real and fake
RFID tags contain legitimate information.
In Figure 3.1, the IDs and keys can be stolen. In addition, one of the readers installed
at a border control point can be compromised, or lost as a result of some attacks. In order
to solve the above problems and make the system secured by preventing counterfeiting, a
communication protocol between an eSeal and an interrogator should guarantee the following
security functions:
• Mutual authentication between an eSeal and readers
• Data confidentiality and data integrity
• Immunity to denial of service
• Replay protection
• Revoking a compromised reader
This can be done by developing a topology and three protocols for providing a suitable
network and secured methods. [104, 105]

73

In summary, RFID systems can be used in customs control points because it has advantages such as anti-smuggling, improving cargo movement security, improving efficiency
of customs clearance, getting real time information on the activities, and savings of cost
and time. [98] In the remainder of this chapter, a developed system model is going to
be introduced to manage the generation, exchange, storage, use, and replacement of keys.
Also, new proposed protocols will be presented to provide mutual authentication between
tags and readers and between readers and back-end database, to revoke a reader when it is
compromised, and to add a new reader in the system.

3.2

System modeling

In total, there are four levels in the system: enterprise, facilities, reader and supply level.
The enterprise level is the organization wide communication network that runs from the
server down to all the readers. At this level, all tags are created and each tag has a unique
ID and key. Also, it is the repository for all the information. The facilities level could be a
node in the distribution system or the border crossing. The readers are the devices by which
eSeal or tags can be read and they are directly connected back up to the facilities and the
supply level that flow tags.
The eSeal system architecture (The governing information control system) consists of a
main server (enterprise level) and a set of sub-servers (facilities level). Each sub-server is
connected to a number of RFID readers. Further, every reader may read thousands of tags
inside the network. The goal for the system is to provide mutual authentication between
servers-readers and readers-tags. Because all tags in this system have a unique key, the whole
keys should be stored in the main server and then any authenticated reader in the system

74

can get a key of a read tag from the main server. In addition, if there is a compromised
reader, it will be revoked in an efficient way from the whole system without affecting the
rest of the system by developing a model to manage this problem.
Before starting to explain the developed protocols, a system model should be designed to
solve what computer scientists refer to as “key management problems” and to manage the
communication function between servers and RFID readers by providing privacy, integrity,
and authentication. As seen in Figure 18, a tree key graph model can be used to implement
the RFID system in customs control points because a compromised reader may need to be
revoked from the system or a new reader may need to be added in the system. [96] New
keys will be created and shared to facilitate the process.

3.2.1

The Tree Key Graph

The system is designed based on the concept of a tree key graph because it is scalable to large
groups. [96] Their secure group system consists of a set of users (U), a set of keys (K), and R
which is the binary relation between U and K (user-key relation which means that the (user,
key) is in R). See Figure 3.2. Each user has an individual key which is shared only with the
main server; for example: u1 has the key k1 . Also, there is the main key, which is k1−9 , and
the group keys, such as k123 , k456 , and k789 . When the server sends a message to the first
group, it will be encrypted using k123 . In the border customs scenario, the main key would
be the government master key, the group key would be the individual custom station key
and the user keys would be the readers keys used at each customs stations. The advantage
of the group keys is to reduce the number of messages that the server sends because each
message sent adds time to the encryption and decryption functions and therefore slows the
system down.
75

K1-3

Group 1 key

K1-9

Main key

K4-6

Group 2 key

U2
(K2)

U3
(K3)

Group 3 key

User keys

User keys
U1
(K1)

K7-9

U4
(K4)

U5
(K5)

U6
(K6)

U7
(K7)

U8
(K8)

U9
(K9)

Figure 3.2: A tree key graph [96]
Every time a user joins or leaves, new keys are created and sent to the client. The
number of messages that the server sends depends on the number of users the system has.
For example, if u9 is revoked from the system, the server creates a new group key instead of
k789 and main key instead of k1−9 . Then, these keys will be encrypted before sending them.
In Figure 3.2, when u9 leaves the system, the server will send the following messages: a new
group key k78 to the u7 and u8 encrypted by using their individual keys (k7 and k8 ), and
a new main key k1−8 to the first group (1,2,3) and second group (4,5,6) encrypted by k123
and k456 respectively. Then, it will send another message which has the main key (k1−8 ) to
the third group (7,8) encrypted by a new k78 . The same above process will be applied when

76

a new user joins the system.
To construct and send the rekey messages within the system, three different approaches
are considered [70].
1. User Oriented Rekeying: The idea of user oriented rekeying is that for each user,
the server constructs a rekey message that contains the new keys needed by the users
and encrypts them using a key already held by the users.
For joining the system, if a user u10 is added to the system with the group (7, 8, 9),
the following messages are going to be sent by the main server:
• The server sends to the users (1-6) the message (k1−10 ) encrypted by k1−9
• The server sends to the users (7,8,9) the message (k1−10 , k78910 ) encrypted by
k789
• The server sends to the user (10) the message (k1−10 , k78910 ) encrypted by k10
For a user leaving the system, for example u9 , the following messages are going to be
sent by the server:
• The server sends to the users (1,2,3) the message (k1−8 ) encrypted by k123
• The server sends to the users (4,5,6) the message (k1−8 ) encrypted by k456
• The server sends to the user (7) the message (k1−8 , k78 ) encrypted by k7
• The server sends to the user (8) the message (k1−8 , k78 ) encrypted by k8
2. Key Oriented Rekeying: Each new key is encrypted individually. The server constructs multiple rekey messages as per the needs of each subgroup. The users of each
group receive a rekey message containing precisely the new keys that each group needs.
77

For joining the system, if a user u10 is added to the system with the group (7,8,9), the
following messages are going to be send by the server:
• The server sends to the users (1-9) the message (k1−10 ) encrypted by k1−9
• The server sends to the user (10) the message (k1−10 ) encrypted by k10
• The server sends to the users (7,8,9) the message (k78910 ) encrypted by k789
• The server sends to the user (10) the message (k78910 ) encrypted by k10
For a user leaving the system, for example u9 , the following messages are going to be
sent by the server:
• The server sends to the users (1,2,3) the message (k1−8 ) encrypted by k123
• The server sends to the users (4,5,6) the message (k1−8 ) encrypted by k456
• The server sends to the user (7) the messages (k1−8 ) encrypted by k78 , and (k78 )
encrypted by k7
• The server sends to the user (8) the messages (k1−8 ) encrypted by k78 , and (k78 )
encrypted by k8
3. Group Oriented Rekeying: Here the server constructs a single rekey message containing all new keys. This rekey message is then multicast to the entire group. This
technique has a number of advantages compared to the other techniques. One, there is
no need for a subgroup multicast. Second, with fewer rekey messages, the number of
messages are reduced. Also, the total number of bytes transmitted by the server per
join/leave request is less than those of the other approaches. [96]

78

For joining the system, if a user u10 is added to the system with the group (7,8,9), the
following messages are going to be sent by the server:
- The server sends to the users (1-9) the messages (k1−10 ) encrypted by k1−9 , and
(k78910 ) encrypted by k789 - The server sends to the user (10) the message (k1−10 ,k78910 )
encrypted by k10
For a user leaving the system, for example u9 , the server sends to the users (1-8) only
one single rekey message (k1−8 )k123 , (k1−8 )k456 , (k1−8 )k78 , (k78 )k7 , and(k78 )k8

Enterprise Level

Facilities Level

K1
R1

K1-9

K1-3

K2

K4-6

K3

R2

K4
R3

K7-9

K5

K4
R4

R5

R6

K7
R7

Readers and supply Level

Figure 3.3: Developed RFID System Model

79

K9

K8
R8

R9

3.2.2

The Developed System

The newly developed system in this work is designed based to the concept of the tree key
graphs described above. The system consists of a main server (enterprise level) and set
of subservers (facilities level). Each subserver is connected to a number of RFID readers.
In this application, every customs control point at a border has a subserver that controls
a number of readers. Every reader may read thousands of tags inside the network. It is
possible to add new readers to the server and remove existing or compromised readers from
the server. This is mainly done by using the concept of shared keys within the network.
Earlier, this situation was discussed. The border customs has the government master key
(the main key), and the individual custom station key (the group key). The main server
has the main key, which is shared between the subservers. For example, in Figure 3.3, the
main key is k1−9 . In addition, a shared key (k123 ) is the group key between the subserver
and readers R1 ,R2 and R3 . Also, each reader and tag has their own unique key that helps
to identify them. The technique that is used in this application is group-oriented rekeying
for three reasons. One, there is no need for a subgroup multicast. Second, with fewer rekey
messages, the numbers of messages are reduced. Also, the total number of bytes transmitted
by the server per join/leave reader is less than those of the other approaches.
When new readers are added to the system or when existing readers are removed from
the system, the security within the system should not be compromised. This is achieved by
effectively changing the keys at each and every point in the network. To securely distribute
the new keys within the network, the server sends the rekey messages to the readers. A
rekey message contains one (or more) encrypted new keys that the readers use to decrypt
with appropriate keys in order to get the new keys.

80

As mentioned earlier, there are four levels in the system: enterprise, facilities, reader and
supply level. The goal for the system is to provide mutual authentication between serversreaders and readers-tags. Because all tags in this system have a unique key, the whole keys
should be stored in the main server and then any authenticated reader in the system can
get a key of a read tag from the main server. There are three scenarios on which the system
can be applied. These three scenarios can be described and applied in the system.
1. A good reader reading a good tag
2. A good reader reading a bad tag
3. A bad reader reading either a good tag or a bad tag

Reader 1

Antenna

LLRP

Server

LLRP
Reader 2

Antenna

Reader n

Antenna

LLRP

Figure 3.4: LLRP endpoint [99]

As described in Chapter One, the Low Level Reader Protocol (LLRP) provides authenticated communication between the server and the reader. [99] See Figure 3.4. It provides the
format and procedures for communication between them. In addition, the LLRP provides
81

control of RFID air protocol operations timing and access to air protocol command parameters. There will be messages (used by LLRP) from the server to the readers for managing,
controlling and updating the operations at the readers. Also, there will be messages from
the readers to the server for sending the status of the readers as a report. After receiving
reports from the readers, the server configures the readers to enable or disable some other
events such as buffer overflow.

3.3

The Communication Between Readers and Servers

Therefore, there will be an interchange of message between readers and the server to check
the health of the readers in the system. For example, GET READER CONFIG is a command that is issued by the server to the reader to get the information related to the reader’s
configuration. This command has some interested parameters that the reader will return. If
there is an error, some codes will be sent to the server. Also, ReaderEventNotificationData
and ReportBufferOverflowErrorEvent are parameters sent by the reader to notify the server
about the connection establishment (Transmission Control Protocol and Internet Protocol
(TCP/IP connection)) or some critical events like fault-detection and buffer overflow. Another parameter sent by the reader is ReaderExceptionEvent, which is sent if an unexpected
event has occurred on the reader. In addition, if an antenna has an issue (connected or
disconnected), the reader will send a report (AntennaEvent Parameter) to the server telling
about this issue. [99]
Another aspect for the communication between the reader and the server is there will be
two channels between the server and the reader: an alarm channel and command channel.
The server issues requests and sends them by the command channel to the readers and also

82

receives a response from the readers. This is done to monitor the health of the readers. The
response will be either a normal or an error response. The alarm channel carries messages
issued by the readers to inform the server about the health of the reader. For example, if
the server asks the reader to get the serial number of the reader, and the reader cannot find
it, the error message “ERROR UNKNOWN” is sent by the reader. [100]
There are two types of errors: communication errors and command execution errors. The
communication errors can occur when the server tries to issue a command to the reader, but
the reader is not responding, or the reader sends a notification to the server. However, the
command execution errors occur as a response to commands sent from the server to the reader
in the event of errors. These errors could be parameter missing, parameter length extended,
parameter invalid format, and so on. If errors happen in a reader such as a buffer overflow,
the reader will send a notification to the server. [100]
In the above cases, when the enterprise server receives the error notification, it will
determine if the issue that happened to the reader is critical, or not. If it is critical, this
reader will be revoked from the entire system. If not, it will handle and maintain the errors.
In the next section, three proposed protocols that provide mutual authentication and
handle a compromised reader are going to be introduced briefly.

3.4
3.4.1

The Developed Protocols
First Protocol

In this protocol, each reader has an individual key and also a shared key with their servers.
Every RFID system consists of a database (the server), which holds the information about
all the authenticated tags and readers under that network. All RFID tags have a unique ID,
83

which is the numeric representation of the individual item, name (item type), and a secret
key (Kt ). The tag memory is logically separated into four distinct banks: reserved memory
(at least 32 bits for storing the EPC information such as password), EPC memory (at least
496 bits for storing EPC information), TID memory (at least 32 bits storing tag identifier),
and user memory (for storing information related to the application of a tag).
The main server (enterprise level) might have a number of subservers (facilities level)
based on the capacity and the size of the network. All RFID readers (authenticated ones)
are connected to the subservers via a group key (which is the common key between the
readers and the server in a particular network).

Figure 3.5: First Protocol

84

When a reader sends a query to a tag for getting that tag’s information, the tag sends
its ID, encrypted with a random number, generated from the tag, using its key. The reader
in turn communicates with the subserver in the network and sends the encrypted message
using the group key to get the key of the tag. The subserver contacts the main server by
using the shared key and validates the tag and finds out the corresponding matching key for
the tag from its database. This key is encrypted by using the group key and then sent to
the reader, that can now read the information on the tag. See Figure 3.5.
Only an authenticated reader can get the key of a tag. This way, compromised readers
cannot access the tags. When such a compromised reader tries to access the database, it
fails because it does not have a valid key to communicate with the main database. By using
keys, both the tags and readers are authenticated, and compromised readers are not allowed
in the network.
This method makes sure that compromised readers cannot read the tag information and
also provides mutual authentication between readers and tags and between readers and the
back-end database in the system. A tag comes in contact with a reader; the reader validates
the tag by sending a query to the tag. Now the tag generates and sends a random nonce number (nt ) to the reader. Also, it sends its unique ID which is encrypted with the random number by using its key Kt . Before the reader sends the tag-encrypted ID, which is (ID, nt )Kt ,
to the server, it will encrypt the message using the group key ((ID, nt )Kt )groupkey . Also, it
sends the random number of the tag (nt ) along with the ((ID, nt )Kt )groupkey . The server
has all information about the tags and even their ID. So, the server first will decrypt the
message by using the same shared group key that validates the reader. Then, it will search
in database to find the key of the tag and then encrypts the key of the tag by using the
shared group key. If the reader is genuine, the server will send the encrypted key of the tag
85

to the reader, which is (Kt )Groupkey . The reader in turn decrypts the message to get the
key Kt and sends the hash of the nonce number and the key [i.e., H1 (nt , Kt )] along with its
random number (nr ) to the tag. Because the tag knows the key Kt , it can verify that the
hash result received from the reader is valid. The tag will then send H2 , which is the hash
of (nr , Kt ) to the reader. The reader authenticates the tag by verifying the validity of the
hash result [hash(nr , Kt )] received from the tag. See Figure 3.5.
When a reader is compromised on the network, the group key and the main key will be
changed by using one of the rekeying strategies; the enterprise server will change the keys
and then distribute them to the facilities level. In case a compromised reader tries to connect
with the server to get the key of the tag, it encrypts the tag’s ID with its old group key
(since the compromised reader does not have the new group key) and sends it to the server.
However, the server will not be able to decrypt the message because there is a mismatch of
the group keys between the server and reader.

Table 3.1: The Steps For The Authenticated Reader in the first protocol
Server

Reader
1)
- Send query to the tag
- Encrypt (ID, nt ) using Kt
- Send (ID, nt )Kt and nt
3)
- Encrypt (ID, nt ) using the
group key
- Send [(ID, nt )Kt ]groupKey
to the server

4)
- Using the group key for
that reader
- Decrypt the message
[(ID, nt )Kt ]groupKey
Continued on next page
86

Tag
2)
- Generate nt

Table 3.1 (cont’d)
Server
- Add nt with all stored ID
- Encrypt the results with
the keys for each tag
- If the result matches the
received one
o Get the key Kt
o Decrypt (Kt ) using group
key
o Send (Kt )groupKey

Reader

Tag

5)
- Decrypt (Kt )groupKey
- Generate nr
- Calculate hash for Kt and
nt
- Send the hash along with
nr to the tag
6)
– Calculate hash for Kt and
nt
- Compare the result with
the one received
- If not the same:
o The reader is not authenticated
o Abort
- If the same:
o Calculate hash for Kt and
nr
o Send hash(Kt , nr ) to the
reader
7)
Calculate hash for for Kt
and nr
- Compare the result with
the received one
- If the same:
o The tag is authenticated
o Getting more information
from the tag
- If not the same:
o The tag is not authenticated
Continued on next page
87

Table 3.1 (cont’d)
Server

Reader
o Abort

Tag

Table 3.2: The Steps For The Compromised Reader in the first protocol
Server

Reader
1)
- Send query to the tag
- Encrypt (ID, nt ) using Kt
- Send (ID, nt )Kt and nt
3)
- Encrypt (ID, nt ) using the
group key (old one)
- Send [(ID, nt )Kt ]groupKey
to the server

Tag
2)
- Generate nt

4)
- Using the group key (new
one) for that reader
- Decrypt the message
[(ID, nt )Kt ]groupKey
- The server cannot do that
because of is matching
o Abort

Tables 3.1 and 3.2 shows the steps for the two situations (a compromised reader and a
genuine one) explained above.

3.4.2

Second Protocol

In this protocol, the concept is similar to the first protocol, but the number of times a reader
contacts the server is reduced. See Figure 3.6. Therefore, the second protocol is defined by
sharing the key between the server and the all tags. Also, the server will send an encrypted

88

message, from which a reader cannot get the information. The result is that only the system
tags will receive the information.

Figure 3.6: Second Protocol

Initially, all the tags have a shared key (Kt ) between the server and the tags. In addition,
every reader has a privacy bit (either 0 or 1) assigned by the server. The privacy bit 0
indicates the reader is not compromised and 1 indicates the reader is revoked. When a tag
is read by the reader, the reader will send a random number (m) and the encrypted message
(using Kt ) that is received from the server, which has the privacy bit and the unique key of
the reader Kr . This message will be changed only when one of the readers is added to or
revoked from the system.

89

Then, the tag checks the privacy bit of the reader after decrypting the message using
the key Kt. If the privacy bit is zero, then the tag will send the encrypted random number
(m) received from the reader using the key of the reader which is Kr . After verifying the
(m) by the reader, the tag will communicate with the reader using the key of the reader Kr
that is sent inside the decrypted message. If the privacy bit is one, the tag will not send
any more information to the reader. However, when one of the readers is compromised, the
server will change the message that has the privacy bit and the key of the reader, and will
send this new massage. The server will also change the privacy bit of the revoked reader to
1 to prevent the revoked readers from reading the tags again.
As shown in Figure 3.6, when a tag receives a query from the reader, the reader sends
out the message (privacy bit, Kr )Kt along with the random number (m). The tag checks
the privacy bit first after decrypting the message using Kt. There are two cases:
1. If the privacy bit is 1, the tag finds that the reader is revoked from the system and the
tag will stop sending its information to the revoked reader.
2. If the privacy bit is 0, the tag will encrypt the random number (m) using Kr , (m)Kr
and then the reader will verify the m.
The advantage of this protocol is that readers avoid contacting the server every time
a tag is read, which increases the efficiency of the system. The server is contacted only
when new readers are added or existing readers become compromised. This increases the
overall efficiency of the system as it reduces the number of times the servers are contacted.
The tag calculates only one encryption function and one decryption function to check the
authenticity of the reader, which means reduced cost and timed compared to the other
protocols. In addition, in the closed-system, a reader doesn’t have a shared key with a tag.
90

(This is different from what is usually assumed by the research community where a reader is
often assumed to have a shared key with a tag). This is reasonable; considering a terminal
reader in the Airport customs process, there is no way for a reader to have shared keys with
the RFID passports of travelers from different places/countries.
Table 3.3: The Steps For The Authenticated Reader in the second protocol
Server

Reader
1)
- Send query to the tag
2)
- Generate random number (m)
- Send (privacybit, kr )kt which is
received from the server

Tag

3)
- Decrypt the message to get privacy bit and kr
- Check the privacy bit, If bit=0
- Encrypt (m) using kr
4)
- Verify m by decrypting (m)kr

Table 3.4: The Steps For The Compromised Reader in the second protocol
Server
1) It changes the shared key
for the group of readers and
change the privacy bit to 1
for the compromised reader

Reader

Tag

2)
- Send query to the tag
3)
- Generate random number
(m)
Continued on next page

91

Table 3.4 (cont’d)
Server

Reader

Tag
- Send (privacybit, kr )kt
which was received from the
server after it is revoked
from the system
4)
- Decrypt the message to
get privacy bit and kr
- Check the privacy bit, If
bit=1
- The connection is aborted

Table 3.3 and 3.4 show the steps for the connection between the authenticated reader
with the tag and the compromised reader with the tag, respectively. When the system finds
a compromised reader, the server will change the privacy bit of that reader from 0 to 1.
Also, it encrypts a message again that includes the privacy bit and the key to that reader
and stores it to the compromised one.

3.4.3

Third Protocol

In this protocol, there will be another concept to provide authentication and privacy between
tags and readers using a trusted third party. Man-in-the-middle attacks and the symmetry
in the authentication protocols can be avoided with this protocol. In addition, a session
key is established in this protocol. A session key that is generated by the server will be
used to encrypt and decrypt messages between the tag and the reader using symmetric key
cryptography.
When the reader sends a query and random number nr to the tag, the tag will send the
following: its ID with the random number nt encrypted by using its key, which is Kt , the
92

Figure 3.7: Third Protocol
encrypted message of nr with reader’s name (R) using the key of the tag (nr , R)kt , along with
the random number (nt ). In turn the reader will forward the received messages ((ID, nt )kt
and (nr , R)kt ) to the server, sending the random number (nt ) too. Because the server knows
the Kt of the tag, it will decrypt the message to get nr and R. See Figure 3.7.
The server will then reply to the reader with two messages. The first one has nr , ID and
Ktr , which is the session key used between the tag and the reader, (nr , ID, Ktr )groupkey .
These three entries are encrypted using a group key that is shared between the subserver
and the readers in that server. The second message has R, and Ktr is encrypted using Kt ,
((R, Ktr )kt ). When the reader receives these two messages from the server, it will decrypt

93

the first message, which is (nr , ID, Ktr )groupkey , because it has already the group key to
get the session key and also to authenticate the tag. The reader will forward the second
message, which is (R, Ktr )Kt , to the tag. Also, when the reader gets the session key from
the first message, it will encrypt the random number (nt ), which is received from the tag,
and send it to let the tag authenticate the reader, using the session key Ktr .
Therefore, the tag will receive two messages (R, ktr )kt and (nt )ktr . The tag will decrypt
the first message to get the session key and verify the reader. Then, it will decrypt the
second one using the session key to verify and authenticate the reader. In this protocol, the
authentication and privacy will be provided with higher security. As a benefit, the reader will
never know the tag key. As a result, this prevents key leakage from a compromised reader.
Table 3.5 illustrates the steps for the connection between a reader, tag and the server in
authenticated situation.

Table 3.5: The Steps For The Authenticated Reader in the third protocol
Server

Reader
1)
- Send query to the tag
- Send random nr

Tag

2)
- Generate random number
(nt )
- Encrypt ID and nt using
its key kt
- Encrypt R and nr using kt
- Send them to the reader
3)
- Forward the received two
encrypted messages to the
server
4)
- Decrypt IDkt
Continued on next page
94

Table 3.5 (cont’d)
Server
- Decrypt (nr , R)kt
- Create two encrypted messages:
o (nr , ID, ktr )groupkey
o (R, ktr )kt
- Send them to the reader

Reader

Tag

5)
- Decrypt the message
(nr , ID, ktr )groupkey to get
the session key
- Encrypt (nt ) using the session key Ktr
- Forward (R, ktr )kt to the
tag
6)
- Decrypt (R, ktr )kt to get
the session key
- Decrypt (nt )ktr to verify
m

95

Chapter 4
Security and Time Analysis with
Conclusion and Future works
RFID tags have limited memory capacity. Hence, there is a strong need for new lightweight
cryptographic primitives that can be supported by low-cost RFID tags. The speed and
simplicity of an algorithm are usually qualifying factors; low complexity of the primitives
has a first place importance in this research. It also determines how attacks may happen in
an RFID system and what possible solutions can keep that from happening. After developing
and implementing the three protocols and the topology, security analysis and time analysis
are explained in this chapter.

4.1

Security analysis

In the three proposed protocols, there are no assumptions like the previous proposed solutions. Two situations have been handled: a compromised reader and a compromised tag by
providing a mutual authentication and privacy between them and the server. The following
subsections provide security analysis for each protocol.

96

4.1.1

Analyzing the first protocol

This protocol makes sure that compromised readers have no way to get the information of
the tag and also provides mutual authentication between readers and tags in the system. A
tag generates and a random nonce number (nt ) to the reader. In this way, replay attacks
cannot be applied because (ID, nt )Kt is different in each communication. The reader will
be authenticated when it encrypts the message using the group key ((ID, nt )Kt )groupkey
because this key is changed when one of the reader is revoked from the system. If the
reader is genuine, the server will send the encrypted key of the tag to the reader, which
is (Kt )Groupkey . The mutual authentication between readers and tags can be achieved by
verifying the two hash functions H1 and H2 .
In this protocol, mutual authentication and privacy have been achieved. Also, readers
will be verified. So, if there is a compromised reader, it will never get the private information
from the tag.
Security Analysis
• Mutual authentication The mutual authentication is achieved between readers and
tags and between readers and the server. The authentication between readers and tags
is achieved by verifying hash functions hash(nt , Kt ) , hash(nr , Kt ) . Also, the reader
has another opportunity to authenticate the tag (before verifying the hash function)
by getting the key of that tag from the server encrypted using the group key. The
reader is authenticated by the server from the message ((ID, nt )Kt )groupkey .
• Integrity The integrity of this protocol can be achieved in all the steps. The integrity
of communication is ensured by generating random numbers and encrypting the contents of the messages between reader, tags, and the servers. So, even if an attacker
97

gets all the transferred messages, the real messages cannot be revealed.
• Replay Attackers can record the messages between readers and tags and between the
server and readers. They can use later these messages for communications. However,
because tags generate random number (nt ) and readers generate also random number
(nr ), this attack can be prevented. In this case, attackers cannot replay previous
messages for communication.
• Man-in-the-middle An attacker can modify the messages or inject new messages
into the channels. This kind of attack is prevented in this protocol because mutual
authentications are achieved and random numbers are generated.
• Compromised readers A compromised reader is prevented to get the private information in the system by revoking it from the system and updating the main and
group keys. So, when a compromised reader gets the message (ID, nt )Kt ) from the
tag, it will encrypt the message using the old group key not the updated one. Then,
the server will know if this reader is an authenticated one or not by decrypting the
message ((ID, nt )Kt )groupkey using the updated key. In this case, the server will not
send the message (Kt )Groupkey to that reader.

4.1.2

Analyzing the second protocol

The number of times a reader contacts the server is reduced in this protocol. This is achieved
by sharing the key between the server and the all tags. Also, the server will send an encrypted message, from which a reader cannot get the information. Just the tags can get the
information. This is different from what is usually assumed by other research where a reader
is often assumed to have a shared key with a tag. When one of the readers is compromised,
98

the server will change the message (privacy bit, Kr )Kt , which has the privacy bit and the
key of the reader, and then it will send this new massage. The server will also update the
privacy bit of the revoked reader to 1 to prevent the revoked readers from reading the tags
again.
Security Analysis
• Mutual authentication The mutual authentication is achieved between readers and
tags and between readers and the server. The authentication between readers and tags
is achieved by verifying the message (privacy bit, Kr )Kt and (m)Kr . so, when the tag
receives the message (privacy bit, Kr )Kt , it will decrypt it and checks the privacy bit
to see if this reader is authorized by the server or not. If yes, then the reader can verify
the next message received from the tag, which is (m)Kr .
• Integrity The integrity of this protocol can be achieved. The integrity of communication is ensured by generating random numbers and encrypting the contents of the
messages between reader and tags. So, even if an attacker gets all the transferred
messages, the real messages cannot be revealed. The key leakage from a compromised
reader is prevented in this protocol because the reader never knows the key of the tags.
• Replay Attackers can record the messages between readers and tag. They can use
later these messages for communications. However in this protocol, readers generate
a random number (m). In this case, attackers cannot replay previous messages for
communication because of the random number.
• Man-in-the-middle An attacker can modify the messages or inject new messages
into the channels. This kind of attack is prevented in this protocol because mutual
authentications are achieved and random numbers are generated.
99

• Compromised readers A compromised reader is prevented from getting the private
information in the system by revoking it from the system and updating the message
(privacy bit, Kr )Kt . In this case, the tag will never send private information to the
compromised reader when it checks the privacy bit and sees that it is 1.

4.1.3

Analyzing the third protocol

Without a trusted third party, readers and tags share a secret key. However, with a trusted
third party, a reader and a tag share a secret key with the third party. The third protocol
is developed with this idea. There are many threads such as: inject a new message into
a channel, modify or delete a message in a channel, or replay and old message. The idea
of using trusted third party is to reduce the number of keys. The message (nr , R)kt is
send to the reader by the tag. Using R is for preventing (nr )kt from being reused by
attacker. The message ((ID, nt )kt is to prevent sniffing the ID of the tag for tracking
that tag. Also, the message (nr , ID, ktr )groupkey , which is sent by the server, is to make
the reader authenticate the tag, which in turns will authenticate the reader when receiving
the two messages (R, ktr )kt and (nt )ktr . In addition, the message (nr , ID, ktr )groupkey is
used to make sure that the compromised reader cannot decrypt it because it does not have
the updated group key. After providing the mutual authentication between the reader and
the tag with the server, the transferring messages between the tag and the reader will be
encrypted using the session key that is established by the server. This is used to ensure the
security of the communication between the tag and the reader. This key is used for both
encryption and decryption.
In this protocol, the mutual authentication and privacy have been achieved. Also, readers
will be verified. So, if there is a compromised reader, it will never get the private information
100

from the tag.
Security Analysis
• Mutual authentication In third protocol: the tag-to-reader authentication is achieved
when the reader verifies the value of (nr , ID, ktr )groupkey , which includes the random
number sent by the reader (nr ), the ID of the tag , and the session key (Ktr ) that will
be used for the communication between the reader and the tag. When the server sends
the ID of that tag, it means that this tag is an authenticated one. The reader-to-tag
authentication is achieved by receiving these two messages (R, ktr )kt and (nt )ktr . The
first message includes the R, which represents the reader, and the session key. So,
when the tag receives these two messages, it will know that the reader R is an authenticated one. Then, the tag will use the session key to verify its random number (nt ). In
addition, the readers can be an authenticated by the server by sending R to the server,
which in turn send the message (nr , ID, ktr )groupkey . If the reader is authenticated
one, it can use the group key to decrypt the message and to get the information. If
not, it cannot decrypt the message (nr , ID, ktr )groupkey using the old key because after
revoking the compromised reader from the system, the server will update the group
key.
• Integrity The integrity of this protocol can be achieved in all the steps. The integrity
of communication is ensured by generating random numbers and encrypting the contents between reader, tags, and the server. In addition, the integrity of this protocol
can be completed by using the session key that is established by the third party. This
key will be used between the reader and the tag after the mutual authentication is
satisfied to encrypt all the messages that are transferred between them. By using a
101

session key, the key leakage from a compromised reader is prevented with this protocol.
• Replay An attacker can record the messages between readers and tag and between
backend database and readers. They can use later these messages for communications.
However, because both tags and readers generate the random numbers (nt ) and (nr )
respectively, this attack can be prevented. In this case, an attacker cannot replay
previous messages for communication.
• Man-in-the-middle An attacker can modify the messages or inject new messages
into the channels. This kind of attack is prevented in this protocol because mutual
authentications are achieved and random numbers are generated. Also, symmetry of
the transferred messages in this protocol is avoided.
• Compromised readers A compromised reader is prevented in this protocol from
getting the private information in the system by revoking it from the system and
updating the group key. In this case, the compromised reader cannot decrypt the
message (nr , ID, ktr )groupkey using the old key because after revoking the compromised
reader from the system, the server will update the group key.
Table 4.1: The goals of the three developed protocols that have been achieved
❳❳❳
❳❳❳
Goal
❳
Protocol ❳❳❳❳❳

First
Second
Third

Mutual
authentication
√
√
√

Integrity
√
√
√

Replay
√
√
√

Man-inthe-middle
√
√
√

Compromised reader
√
√
√

Table 4.1 summarizes the goals that have been achieved in each developed protocol, but
102

in different ways.

4.2

Time analysis of the topology

A program for implementing the topology was written in Python language. A main server
and three sub servers are implemented to check the time consumed in deleting a compromised
reader from the system. The time is calculated in seconds. Figures 4.1 and 4.2 show the
results of deleting a reader from the system.

0.0006

0.0005

Time in second

0.0004

0.0003

0.0002

0.0001

0
0

5

10

15

20

25

30

35

40

Number of deleted readers

Figure 4.1: The execution time for deleting a compromised reader form the system

Figure 4.1 shows the execution time for deleting a reader sequentially. The execution
time decreases every time one of the readers is deleted form the system. The readers were
deleted from the first sub server and then from the second and so on.

103

In the second case, the readers are deleted from the system randomly. In Figure 4.2, the
execution time for deleting a reader decreases every time one of the readers is deleted from
the system. In this case, also, the time decreases similar to the first case.

0.0006

0.0005

Time in second

0.0004

0.0003

0.0002

0.0001

0.0000
0

5

10

15

20

25

30

35

Number of deleted readers

Figure 4.2: The execution time for deleting a compromised reader form the system randomly

4.3

Operations in an RFID tag

A tag in each protocol applies some steps for providing privacy and mutual authentication.
The number of steps is different in each protocol. In this section the operations that are
executed by a tag are defined: The operations that are executed in the first protocol by the
tag:
1. Generating random number
2. Applying encryption function
3. Applying hash function
104

4. Making compassion
5. Applying another hash function
The operations that are applied in the second protocol by the tag:
1. Applying decryption function
2. Applying encryption function
3. Checking the privacy bit
The operations that are applied in the third protocol by the tag:
1. Generating random number
2. Applying encryption function
3. Applying another encryption function
4. Applying decryption function
5. Applying another decryption function
6. Checking random number
Based on the above operations that are applied by the tag, the time is calculated in each
protocol as described in the next section.

105

4.4

Time analysis

In each protocol, the time is calculated in the tag to check how fast it is to execute operations for providing privacy and mutual authentication with readers. The time is calculated
in micro seconds. MSP430-P2274, which is a mixed-signal microcontroller from Texas Instruments, is used to calculate the time. MSP430-P2274 is used for low cost and low power
consumption embedded applications. [108] Its function the same as RFID tag. In addition,
IAR, which is an integrated development environment with C/C++ compiler is used in this
research to generate faster codes and to debug the microcontroller. It is powerful with the
shortest possible execution times. It is user friendly; it incorporates a compiler, an assembler, a linker and a debugger in one integrated development environment (IDE). This gives
an uninterrupted workflow. Three codes for each protocol are written in C programming
language. In each code, the time is calculated for the all operations that are run in the ‘tag’.
The following sections provide more details about the time in each protocol:

4.4.1

The time calculated in the first protocol

A tag performs five operations, as described above, in the first protocol. These are:
1. Generating random number
2. Applying encryption function
3. Applying hash function
4. Making compassion
5. Applying another hash function

106

For each operation, the time is calculated. Figure 4.3,4.4,4.5, and 4.6 show the results.
16000
14000

Time in Microsecond

12000
10000

Rt=Rr=2
Rt=3, Rr=2

8000

Rt=4, Rr=2
Rt=5, Rr=2

6000

Rt=6, Rr=2

Rt=7, Rr=2

4000
2000
0

0

100

200

300

400

500

600

Number of executions

Figure 4.3: Times, while increasing the number of digits in RFID tags only in the first
protocol

Figure 4.3,4.4,4.5, and 4.6 show six cases based on the number of digits in generating the
random number. Every time the number of digits increases, the time to run the calculation
also increases. In this protocol, two random numbers are generated, one by a tag (Rt ) and
the other one by a reader (Rr ).
In Figure 4.3, the random numbers started with two digits and then increased the number
of digits in the random number that is generated by the tag only in each case. The code for
the first protocol was run 500 times to check the stability of it.

107

In Figure 4.4, the number of digits in the random number generated by the reader is
increased, but not for the tag. Both random numbers started with two digits. The code for
the first protocol was also run 500 times to check the stability of it.

16000
14000

Time in Microsecond

12000
10000

Rt=Rr=2
Rr=3, Rt=2

8000

Rr=4, Rt=2
Rr=5, Rt=2

6000

Rr=6, Rt=2
Rr=7, Rt=2

4000
2000

0
0

100

200

300

400

500

600

Number of executions

Figure 4.4: Times, while increasing the number of digits in RFID readers only in the first
protocol

108

25000

Time in Microsecond

20000

Rt=Rr=2

15000

Rt=Rr=3

Rt=Rr=4
10000

Rt=Rr=5
Rt=Rr=6
Rt=Rr=7

5000

0

0

100

200

300

400

500

600

Number of executions

Figure 4.5: Times, while increasing the number of digits in both RFID tags and readers
equally in the first protocol

In Figure 4.5, the number of digits in the random number is increased equally in both
of the tag and the reader. Both random numbers started with two digits. The code for the
first protocol was also run 500 times to check the stability of it.

18000
16000

Time in Microsecond

14000
12000
10000
8000
6000
4000
2000
0

0

100

200

300

400

500

600

Number of executions

Figure 4.6: Times,with random digits in both RFID tags and readers in the first protocol
109

In Figure 4.6, the number of digits in the random number generated by the tag and the
reader were generated randomly. The code for the first protocol was also run 500 times. The
times in the figure are random, based on the number of digit in the random number that is
generated by the tag and by the reader in every run.

4.4.2

The time calculated in the second protocol

A tag performs three operations, as described above, in this protocol. These are:
1. Applying decryption function
2. Applying encryption function
3. Checking the privacy bit
For each operation, the time is calculated. Figure 4.7 and 4.8 show the results.
2500

Time in Microsecond

2000

2 digits

1500

3 digits
4 digits
5 digits

1000

6 digits

7 digits
500

0

0

100

200

300

400

500

600

Number of executions

Figure 4.7: Times, while increasing the number of digits of the random number in RFID
readers in the second protocol

110

Figure 4.7 and 4.8 show six cases based on the number of digits in the random number.
Every time the number increases, the time also increases. In this protocol, only one random
number is generated by the reader. In Figure 4.7, the number of digits in the random number
was increased by one, starting with two digits. The code for the second protocol was run
500 times to check the stability of it.
2500

Time in Microsecond

2000

1500

1000

500

0

0

100

200

300

400

500

600

Number of executions

Figure 4.8: Times, while generating random numbers with random digits in RFID readers
in the second protocol

In Figure 4.8, the number of digits in the random number generated by the reader is
generated randomly. The code for the second protocol was also run 500 times. The times in
the figure are random, based on the number of digit in the random number that is generated
by the reader in every run.

4.4.3

The time calculated in the third protocol

A tag performs six operations, as described above, in this protocol. These are:
1. Generating random number
111

2. Applying encryption function
3. Applying another encryption function
4. Applying decryption function
5. Applying another decryption function
6. Checking random number
For each operation, the time is calculated. Figure 4.9,4.10,4.11, and 4.12 show the results.
20000
18000

16000

Time in Microsecond

14000
Rt=Rr=2

12000

Rt=3,Rr=2

10000

Rt=4,Rr=2
Rt=5,Rr=2

8000

Rt=6,Rr=2

6000

Rt=7,Rr=2

4000
2000
0

0

100

200

300

400

500

600

Number of executions

Figure 4.9: Times, while increasing the number of digits in RFID tags only in the third
protocol

4.9,4.10,4.11, and 4.12 show six cases based on the number of digits in the random
number. Every time the number increased, the time also increased. In this protocol, two
random numbers were generated, one by the tag (Rt ) and the other one by the reader (Rr ).
In Figure 4.9, the number of digits in the random number generated by the tag is increased,
but not for the reader. Both random numbers started with two digits and then increased
112

them by one in every run. The code for the third protocol was run 500 times to check the
stability of it.
In Figure 4.10, the number of digit in the random number generated by the reader is
increased, but not for the tag. Both random numbers started with two digits and then
increased them by one in every run. The code for the third protocol was also run 500 times
to check the stability of it.

18000
16000

Time in Microsecond

14000
12000
Rt=Rr=2
10000

Rr=3,Rt=2
Rr=4,Rt=2

8000

Rr=5,Rt=2
6000

Rr=6,Rt=2

Rr=7,Rt=2

4000
2000
0
0

100

200

300

400

500

600

Number of executions

Figure 4.10: Times, while increasing the number of digits in RFID readers only in the third
protocol

In Figure 4.11, the number of digits in the random number was increased equally in both
of the tag and the reader. Both random numbers started with two digits. The code for the
third protocol was also run 500 times to check the stability of it.

113

30000

Time in microsecond

25000

20000
Rt=Rr=2
Rt=Rr=3

15000

Rt=Rr=4
Rt=Rr=5

10000

Rt=Rr=6
Rt=Rr=7

5000

0

0

100

200

300

400

500

600

Number of executions

Figure 4.11: Times, while increasing the number of digits in both RFID tags and readers
equally in the third protocol

25000

Time in Microsecond

20000

15000

Series1

10000

5000

0

0

100

200

300

400

500

600

Number of executions

Figure 4.12: Times, with random digits in both RFID tags and readers in the third protocol

In Figure 4.12, the number of digits in both random numbers (Rt and Rr ) was generated
randomly. The code for the third protocol was also run 500 times. The times in the figure
are random based on the number of digits in the random number generated by the tag and
114

by the reader in every run.
In summary, there are discrete time variables based on the number of steps and the
number of digits in the random number. As random numbers increase, the time to read
increases. Therefore, the number of communication transactions will be bounded within a
short time period by the number of tags in the field.
Table 4.2 shows the time range in microseconds of the second protocol; a reader generated
the random number for this protocol. Tables 4.3 and 4.4 show the time range in microseconds
of the first and the third protocols in two situations (the number of digits in a reader increased
without increasing the digits in a tag, and the number of digits in a tag increased without
increasing the digits in a reader) respectively. Table 4.5 shows the time range in microseconds
of the first and the third protocols when the number of digits in the random number of a
tag and reader were equal.
Table 4.2: The time duration in microseconds of the second proposed protocol
Protocol
2nd

2 digits
840-842

3 digits
1068-1071

4 digits
1298-1302

5 digits
1528-1535

6 digits
1763-1768

7 digits
1995-2004

Table 4.3: The time duration in microseconds of the first and third proposed protocols when
the number of digit in a reader increases
Protocol
1st
3rd

Rt =2, Rr =3
7871-7906
10903-10934

Rt =2, Rr =4
9499-9549
12384-12430

Rt =2, Rr =5
11134-11192
13865-13926

115

Rt =2, Rr =6
12771-12826
15350-15399

Rt =2, Rr =7
14398-14492
16834-16910

Table 4.4: The time duration in microseconds of the first and third proposed protocols when
the number of digits in a tag increased
Protocol
1st
3rd

Rr =2, Rt =3
7939-7971
10970-11002

Rr =2, Rt =4
9638-9682
12516-12565

Rr =2, Rt =5
11379-11437
14102-14160

Rr =2, Rt =6
13088-13141
15660-15711

Rr =2, Rt =7
14786-14879
17217-17296

Table 4.5: The time duration in microseconds of the first and third proposed protocols when
the random number of a tag and a reader were equal
Protocol
1st

Rt =2,
Rr =2
6246-6366

Rt =3,
Rr =3
9565-9603

3rd

9433-9456

1244512483

4.5

Conclusion

Rt =4,
Rr =4
1289212952
1546815525

Rt =5,
Rr =5
1626716340
1853118610

Rt =6,
Rr =6
1960419686
2157421657

Rt =7,
Rr =7
2296423083
2462324732

The developed topology, security requirements for the eSeal system, and proposed the eSeal
protection protocols are identified in this research. The security analysis shows that the
proposed protocols satisfy the security requirements, which are preventing man-in-the-middle
and reply attacks, providing integrity and mutual authentication, and revoking a compromise
reader from the system. The processing time in an eSeal requires several microseconds
using MSP430-P2274, which is used for a low cost and low power consumption embedded
application. In each protocol, the execution time is measured based on the number of
operations that are performed by an RFID tag and the number of digits in random numbers.
As a result, the execution time in the second protocol is less than the execution times in
the other protocols because the tag performs just three operations. Also, the execution time
in the first protocol is less than the execution times in the third one because the tag in

116

the first protocol performs five operations including one encryption function and two hash
functions. However, the tag in the third protocol performs two encryption functions and two
decryption functions. In all cases, the time jump was linear based on the number of digits
in the random number.

Uses 8 digits

Uses 5 digits

Port B

Port A

Figure 4.13: Customs using RFID system with secured protocols

The execution time for the tag depends on how many operation it performs. In addition,
it depends on the number of digits of the random number. As found, when the number of
digits of the random number increases, the execution time of the tag also increases. However,
changing the number of digits of the random number gives another benefit of security in the
system. After changing the number of digits in every run, the outputs from the tag are also
changed. For example, in Figure 4.13 port A and port B can have the same program but
different number of digits. If an attacker sniffs the output between the tag and the reader
in port A, s/he cannot match this output with the sniffed output in port B. Remember that
the real data are the same in both ports, but after applying the functions (encryption or

117

hash) the output is different. If an attacker acquired a suitable antenna, an RF receiver,
and a method to sample and record the data, s/he cannot execute an eavesdropping attack.
So, a tracking attack between ports is prevented. This means that the protocols provide
another benefit of security in customs application. When there is a shipment coming in to
a warehouse, the tag will be scanned and operate with this security protocol to designate
whether the shipment is legitimate, or not.
Hong Kong customs may use a lower number of random digits. Because there is counterfeiting, Hong Kong customs might require higher security. Therefore, they use a higher
number of random digits.
If enterprises, such as ports, are looking for lowest cost and high speed operations, they
can choose the second protocol.If they are looking for highest security and irrespective of
cost, they can choose the third protocol. If they are looking for the middle ground, with low
cost and medium security,they can choose the first protocol.
In summery, the following table 4.6 demonstrates the applicability of the three developed
protocols

Table 4.6: The differences between the three developed protocols
1st protocol
A tag performs five operations
It is selected when enterprises are looking for the
middle ground, with low
cost and medium security
Continued on next page

2nd protocol
A tag performs three operations
It is selected for lowest cost,
high speed operations

118

3rd protocol
A tag performs six operations
It is selected for highest security, irrespective of cost

Table 4.6 (cont’d)
1st protocol
It provides integrity by generating random numbers
and encrypting the contents
of the messages between
reader, tags, and the server
using the key of a reader
and a tag
It handles a compromised
reader because the server
sends an encrypted message
using an updated group key

It prevents man-in-themiddle attack because
mutual authentication was
achieved between readers
and the server; readers and
tags, and random numbers
were generated
It prevents replay attack
because attackers cannot
replay previous messages
for communication

Authorized readers knows
the key of the tag
The server does not generate a session key used between readers and tags. It
sends the key of a tag
The tag performs one encryption and two hash functions
Continued on next page

2nd protocol
It provides integrity by generating random numbers
and by using the shared
key between tags and the
server to encrypt the contents of the messages between reader, tags, and the
server
It handles a compromised
reader because the server
updates the message that
has a privacy bit from 0 to 1
after finding a compromised
reader
It prevents man-in-themiddle attack because
mutual
authentication
was achieved and random
numbers were generated

It prevents replay attack
because attackers cannot
replay previous messages
for communication and because of the random number
Authorized readers cannot
know the key of the tag
The server does not generate a session key. It uses
the shared key between the
server and the tags
The tag performs one encryption function and one
decryption function

119

3rd protocol
It provides integrity by using the session key that
is established by the third
party to encrypt all the
messages that are transferred between them

It handles a compromised
reader because the server
sends an encrypted message
that has a session key using a valid and an updated
group key
It prevents man-in-themiddle attack because
mutual
authentication
was achieved and random
numbers were generated.
Also, symmetry of the
transferred messages in this
protocol is avoided
It prevents replay attack
because attackers cannot
replay previous messages
for communication and because of the random number
All readers cannot know the
key of the tag. They use a
session key
The server generates a session key. It does not use the
keys of readers and tags
The tag performs two encryption functions and two
decryption functions

Table 4.6 (cont’d)
1st protocol
There is a connection between the server and readers every time readers contact tags

A reader performs one encryption function, one decryption function, and two
hash functions
If there is a compromised
reader, the reader can get
the encrypted messages
even if it cannot decrypt
them

4.6

2nd protocol
The number of times a
reader contacts the server
is reduced. Only when one
of readers in the system
is compromised does the
server contact the reader
A reader performs one decryption function only

3rd protocol
There is a connection between the server and readers every time readers contact tags

If there is a compromised
reader, the reader cannot
get any information from
the tags and servers even if
they are encrypted

If there is a compromised
reader, the reader can get
the encrypted messages
even if it cannot decrypt
them

A reader performs one decryption function and one
encryption function

Future Works

Ports worldwide, like Hong Kong, need to be monitored for security requirements based
on increased counterfeiting or increased shipments movement. In this research, encryption,
decryption, and hash functions are used for providing security. There are many types of
algorithms for these functions. Therefore, different algorithms for these functions can be
applied in the system to figure out the calculated time and compare them with the group of
time that are calculated in this research. Also, the results from this research can be compared
with other research, such as Chiew’s works in evaluating the time cost using different types
of encryption, decryption, and hash functions. [106] The calculation times are gotten when
the functions are applied within the tag. In future works, this should be expanded to include
all units (readers and servers) in order to analyze the time for the whole system. The path

120

way to do this has been created in this research.

121

REFERENCES

122

REFERENCES
[1]

Bing Liang, Security And Performance Analysis For RFID Protocols,2010.

[2]

Melanie Rieback, Bruno Crispo, and Andrew Tanenbaum. The Evolution of RFID
Security, IEEE Pervasive Computing, 5(1):62?69, January?March 2006.

[3]

Alex Macario, MD, MBA; Dean Morris, MBA; Sharon Morris, RN, BSN, CNOR
, Initial Clinical Evaluation of a Handheld Device for Detecting Retained Surgical
Gauze Sponges Using Radiofrequency Identification Technology, 2006.

[4]

Timothy Hay, Using RFID To Track Surgical Sponges Left In The Body, August
31,2010.

[5]

Ari Juels. RFID Security and Privacy: A Research Survey, IEEE Journal on Selected
Areas in Communications, 24(2):381?394, February 2006.

[6]

Pawel Rotter, A Framework for Assessing RFID System Security and Privacy Risks,
IEEE Pervasive Computing, 7(2):70?77, June 2008.

[7]

Michael O. Leavitt, An integrated strategy for protecting the nation?s food supply,
Department of Health and Human Services/ Food and Drug Administration, Nov.
2007.

[8]

Rudolf, P. M. and I. B. G. Bernstein, Counterfeit Drugs, New England Journal of
Medicine Volume 350:1384-1386, Number 14,2004.

[9]

Acheson, D., Food Protection, Food Safety and Food Defense, Association of Food and
Drug Officials (AFDO) Annual Conference. San Antonio, Texas, Association of Food
and Drug Officials (AFDO), 2007.

[10]

Mark Roberti, DOD Releases Final RFID Policy, 2005.

[11]

Defense Industry Daily staff, RFID Technology: Keeping Track of DoD?s Stuff, Jul
13, 2010.

[12]

David Molnar and David Wagner, Privacy and Security in Library RFID: Issues,
Practices, and Architectures, In Birgit Pfitzmann and Peng Liu, editors, Conference
123

on Computer and Communications Security ? ACM CCS, pages 210?219, Washington,
DC, USA, ACM, ACM Press, October 2004.
[13]

Melanie Rieback, Bruno Crispo, and Andrew Tanenbaum, The Evolution of RFID
Security, IEEE Pervasive Computing, 5(1):62-69,January-March 2006.

[14]

Liu, A. and Bailey, L., PAP: A Privacy and Authentication Protocol for Passive RFID
Tags, 2005.

[15]

Michael Grimalia, RFID Security Concerns, ISSA Journal, Feb 2007.

[16]

Dirk Henrici, RFID Security and Privacy: Concepts, Protocols, and Architectures,
Volume 17, pp 7-56, 2008.

[17]

Ken Traub, Felice Armenio, The EPCglobal Architecture Framework, Dec 2010.

[18]

Intermec, The Write Stuff: Understanding The Value of Read/ Write RFID Functionality, 2004.

[19]

Klaus Finkenzeller, RFID HANDBOOK: Radio Frequency Identification Fundamentals and Applications, 2000.

[20]

Genn W. Lee, The Auto ID Book: Bar Coding and Automatic Identification Technologies, Informatics, Inc, 1998.

[21]

Chris Riley, David Benyon, Graham I. Johnson and Kathy Buckner, Security in context: investigating the impact of context on attitudes towards biometric technology,
BCS ’10 Proceedings of the 24th BCS Interaction Specialist Group Conference Pages
108-116 , Sept 2010.

[22]

Alexander P.Pons and Peter Polak, Understanding User Perspectives on Biometric
Technology, Communication of the ACM, Volume 51, Issue 9, Sept 2008.

[23]

Won J. Jun, Smart Card Technology Capabilities, July 8, 2003.

[24]

Daniel Kouril, Ludek Matyska, and Michal Prochazka, Improving Security in Grids
Using the Smart Card Technology, IEEE Computer Society, September 2006.

[25]

Husemann, D. The smart card: don?t leave home without it, IEEE Concurrency 7, 2,
24?27 (April?June 1999).
124

[26]

Katherine M. Shelfer, and J. Drew Procaccino, Smart Card Evolution, Communications of the ACM , Volume 45 Issue 7, July 2002.

[27]

Thomas Roder, Smart Cards Solutions: Bringing Value to Citizens, May 2012.

[28]

Fletcher, P. , Europe holds a winning hand with smart cards, Electronic Design 47, 1
, 106, (Jan. 11, 1999).

[29]

A Summary of RFID Standards, RFID Journal, Inc, 2005.

[30]

RFID Standards, RFID in Europe, 2012.

[31]

EPCglobal, EPCglobal Tag Data Standards Version 1.5, EPCglobal Ratified
Standard, August 2010, http://www.gs1.org/gsmp/kc/epcglobal/architecture/
architecture_1_4-framework-20101215.pdf

[32]

Divyan M. Konidala, Woan-Sik Kim, and Kwangjo Kim, Security Assessment of EPCglobal Architecture Framework, April 2006.

[33]

Mark Brown, Sam Patadia, and Sanjiv Dua, Mike Meyer?s Certification Passport/
RFID+Certification, McGraw-Hill Professional Publishing, 1st edition, 2007.

[34]

Parikh, D. and Jancke, G., Localization and Segmentation of A 2D High Capacity
Color Barcode, IEEE , 2008.

[35]

GS1 DataMatrix, An Introduction and technical overview of the most advanced GS1
Application Identifiers compliant symbology, 2011.

[36]

Bo Chen , Xiao-hui Qiang , and Ling Yu, A lightweight color barcode algorithm and
application in mobile e-commerce,Computer Science and Education (ICCSE), IEEE
8th International Conference on , 2013.

[37]

Maeva, A. and Severin, F., High Resolution Ultrasonic Method for 3D Fingerprint
Recognizable Characteristics In Biometrics Identification , IEEE International, 2009.

[38]

Zhen Liang , Fei Tan , and Zheru Chi, Video-based biometric identification using eye
tracking technique, Signal Processing, Communication and Computing (ICSPCC),
2012 IEEE International Conference , p (728-733), 2012.

125

[39]

Taherdoost, H. and Masrom, M., An examination of smart card technology acceptance
using adoption model , Information Technology Interfaces, 2009. ITI ’09. Proceedings
of the ITI 2009 31st International Conference on , p (329-334), 2009.

[40]

Smart Card Alliance, Benefits of Smart Cards versus Magnetic Stripe Cards for
Healthcare Applications, 2011-2012.

[41]

Smart Card Alliance, Smart Card Technology in U.S.Healthcare:Frequently Asked
Questions, A Smart Card Alliance Healthcare Council Publication, Sept 2012.

[42]

Christopher J. Dyball and Terri Lichtenstein, Optical memory cards provide secure
identification, LaserFocusWorld, International Resource for Technology and Applications in the Global Photonics Industry, Nov 2001.

[43]

Yu-Yi Chen and Meng-Lin Tsai, The Study on Secure RFID Authentication and Access
Control, Current Trends and Challenges in RFID, Prof. Cornel Turcu (Ed.), ISBN:
978-953-307-356-9, InTech, DOI: 10.5772/20750, 2011.

[44]

Robinson, G. and Clarke R., Effects of Radio Frequency Noise on Communication
Capabilities of Ultra-High Frequency Passive Transponder, Master thesis, Michigan
State University, Packaging, 2010.

[45]

Thamae, L.Z. , Wu, Z. and Konrad, W, Propagation characteristics of a 2.45 GHz
microwave radio frequency identification system ,IET Journals, vol 3, issue 1, 2009.

[46]

S.P. Singh, M. McCartney, J. Singh, and R. Clarke, RFID research and testing
for package of apparel, consumer goods and fresh produce in the retail distribution
environment, Packaging Technology and Science, P 91-102,2008.

[47]

GS1 General specification, version 13.1, Issue 2, Jul 2013.

[48]

MICROSCAN, Machine Vision and Auto ID/ Track, Trace and Control Solution,
2013.

[49]

Yud-Ren Chen, Kuangkin Chao, and Moon S Kim, Machine vision technology for
agricultural application, Computers and Electronics in Agriculture,Vol 36, Issues 2?3,
Pages 173?191, November 2002.

[50]

Sinopec group, Industrial Lubricant Manufacturer Relies on Machine Vision for Product Quality and Traceability , Food and Beverage Packaging, 2013.

126

[51]

Wayne E. Bailey, Fourth-Generation Sweet Potato Supplier Incorporates Traceability
System in its Packing Line for PTI Compliance, Food and Beverage Packaging, 2013.

[52]

Ward, M and Kranenburg, R., RFID: Frequency, standards, adoption and innovation,
JISC Technology and Standards Watch, May 2006.

[53]

AIM Inc., Linear Symbologies, https://aimglobal.site-ym.com/?page=Linear_
symb.

[54]

RFID Adoption and Implication, A Sectoral e-Business Watch study by IDC / Global
Retail Insights, study report, No. 07 /Sept 2008.

[55]

The RFID Network,
Omni-ID Unveils The Ultimate Passive Tag
with
135
foot
Read
Range,
http://rfid.net/product-listing/
reviews/183-omni-id-unveils-the-ultimate-passive-tag-with-135
-foot-read-range,Oct 2011.

[56]

Stephen A. Weis, Sanjay E. Sarma, Ronald L. Rivest and Daniel W. Engels, Security
and Privacy Aspect of Low-Cost Radio Frequency Identification Systems, Security in
Pervasive Computing, Vol 2802, 2004, pp 201-2012.

[57]

Smart Border Alliance, RFID Security and Privacy , RFID Security and Privacy white
paper, http://www.dhs.gov/xlibrary/assets/foia/US-VISIT_RFIDattachE.pdf

[58]

Gary Stoneburner, Computer Security, Underlying Technical Models for Information
Technology Security, NIST, Dec 2001.

[59]

M. Rieback, RFID Malware: Truth vs myth, IEEE Computer Society, 2006.

[60]

Rieback, M.; Simpson, P.; Crispo, B. and Tanenbaum, A., RFID Malware: Design
Principles and Examples, IEEE PErCom, 2006.

[61]

Rieback, M.; Simpson, P.; Crispo, B. and Tanenbaum, A., Is Your Cat Infected with
a Computer Virus, IEEE PErCom, 2006.

[62]

Michael Grimalia, RFID Security Concerns, ISSA Journal, Feb 2007.

[63]

Pedro Peris-Lopez, Julio Cesar Hernandez-Castro, Juan M. Estevez-Tapiador, and
Arturo Ribagorda, Attacking RFID Systems, in Wireless Networks and Mobile Communications Series: Security in RFID and Sensor Networks, CRC Press, ch. 2, pp.
29-48, 2009.
127

[64]

Alexei Czeskis, Karl Koscher, Joshua R. Smith, and Tadayoshi Kohno, RFIDs and
Secret Handshakes: Defending Against Ghost-and-Leech Attacks and Unauthorized
Reads with Context-Aware Communications, ACM, Oct 2008.

[65]

Chia-hung Huang, An Overview of RFID Technology, Application, and Security/Privacy Threats and Solutions, Spring 2009.

[66]

RFID Security, The Government of Hong Kong Special Administrative Region, Feb
2008.

[67]

Rishab Nithyanand, Gene Tsudik, and Ersin Uzun, Readers Behaving Badly/ Reader
Revocation in PKI-Based RFID Systems, Cryptoloy ePrint Archive, Report 2009 /
465, 2009.

[68]

Juels, A., Minimalist Cryptogarphy for Low-Cost RFID Tags, RSA Laboratories. Bedford, 2005.

[69]

Juels, A, Strengthening EPC tags against Cloning, Proceedings of the 4th ACM workshop on wireless security, 67-76, Sept 2005.

[70]

Peris-Lopez, P.; Hernandez-Castro, J.; Estevez-Tapiador, J. and Ribagorda, A., A
Minimalist Mutual-Authentication Protocol for Low-cost RFID Tags, Academic Press,
Inc, 2006.

[71]

Rivest, R., The RC5 Encryption Algorithm, Proceedings of the Second International
Workshop on Fast Software Encryption, 1994.

[72]

Smith, J; Sample, A.; Powlwdge, P.; Roy, S. and Mamishev, A., A WirelesslyPoweredPlatform for Sensing and Computation, Springer-Verlag Berlin Heidelberg
,2006.

[73]

Bailey, Daniel, and Juels, Ari, Shoehorning Security into the EPC Standard, International Conference on Security in Communication Networks - SCN 2006, p. 303-320,
2006.

[74]

Juels, Ari, Vision of RFID security, Proceedings of the IEEE 95(8): 1507-1508. August, 2007.

[75]

University of Arkansas, Fayetteville (2009, November 19). Fingerprinting’ RFID
tags:Researchers develop anti-counterfeiting technology. Science Daily. Retrieved January 14, 2011.
128

[76]

Yao, Y.; Marcialis, G; Pontil, M.; Frasconi, P. and Roli, F., A New Machine Learning
Approach to Fingerprint Classification, Advances in Artificial Intelligent, 2001.

[77]

Chinnappa Gounder Periaswamy, S.; Thompson, D.; Di, J., Fingerprinting RFID
Tags, IEEE Computer Society, October 2010.

[78]

Danev, B.; Heydt-Benjamin, T. and Capkun, S., Physical-layer Identification of RFID
Device, Proceeding SSYM’09 Proceedings of the 18th conference on USENIX security
symposium, 2009.

[79]

Saparkhojayev, N and Thompson, D.R., Matching Electronic Fingerprints of RFID
Tags Using the Hotelling’s Algorithm, Sensors Applications Symposium, IEEE, February 2009.

[80]

PERIASWAMY, SCG.; THOMPSON, D. and ROMERO, H., Fingerprinting Radio
Frequency Identification Tags Using Timing Characteristics, 2010.

[81]

Elaine Barker, William Barker, William Burr, William Polk, and Miles Smid, Computer Security/ Recommendation for Key Management, NIST, July 2012.

[82]

Ari Juels, Ravikanth Pappu, and Bryan Pamo, Unidirectional Key Distribution Access Time and Space with Applications to RFID Security, Proceedings of the 17th
conference on Security Symposium, July 2008.

[83]

Charlie Kaufman, Radia Perlman, and Mike Speciner, Network Security/ Private
Communication in a Public World, Second Edition, 2008.

[84]

Matt Ball, Key Management Summit, IEEE Computer Society, slides show and videos
hosted by MSST, May 2010.

[85]

Key Management Lifecycle, National Institute of Standards and Technology, U.S. Department of Commerce, http://csrc.nist.gov/groups/ST/toolkit/documents/
kms/lifecycle\%20slides\%20\%28b-w\%29.pdf.

[86]

A. Juels, R. L. Rivest, and M. Szydlo, The blocker tag: Selective blocking of RFID
tags for consumer privacy, In Proceeding of the 10th ACM conference on Computer
and Communication security, page 103-111, 2003.

[87]

S. A. Weis, S. E. Sarma, R. L. Rivest, and D. W. Engels, Security and privacy aspects
of low-cost radio frequency identification systems, In Security in Pervasive Computing,
volume 2802 ,pages 201?212, 2004.
129

[88]

Sung-Ho Kim, Hong-Jin Lee, Han-Wool Jung, Beom-Ki Maeng, and Yongsu Park,
IPAP: Improved Privacy and Authentication Protocol for Passive RFID Tags, Network
Infrastructure and Digital Content, IEEE, 2010.

[89]

A. Juels, P. Syverson, and D. Bailey, High-Power Proxies for Enhancing RFID Privacy
and Utility, Privacy Enhancing Technologies (PET) Workshop, pp. 210-226. 2005.

[90]

Alex X. Liu, LeRoy A. Bailey and Adithya H. Krishnamurthy, RFIDGuard: a
Lightweight privacy and authentication protocol for passive RFID tags, Security and
Communication Networks, page 384-393, 2010.

[91]

A. Juels, D. Molnar, and D. Wagner, Security and privacy issues in e-passports,
In Security and Privacy for Emerging Areas in Communications Networks ? SECURECOMM, 2005.

[92]

R. Housley, W. Ford, W. Polk, and D. Solo. RFC 2459: Internet X.509 public key infrastructure Certificate and Certificate Revocation List (CRL) Profile, Network Working Group, April 2002.

[93]

R. Nithyanand, G. Tsudik, and E. Uzun, Readers behaving badly: reader revocation in
PKI-based RFID systems, ESORICS’10 Proceedings of the 15th European conference
on Research in computer security, page 19-36, 2010.

[94]

Pearson, J., Securing the pharmaceutical supply chain with RFID and public key infrastructure (PKI) technologies, in Texas Instruments White Paper, 2005.

[95]

Hakeem, M., Raahemifar, K., and Khan, G., A Novel Key Management Protocol for
RFID System, IWCMC, page 1107-1113, IEEE, 2013.

[96]

Chung Kei, Mohamed Gouda, and Simon Lam, Secure Group Communications Using
Key Graphs ,2000.

[97]

Claire Swedberg, Hong Kong Customs Moves Forward With E-Lock Plans, May 4,
2012.

[98]

Asia-Pacific Economic Cooperation (APEC), Using RFID for customs control transit
containers, submitted by Chinese Taipei, Feb 2009.

[99]

EPCglobal: Low Level Reader Protocol (LLRP), ver 1.1, Oct 13, 2010, http://www.
gs1.org/gsmp/kc/epcglobal/llrp/llrp_1_1-standard-20101013.pdf

130

[100] EPCglobal: Reader Management (RM) ver 1.0.1, May 31, 2007, http://www.gs1.
org/gsmp/kc/epcglobal/rm/rm_1_0_1-standard-20070531.pdf
[101] Kateryna Daschkovska and Bernd Scholz-Reiter, Electronic Seals for Efficient Container Logistics, Dynamics in Logistics, Springer Berlin Heidelberg , pp 305-312, 2008.
[102] Confidex supplies the EPC Gen2 compliant RFID e-seal in multi-million dollar
contract, Jan 11,2011, IT reseller, http://www.itrportal.com/articles/2011/
01/11/6319-confidex-supplies-the-epc-gen2-compliant-rfid-e-seal url-inmulti-million-dollar
[103] World Customs Organization, Customs capacity to fight counterfeits strengthened by
WCO/GS1 cooperation agreement, Brussels, Sept 6 2012, http://www.wcoomd.org/
press/?v=1&lid=1&cid=14&id=316
[104] Dong Kyue Kim, Mun-Kyu Lee, You Sung Kang, Sang-Hwa Chung, Won-Ju Yoon,
Jung-Ki Min, and Howon Kim, Design and Performance Analysis of Electronic Seal
Protection Systems Based on AES, ETRI Journal, vol. 29, no. 6, pp. 755-768, Dec.
2007. http://dx.doi.org/10.4218/etrij.07.0107.0068
[105] T. Drake and J. Reinold, ISO Study: Vulnerabilities and Threats for Container Identification Tags and e-Seals, 2005.
[106] Chiew, K., Yingjiu Li, Tieyan Li, Deng, R.-H., Aigner, M., Time Cost Evaluation for
Executing RFID Authentication Protocols, Internet of Things (IOT) , vol., no., pp.1,8,
Nov. 29 2010-Dec. 1 2010.
[107] Chae, H.-J., D. J. Yeager, J. R. Smith, and K. Fu, Maximalist cryptography and
computation on the WISP UHF RFID tag, in The Conference on RFID Security 2007
(RFIDSec 07), Malaga, Spain, Jul. 11?13, 2007.
[108] MSP430-P2274 development board, Users Manual, Olimex,July 2009.

131