Internet Access in Vehicular Networks Wenchao Xu

Read Anytime Anywhere Easy Ebook Downloads at ebookmeta.com
Internet Access in Vehicular Networks Wenchao Xu
https://ebookmeta.com/product/internet-access-in-vehicular-
networks-wenchao-xu/
OR CLICK HERE
DOWLOAD EBOOK
Visit and Get More Ebook Downloads Instantly at https://ebookmeta.com

Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.
Link Layer Cooperative Communication in Vehicular Networks
Sailesh Bharati Weihua Zhuang
https://ebookmeta.com/product/link-layer-cooperative-communication-in-
vehicular-networks-sailesh-bharati-weihua-zhuang/
ebookmeta.com
Cloud and IoT Based Vehicular Ad Hoc Networks 1st Edition
Gurinder Singh
https://ebookmeta.com/product/cloud-and-iot-based-vehicular-ad-hoc-
networks-1st-edition-gurinder-singh/
ebookmeta.com
Advanced Fiber Access Networks 1st Edition Cedric F. Lam
https://ebookmeta.com/product/advanced-fiber-access-networks-1st-
edition-cedric-f-lam/
ebookmeta.com
Fundamentals of Modern VLSI Devices 3rd Edition Yuan Taur
https://ebookmeta.com/product/fundamentals-of-modern-vlsi-devices-3rd-
edition-yuan-taur/
ebookmeta.com

Unwatchable Nicholas Baer Maggie Hennefeld Laura Horak
Gunnar Iversen
https://ebookmeta.com/product/unwatchable-nicholas-baer-maggie-
hennefeld-laura-horak-gunnar-iversen/
ebookmeta.com
NFPA 101A guide on alternative approaches to life safety
2016 edition. Edition National Fire Protection Association
https://ebookmeta.com/product/nfpa-101a-guide-on-alternative-
approaches-to-life-safety-2016-edition-edition-national-fire-
protection-association/
ebookmeta.com
Dawn of Vengeance A Post Apocalyptic EMP Thriller World
Without Power Book 10 1st Edition Ryan Casey
https://ebookmeta.com/product/dawn-of-vengeance-a-post-apocalyptic-
emp-thriller-world-without-power-book-10-1st-edition-ryan-casey/
ebookmeta.com
Stray Shifter Moon Crossed Wolves 3 1st Edition Aimee
Easterling
https://ebookmeta.com/product/stray-shifter-moon-crossed-wolves-3-1st-
edition-aimee-easterling/
ebookmeta.com
Basic Biology and Clinical Aspects of Inflammation 1st
Edition Robert F. Diegelmann
https://ebookmeta.com/product/basic-biology-and-clinical-aspects-of-
inflammation-1st-edition-robert-f-diegelmann/
ebookmeta.com

Gripping GAAP 2022 1st Edition Catherynn Service
https://ebookmeta.com/product/gripping-gaap-2022-1st-edition-
catherynn-service/
ebookmeta.com

Wenchao Xu
Haibo Zhou
Xuemin (Sherman) Shen
Internet
Access in
Vehicular
Networks

Internet Access in Vehicular Networks

Wenchao Xu • Haibo Zhou
Internet Access in Vehicular
Networks

Wenchao Xu
Department of Computing
Hong Kong Polytechnic University
Hung Hom, Hong Kong
Haibo Zhou
School of Electronic Science
and Engineering
Nanjing University
Nanjing, China
Electrical and Computer Engineering
Department
University of Waterloo
Waterloo, ON, Canada
ISBN 978-3-030-88990-6 ISBN 978-3-030-88991-3 (eBook)
https://doi.org/10.1007/978-3-030-88991-3
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland
AG 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface
Modern vehicles generate a large amount of data for many emerging automotive
applications, such as road safety, traffic management, autonomous driving, and
intelligent transportation system. The Internet of vehicles are expected to enable
effective acquiring, storage, transmission, and computing for such big data among
vehicle users and thus can facilitate better perception of both internal and external
vehicular environments as well as the status of drivers, passengers, and pedestrians.
To efficiently connect mobile vehicles to the Internet and conduct massive infor-
mation exchange among vehicle users and the transportation system, it is essential
to evaluate the Internet access performance via both the vehicle-to-roadside (V2R)
and vehicle-to-vehicle (V2V) paradigms. Besides, in order to support reliable and
efficient Internet access for mobile vehicle users, it is very important to explore
various spectrum resources rather than solely relying on cellular networks, such
as unlicensed WiFi band, TV White Space, opportunistic V2V data path. To set
up effective Internet connection, a practical signaling process between a roadside
access station and vehicle users is necessary, e.g., various management frames
need to be transferred to set up effective Internet connections. As vehicles move
around, it is possible for them to set up opportunistic connections to share the data
contents, bandwidth, and computing capacities, which can enable a variety of novel
computing and communication paradigms that are beneficial to future automotive
applications. In this monograph, we investigate the Internet access procedure and
the corresponding analytical evaluation methods, as well as novel machine learning
paradigms for reliable and robust Internet connectivity on the road. We first
introduce the Internet access of vehicles and propose an analytical framework for
modeling of Internet access performance via the roadside hotspots, considering the
necessary Internet access procedure that comprised of association, authentication,
and network configuration steps, where the access delay and throughput capacity
are evaluated in drive-thru Internet scenario. We then explore the interworking of
different V2X communication paradigms and study the opportunistic assistance
from neighboring vehicles, which apply the V2V communication to conduct
Internet data offloading upon the interworking of V2V and V2R communication,
where the trade-off between the delay and throughput of the V2V assistance is
v

vi Preface
analyzed. In addition, we take a close look at the wireless link management between
the vehicle and Internet access stations. We investigate the V2X channel that is
highly varying and thus difficult to accommodate proper modulation and coding
scheme to satisfy various user quality-of-service (QoS). To deal with such issue, we
apply big data analytics and show that the proposed data-driven and learning-based
methods can greatly reduce the packet drop rate and thus improve the Internet access
performance in terms of both access delay and transmission throughput. Several
case studies are presented to examine the utility of the big vehicular data to enable
the intelligent Internet access. Furthermore, to train the machine learning models
among vehicle users in a distributed manner, we design efficient IoV protocols to
boost the training process, including the rateless coding-based broadcasting scheme
for intelligent model delivery that can enhance the process of collaborative learning
among vehicles, whereby asynchronous federated learning can be conducted for
mobile vehicles with high mobility and opportunistic inter-contacts. We hope that
this monograph will provide inspiration and guidance on further research and
development of the future Internet of vehicles.
We would like to thank Prof. Weihua Zhuang from the University of Waterloo
(UW), Prof. Song Guo from The Hong Kong Polytechnic University, Prof. Nan
Cheng from the Xidian University, and many UW BBCR members for their
contributions in the contents of this monograph and great support to related research
projects. Special thanks are also due to the staff at Springer Science+Business
Media: Susan Lagerstrom-Fife, for her great help in the publication of the mono-
graph.
Hung Hom, Hong Kong Wenchao Xu
Nanjing, China Haibo Zhou
Waterloo, ON, Canada Xuemin (Sherman) Shen

Contents
1 Introduction of Internet Access of Vehicular Networks ................. 1
1.1 Internet of Vehicles Overview .......................................... 1
1.1.1 DSRC............................................................ 3
1.1.2 ISM Band WiFi with Opportunistic Access ................... 4
1.1.3 TVWS with Cognitive Spectrum Access ...................... 5
1.1.4 Cellular IoV ..................................................... 5
1.1.5 Summary ........................................................ 6
1.2 Internet Access Procedure .............................................. 6
1.2.1 Network Detection.............................................. 6
1.2.2 Authentication .................................................. 8
1.2.3 Network Parameters Assignment .............................. 9
1.2.4 Summary ........................................................ 9
1.3 Aim of the Book ........................................................ 10
References ..................................................................... 10
2 Internet Access Modeling for Vehicular Connection..................... 13
2.1 Background and Motivation ............................................ 13
2.2 Delay Analysis of Vehicular Internet Access .......................... 16
2.2.1 System Model................................................... 17
2.2.2 Access Delay Analysis ......................................... 21
2.2.3 Delay Analysis and Simulation ................................ 26
2.2.4 Experiment ...................................................... 30
2.2.5 Summary ........................................................ 36
2.3 Throughput Capacity Analysis of Drive-Thru Internet................ 36
2.3.1 System Model................................................... 37
2.3.2 3D Markov Chain Based Throughput Analysis ............... 39
2.3.3 Simulation Results .............................................. 48
2.3.4 Summary ........................................................ 54
References ..................................................................... 55
3 V2X Interworking via Vehicular Internet Access ........................ 57
vii

viii Contents
3.2 Queueing Model for Opportunistic V2V Assistance .................. 58
3.2.1 System Model................................................... 58
3.2.2 Queueing Model ................................................ 59
3.2.3 Queueing Analysis About V2V Communication ............. 60
3.2.4 Simulation and Discussion ..................................... 65
3.2.5 Summary ........................................................ 70
3.3 Vehicular Offloading via V2X Interworking........................... 70
3.3.1 System Model................................................... 70
3.3.2 Offloading Performance Analysis.............................. 74
3.3.3 Access Delay Approximation .................................. 74
3.3.4 V2V Assistance Analysis....................................... 77
3.3.5 Simulation and Verification .................................... 78
3.3.6 Summary ........................................................ 81
References ..................................................................... 81
4 Intelligent Link Management for Vehicular Internet Access ........... 83
4.2 Reinforcement Learning Based Link Adaptation for
Drive-Thru Internet ..................................................... 84
4.2.1 System Model and Problem Formulation...................... 86
4.2.2 Problem Formulation ........................................... 88
4.2.3 RL Based RA Design........................................... 89
4.2.4 Performance Evaluation and Discussion ...................... 92
4.2.5 Experiment Setup ............................................... 92
4.2.6 Performance Evaluation ........................................ 94
4.2.7 Feasibility Analysis............................................. 96
4.2.8 Summary ........................................................ 100
4.3 Deep Learning Classifier Enabled Rate Adaptation for
802.11af TVWS Vehicular Internet Access............................ 100
4.3.1 System Model................................................... 100
4.3.2 Problem Formulation and DL Solution ........................ 102
4.3.3 Evaluation of the TSC for TVWS RA ......................... 104
4.3.5 Summary ........................................................ 110
4.4 Autonomous Rate Control for More Categories of Vehicles.......... 110
4.4.1 System Model................................................... 111
4.4.2 Problem Formulation and DRL Based RC .................... 113
4.4.4 Summary ........................................................ 121
4.5 Intelligent Rate Control for Internet of Maritime Vehicles............ 122
4.5.1 Related Works .................................................. 124
4.5.2 System Model................................................... 125
4.5.3 Proactive NARXNN Forecaster Based MCS Selection ....... 129
4.5.5 Summary ........................................................ 140
References ..................................................................... 141

Contents ix
5 Intelligent Networking enabled Vehicular Distributed Learning....... 145
5.2 Rateless Coding Enabled Broadcasting for Vehicular
Federated Learning...................................................... 147
5.2.1 System Model................................................... 148
5.2.2 BAP for Vehicular Cooperative Learning ..................... 149
5.2.3 Convergence Analysis .......................................... 150
5.2.5 Summary ........................................................ 156
5.3 Opportunistic Collaborated Learning Over Intelligent
Internet of Vehicles ..................................................... 156
5.3.1 System Model................................................... 157
5.3.2 Opportunistic Collaborated Learning via V2R Interaction ... 158
5.3.3 Convergence Analysis .......................................... 159
5.3.4 Summary ........................................................ 161
References ..................................................................... 161
6 Conclusions and Future Workers .......................................... 163
6.1 Conclusions ............................................................. 163
6.2 Future Directions........................................................ 164

Acronyms
AAA Authenticate, authorization, and accounting
AI Artificial intelligence
AP Access point
AS Aggregation server
BAP Broadcasting enabled asynchronous parallelization
BS Base station
COTS Commercial off-the-shelf
CSI Channel state information
DCF Distributed coordination function
DL Deep learning
DRL Deep reinforcement learning
DSRC Dedicated short-range communication
FEL Federated edge learning
FL Federated learning
IoV Internet of vehicle
ITS Intelligent transportation system
MAC Medium access control
MCS Modulation and coding scheme
ML Machine learning
NLOS Non-line-of-sight
OBU Onboard unit
RA Rate adaptation
RSSI Received signal strength indicator
RSU Roadside unit
SNR Signal-to-noise ratio
TSC Time series classification
TVWS TV white space
xi

xii Acronyms
UAV Unmanned aerial vehicle
V2I Vehicle-to-infrastructure
V2R Vehicle-to-roadside
V2V Vehicle-to-vehicle
V2X Vehicular-to-everything
VANET Vehicular network

Chapter 1
Introduction of Internet Access
of Vehicular Networks
Connected vehicles are changing the modern transportation. Based on the wireless
communication between vehicles with sophisticated radio interfaces, vehicles in
the mobility world can exchange information with neighbors as well as remote
transportation center, which can enable the vehicle to understand both the in-
vehicle status and the road situation. Based on such capability, a lot of smart road
applications can be realized, including the safety-related application, intelligent
transportation system (ITS) and in-vehicle infotainment, etc. The Internet access for
vehicles can further extend the spatial scope and temporal range of the vehicular
communication, which can help all road users to conduct both the long-term
evaluation and short-time response to all situations. In this chapter, we first introduce
the overview of Internet of vehicles (IoV), then we present the Internet access
procedure for a vehicle to connect to a wireless access station that deployed along
the roadside. We then explain the aim of the book, covering the topic of Internet
access performance evaluation, data traffic offloading, Internet link management
and intelligent machine learning (ML) paradigm over IoV.
1.1 Internet of Vehicles Overview
To provide the ubiquitous network connection for automobiles, traditional vehicular
network, which are supported with the Vehicle-to-Everything (V2X) communi-
cation technology, are evolving to the Internet of Vehicles (IoV), which greatly
expand the time scope and the spatial range of traditional vehicular networks.
Literature status quo show that great progress have been made toward the robust
and high performance networking between vehicles and vehicles, vehicles and
roadside unit. However, the emerging automotive applications, such as autonomous
vehicles, intelligent transportation system, etc., have raised new requirements
and challenges for future connection vehicles [1]. To satisfy the need of future
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
W. Xu et al., Internet Access in Vehicular Networks,
https://doi.org/10.1007/978-3-030-88991-3_1
1

2 1 Introduction of Internet Access of Vehicular Networks
automotive applications, IoV need to focus on more aspects of the road services
that beyond the traditional VANET’s capabilities, which mostly deals with the road
safety issues. The original V2X technology is called the Dedicated Short-Range
Communications (DSRC) that developed over the IEEE 802.11p protocol revised
from the original IEEE 802.11a standard. The idea is to migrate the success of the
WiFi network, i.e., take advantage of its simplicity, high throughput of the IEEE
802.11 based networking to enable basic vehicular wireless connections to support
the safety message exchanging. Since the publication of the DSRC in 1999, various
V2X technologies have been proposed, which can be divided into two categories,
i.e., the IEEE 802.11 V2X and the cellular based V2X (C-V2X).
As shown in Fig. 1.1, IoV is urged to provide widely and real-time network
access for emerging automotive applications, including cooperative automatic
driving, intelligent traffic control, and collaborative environmental perception, etc.,
which require extensive data transmission and information exchange that beyond
the capability of conventional connected vehicles. To better assist future road
applications, IoV has to deal with the challenges such as dynamic topological
variations, extended network scale, interrupted radio connection, spectrum scarcity
and computation resource limitation, etc. We focus on the Internet connection
Fig. 1.1 The illustration of Internet of Vehicles

1.1 Internet of Vehicles Overview 3
Fig. 1.2 802.11 V2X
spectrum access types
Table 1.1 802.11 V2X comparison
802.11 Access Link capacity
V2X type spectrum Coverage Penetration (approx.) Cost
DSRC Dedicated <1 km Poor Less than
10 Mbps
High
WiFi Unlicensed 100–300 m Poor Up to 1 Gbps
(802.11ac)
Low
TVWS Cognitive Around 10 km Good 10 Mbps High
between vehicle users all road users, to verify the capability to support the
transmission, computing, storage of the big IoV data.
There are two camps of Internet access technology, i.e., IEEE 802.11 IoV and
cellular IoV. Compared with cellular IoV, the IEEE 802.11 IoV has attracted wide
attentions due to its great advantages, i.e., to bring the success of WiFi networks
to vehicular condition. There are three types of IEEE 802.11 IoV branches based
on the way of accessing to the spectrum resource, as shown in Fig. 1.2, i.e., DSRC,
WiFi and TVWS, whose detailed comparisons are given in Table 1.1.
1.1.1 DSRC
DSRC is the initial standard for vehicular networks that stands for the Dedicated
Short Range Communications published in 1999. The standard is an amended
version from the original WiFi protocol, i.e., the IEEE 802.11a standard. By revising
the PHY layer parameters and fix the backoff window to prevent the excessive
channel access delay. The revision is called IEEE 802.11p, which works at the
5.9 GHz band and only uses half of the IEEE 802.11a’s bandwidth to support
same-level modulation and coding scheme [2]. The DSRC protocol works at the
ad-hoc mode, i.e., vehicles are the ad-hoc node that can communicate with each
other directly via the V2V channel, and vehicles can also directly communicate
with the roadside units via the V2R channel. DSRC is mandatory in north America

for new manufactured vehicles after 2016. There are a lot of safety applications
developed based on DSRC, including collision alert, emergency event informing,
etc. [3, 4]. Both of the single-hop and multi-hop DSRC connections are utilized
to enable ITS application to improve road efficiency [5]. Due to the limited
bandwidth and communication range, DSRC cannot always support the high data
rate transmission, and the connection is often interrupted due to the non-line-of-
sight (NLOS) condition and highly dynamic network topology [6]. The access to
DSRC channel would suffer significant delay due to the inefficiency of MAC layer
access in dense conditions. Recent studies show that DSRC can support end-to-end
transmission of 50–100 ms latency, which can support current safety applications
well such as brake alert, and traffic efficiency enhancement. However, for some
advanced vehicular applications such as remote driving, DSRC may be not feasible
as they require more stringent transmission latency [7] (Table 1.1).
1.1.2 ISM Band WiFi with Opportunistic Access
Directly applying the ISM band WiFi to vehicular conditions is expected to
continue their success on road due to its unlicensed spectrum advantage and the
high performance. The utility of WiFi networks has been well proved by both
measurement and analysis [8, 9]. In 2004, Ott et al. setup a 802.11b at 2.4 GHz
WiFi AP on roadside to provide Internet access for the drive-by vehicles, which is
referred to as the ‘drive-thru Internet’. It is shown that considerable throughput can
be achieved for both UDP and TCP traffic. Such paradigm has been used for many
data applications, e.g., vehicle data offloading [10, 11], content caching [12, 13],
data delivery [14], etc. Due to the limited coverage range, the connection to roadside
WiFi networks would be interrupted. Cheng et al. analyzed the trade-off between
the data task fulfillment delay and the offloading efficiency by opportunistically
transmitting to a series of WiFi APs along the road when drive through the coverage
areas [15]. Beside such opportunistic V2R communications, WiFi is also applied to
enable the V2V communications, such as V2V content sharing [16].
The advantage is that WiFi is always evolving. First, the new generation
WiFi network would always introduce new technologies to improve the link rate,
e.g., channel binding, massive MIMO, higher order modulation, etc. [17]. For
example, the 802.11ac WiFi is measured in both V2V and V2I communication
and outperforms the legacy 802.11n protocol [18]. In addition to the data plane
improvement, new control functions are introduced to the next generation WiFi.
To enable smooth AP switch when the vehicle drives through a series of roadside
WiFi networks, 802.11r is issued to reduce the number of management frames
during the handover process [19]. The new hotspot 2.0 specification has enabled
the Authenticate, Authorization, Accounting (AAA) functions to WiFi and access
framework including the automatic association, secure communication, and better
interworking with backhaul networks [20]. Furthermore, the cost of deploying WiFi
networks is low. It is agile to build a roadside AP based on inexpensive hardware

1.1 Internet of Vehicles Overview 5
and open source software. Thus, it is even possible to deploy an array of WiFi
transceivers along the roadside and take advantage of the path diversity. Experiment
in [21] has shown that the link performance can be greatly improved for both UDP
and TCP traffic. To apply WiFi technology to vehicular networks, there would be
some issues cause of the high mobility. However, consider economical efficiency
and network performance, the advantages apparently outweigh disadvantages, not
to mention that future WiFi networks are keeping improving in terms of higher link
rate, mobility support, and roaming ability, etc.
1.1.3 TVWS with Cognitive Spectrum Access
Both of the DSRC and WiFi networks have poor propagation range and penetration
capability due to the high carrier frequency. To avoid the frequent handover and
further improve the network bandwidth, the vacant TV spectrum between 470 to
790 MHz are re-utilized as they are often no longer used when Internet becomes
the main source of information and entertainment. In 2014, IEEE published the
802.11af standard to support sharing the TVWS spectrum for cognitive secondary
users [22]. To investigate the feasibility and check the efficiency of vehicular access
to TVWS, Zhou et al. applied the 802.11af TV access system for both V2V and
V2I communication guided by spectrum usage according to a geolocation database
[23, 24]. The field measurements in [25] show that 802.11af transceivers can achieve
beyond 9 Mbps over 6 km. There still remains critical issues for vehicular TVWS
access. First is that the secondary users could not access if primary users occupy the
TV spectrum, which means that the usability of TV band is not guaranteed. Since
the coverage area is much bigger compared with the normal WiFi networks, number
of co-associated vehicle users would likely be large, which leads to intensive media
access contention and severe congestion [26]. Moreover, the cost of setting up a
TVWS not only requires expensive investments, but also the spectrum permission
from regulation office, which further limits its usage.
1.1.4 Cellular IoV
Cellular networks can provide ubiquitous coverage, seamless handover, high band-
width access, reliable and secure message transmission, and thus has great potential
to support enormous vehicular communication applications and services. Cellular
IoV standardization can be divided into three phases. At the first stage, LTE
network is used to support basic V2X applications since 2015 within 3GPP Release
14 [27]. 3GPP defines 27 typical use cases encompassing V2V, V2I, V2P, and
V2N applications, and the requirements for 7 typical scenarios. It has confirmed
to adopt the PC5 interface and Uu interface based LTE network to support IoV
services, and enhancement from physical layer structure, resource selection and

allocation, and synchronization have been developed by the 3GPP RAN work
group (WG). In Rel-14, V2X is mostly to provide data transport service for basic
road safety service such as cooperative awareness messages (CAM), basic safety
message (BSM), or decentralized environmental notification messages (DENM),
and so on. Enhancement has been involved in Rel-15 [28] to support advanced
IoV scenarios [29], such as vehicle platooning, advanced driving, extended sensors,
remote driving, etc. In Rel-15, the WI “V2X phase 2 based on LTE” introduces
some key functionalities to support advanced IoV services in a fully backward
compatible manner with Rel-14 V2X, including carrier aggregation (CA) for mode-
4, higher order modulation (e.g., 64-QAM), radio resource pool sharing between
mode-3 and mode-4 UEs, shorten transmission time interval (TTI), and reduction
of the maximum time between packet arrival at Layer 1–10 ms (which was 20 ms
in Rel-14) and resource selected for transmission. 3GPP has also started 5G NR-
V2X recently [30], to evaluate the enhanced V2X services by defining simulation
scenarios, performance metrics, channel modeling, spectrum, etc. Chinese telecom
companies, such as Datang and Huawei, are among the main force in the 3GPP V2X
standardization and development of several LTE-V2X applications [31]. From 2016,
the newly established 5G Automotive Association (5GAA) and the Next Generation
Mobile Networks Alliance (NGMN) [32] have developed V2X solutions to support
connected cars and road safety applications.
1.1.5 Summary
Both of IEEE 802.11 IoV and cellular IoV will play an important role in future
vehicular applications. We mainly consider the low-cost, high-throughput Internet
access via the unlicensed 802.11 networks in this monograph. To provide Internet
connection to vehicles, a vehicle should finish the Internet access procedure, to find
the AP and get correct configurations to enable effective Internet connection [24].
1.2 Internet Access Procedure
To utilize current IoV radio interface to access to Internet, the access procedure
include three main steps that involve the corresponding main functions, i.e., network
detection, user authentication and network parameters configuration.
1.2.1 Network Detection
Traditional WiFi use beacon frames to exchange the information between the user
and AP. The beacon scheme includes two mechanisms, i.e., passive beaconing and

1.2 Internet Access Procedure 7
Table 1.2 IEEE 802.11 Beacon frames
Type Tag Usage
Link parameters Support rates Physical layer rate/modulation type
Channel status Signal and noise level, spectrum, etc.
802.11 radio RTX frequency index, current data rate, timestamp, etc.
DS-status Mode (ad-hoc/BSS),
HT info 20 MHz/40 MHz bandwidth
Network info QBSS load Channel utilization, admission capacity, etc.
SSID Network name
Interworking Network type and hotpsot 2.0 support
RSN info Authentication method
EAP method Type of EAP
proactive beaconing. In passive beaconing, the AP broadcasts the beacon frames
to nearby users, including the information of link parameters, network ID, etc.,
which are summarized in Table 1.2 [33]. In proactive mode, user broadcast a probe
request frame to search nearby APs, which will then reply the required information
in the prove reply frames. Before access to the roadside WiFi network, the vehicle
should detect the existence of the network via beacon frames exchange or other
query protocols (e.g., the Access Network Query Protocol (ANQP) of Hotspot 2.0)
[34]. This step is essential for vehicles to get the information about the wireless link
parameters, e.g., 802.11 radio channel, supported rates, SSID, etc., and the backhaul
information like authentication method, current load, etc. The information can also
help the vehicle to select a proper nearby AP.
However, sometimes such information is not sufficient for clients to find a proper
AP to associate. First, the clients have limited information about the backhaul
network connected to the AP, e.g., Internet accessibility, the security level, the QoS
mapping support. Secondly, according to the default AP selection policy, the WiFi
clients always choose the one with the largest Received Signal Strength Indicator
(RSSI) among all available nearby APs. Such AP selection policy may result in
improper association, e.g., association to APs without Internet connectivity, or
causing unbalance load distribution and low utilization of available APs [35]. In
order to overcome these problems, hotspot 2.0 are specified that no longer rely on
the SSID solely to identify a WiFi network [36]. Prior to AP association, the WiFi
clients can obtain more information such as the Network Access Identifier (NAI),
the operator information via multiple ways. First, new information elements are
added into the beacon and probe frames, so that the client devices can obtain more
information about the surrounding WiFi networks by listening to beacon frames
or requesting probe response frames. The key added information elements and the
purpose are listed in Table 1.3. Besides, a new query protocol called ANQP is
specified for the clients to obtain further information about the AP or the backhaul
network services [37]. Some of the important elements are also listed in Table 1.3.

Table 1.3 Important information elements in beacon, probe and ANQP frames in Hotspot 2.0
Parameters Sub-field Purpose
Extended capabilities Interworking Indicates if this WiFi network can interwork
with other networks
QoS traffic capability Indicates if the WiFi network can support QoS
mapping between WiFi and external networks
Interworking adv. Access network type Indicates if the network is private or public, is
connected to Internet
Advertisement
protocol ID
Indicate the query protocol ID and the
response length limit
Roaming consortium N/A Indicate the roaming consortium whose
credential can also authenticate with the
current AP
ANQP element WAN metrics Information about Internet connecting,
downlink and uplink speed/load, etc.
NAI Realm data Information about NAI realm name and
authentication method
Domain name Domain information about the operators
1.2.2 Authentication
There are several authentication mechanisms which are applied in different sce-
narios. Webpage/SMS verification are used in some public places such as airports,
malls, etc. However, such mechanism often requires user’s manual interaction to
input the verification code, which is not feasible for vehicle users, who would prefer
the automatic authentication methods. In residential WiFi networks, WPA2-PSK
are often used, where a pre-shared credential is stored in WiFi client devices, which
will automatically perform the authentication process with the AP. However, the
pre-shared credential scheme has some limitations. First, the pre-shared credential
is stored at local system of the AP, which lacks of a credential management entity
that a vehicle user has to negotiate with the AP when need to update the credentials.
WPA2-802.1X is used in commercial WiFi access or enterprises, e.g., eduroam
[38, 39], which support remote management via the authentication server. The
vehicle users are assigned with a certification to better protect the user credentials.
However, the handshake with the remote server requires more management frames
to be exchanged and extra backhaul communication delay is introduced. Hotspot
2.0 also employs the 802.1X protocol in the authentication method to provide
users secure connection. The latest released version also provisions Online SignUp
(OSU) ability which can enable users to automatically select proper plan with
reasonable costs from service providers. This kind of management framework
makes the WiFi networks as easy and secure as cellular networks. WiFi operators
use the authentication step to identify qualified users and prevent unauthorized
users from stealing the network resources, while WiFi users rely on this step to
protect their communication privacy. There are several authentication mechanisms

1.2 Internet Access Procedure 9
which are applied in different scenarios. Commonly, webpage/SMS verifications
are used in some public places such as airports, malls, etc. And WPA2-PSK are
often used in residential areas. WPA2-802.1X is used in commercial WiFi access
or enterprises, e.g., eduroam [38]. From the measurement results in [8], it can be
observed that the delay caused by the access procedure is mainly from the user
authentication step, which takes up the majority part of the management frames
and also introduce possible negotiation overhead with remote authentication server.
Sophisticated authentication schemes usually require more management frames
to be exchanged between the WiFi client and the authentication server, which
consumes longer time and thus the overall throughput that the vehicle can achieve
would be reduced.
Literature works also investigated efficient authentication methods in vehicular
conditions. Chen et al. developed an authentication prototype to support automatic
user authentication and seamless interworking between WiFi and WiMAX [40].
Han et al. analyzed the security of connected vehicles and proposed a verification
method to protect authenticated users [41]. Fu et al. presented a fast authentication
method in heterogeneous WiFi and WiMAX networks to reduce the handover
delay [42]. Bohak et al. proposed a fast authentication mechanism to reduce the
round trip time between the WiFi user and the remote authentication server [43].
However, such scheme requires to distribute the access information to potential
APs, which is difficult in practical deployment. Moustafa et al. applied the 802.11i
protocol to setup reliable data transfer in high way condition [44]. These works
show the importance to find out the overhead of the authentication step for vehicle
users, which can provide guidance for future authentication scheme research and
development.
1.2.3 Network Parameters Assignment
To enable effective Internet connection, the vehicle should set proper network
parameters, such as valid IP address, correct VLAN configuration, etc. DHCP
protocol is often used to dynamically assign the IP address for a WiFi client via
a local or remote DHCP server from an address pool. And the Automatic Private
IP Address (APIPA) protocol is often used to auto configure the IP address if the
DCHP server is not available.
1.2.4 Summary
In this section, we have reviewed the three main steps for a vehicle user to access
to a roadside access station. It is inevitable and thus requires carefully scrutiny over
the whole access procedure and the impact to the Internet access performance.

1.3 Aim of the Book
In this book, we aim to introduce the Internet access in vehicular networks,
considering the practical Internet connection for vehicles via the roadside access
points (APs), which normally involve a set of the management frames for control
functions. We first propose analytical methods to evaluate the Internet access
performance, i.e., the access delay and throughput in Chap. 2. We also consider
using the unlicensed spectrum to offload the data traffic from expensive cellular
networks to economic IEEE 802.11 IoV utilizing the V2X interworking scheme in
Chap. 3. Then we take a deep look at the V2R link management that can adjust
the modulation and coding scheme (MCS) smartly to adapt to the highly dynamic
channel conditions 4. In addition, in Chap. 5 we employ distributed computing
paradigms based on the IoV connectivity for vehicular users, to cooperatively
train machine learning (ML) models for intelligent vehicular applications. The
conclusion and future directions are given in Chap. 6
References
1. W. Xu, H. Zhou, N. Cheng, F. Lyu, W. Shi, J. Chen, X. Shen, Internet of vehicles in big data
era. IEEE/CAA J. Autom. Sin. 5(1), 19–35 (2018)
2. D. Jiang, L. Delgrossi, IEEE 802.11p: towards an international standard for wireless access in
vehicular environments, in VTC Spring 2008-IEEE Vehicular Technology Conference (IEEE,
Piscataway, 2008), pp. 2036–2040
3. D. Jiang, V. Taliwal, A. Meier, W. Holfelder, R. Herrtwich, Design of 5.9 GHZ DSRC-based
vehicular safety communication. IEEE Wirel. Commun. 13(5), 36–43 (2006)
4. Q. Xu, T. Mak, J. Ko, R. Sengupta, Vehicle-to-vehicle safety messaging in DSRC, in
Proceedings of the 1st ACM International Workshop on Vehicular ad Hoc Networks (ACM,
New York, 2004), pp. 19–28
5. Y.L. Morgan, Notes on DSRC & WAVE standards suite: its architecture, design, and charac-
teristics. IEEE Commun. Surv. Tuts. 12(4), 504–518 (2010)
6. F. Lyu, H. Zhu, H. Zhou, L. Qian, W. Xu, M. Li, X. Shen, MoMAC: mobility-aware and
collision-avoidance MAC for safety applications in VANETs. IEEE Trans. Veh. Technol.
67(11), 10590–10602 (2018)
7. G. Naik, B. Choudhury, J. Park, IEEE 802.11bd & 5G NR V2X: evolution of radio access
technologies for V2X communications. IEEE Access 7, 70169–70184 (2019)
8. W. Xu, H. A. Omar, W. Zhuang, X. Shen, Delay analysis of in-vehicle internet access via
on-road WiFi access points. IEEE Access 5, 2736–2746 (2017)
9. W. Xu, W. Shi, F. Lyu, H. Zhou, N. Cheng, X. Shen, Throughput analysis of vehicular internet
access via roadside WiFi hotspot. IEEE Trans. Veh. Technol. 68(4), 3980–3991 (2019)
10. N. Cheng, N. Lu, N. Zhang, X. Zhang, X. Shen, J.W. Mark, Opportunistic wifi offloading
in vehicular environment: A game-theory approach. IEEE Trans. Intell. Transp. Syst. 17(7),
1944–1955 (2016)
11. Y. Chen, N. Zhang, Y. Zhang, X. Chen, W. Wu, X.S. Shen, Energy efficient dynamic offloading
in mobile edge computing for internet of things. IEEE Trans. Cloud Comput. 9(3), 1050–1060
(2021)
12. H. Wu, W. Xu, J. Chen, L. Wang, X. Shen, Matching-based content caching in heterogeneous
vehicular networks, in Proc. IEEE Global Communications Conference (GLOBECOM) (IEEE,
Piscataway, 2018), pp. 1–6

References 11
13. H. Zhou, N. Cheng, J. Wang, J. Chen, Q. Yu, X. Shen, Toward dynamic link utilization for
efficient vehicular edge content distribution. IEEE Trans. Veh. Technol. 68(9), 8301–8313
(2019)
14. Z. Su, Q. Xu, Y. Hui, M. Wen, S. Guo, A game theoretic approach to parked vehicle assisted
content delivery in vehicular ad hoc networks. IEEE Trans. Veh. Technol. 66(7), 6461–6474
(2017)
15. N. Cheng, N. Lu, N. Zhang, X. Shen, J.W. Mark, Opportunistic wifi offloading in vehicular
environment: A queueing analysis, in IEEE Global Communications Conference (GLOBE-
COM) (2014), pp. 211–216
16. H. Viittala, S. Soderi, J. Saloranta, M. Hamalainen, J. Iinatti, An experimental evaluation
of WiFi-based vehicle-to-vehicle (V2V) communication in a tunnel, in IEEE 77th Vehicular
Technology Conference (VTC Spring) (IEEE, Piscataway, 2013), pp. 1–5
17. H.A. Omar, K. Abboud, N. Cheng, K.R. Malekshan, A.T. Gamage, W. Zhuang, A survey
on high efficiency wireless local area networks: next generation WiFi. IEEE Commun. Surv.
Tutorials 18(4), 2315–2344 (2016)
18. V.P. Sarvade, S. Kulkarni, Performance analysis of IEEE 802.11 AC for vehicular networks
using realistic traffic scenarios, in IEEE International Conference on Advances in Computing,
Communications and Informatics (ICACCI) (IEEE, Piscataway, 2017), pp. 137–141
19. M.I. Sanchez, A. Boukerche, On IEEE 802.11 K/R/V amendments: do they have a real impact?
IEEE Wirel. Commun. 23(1), 48–55 (2016)
20. W. Xu, H. Zhou, Y. Bi, N. Cheng, X. Shen, L. Thanayankizil, F. Bai, Exploiting hotspot-2.0
for traffic offloading in mobile networks. IEEE Netw. 32(5), 131–137 (2018)
21. Z. Song, L. Shangguan, K. Jamieson, WiFi goes to town: rapid picocell switching for wireless
transit networks, in Proc. ACM SigCom (2017), pp. 322–334
22. A.B. Flores, R.E. Guerra, E.W. Knightly, P. Ecclesine, S. Pandey, IEEE 802.11 AF: a standard
for TV white space spectrum sharing. IEEE Commun. Mag. 51(10), 92–100 (2013).
23. H. Zhou, N. Zhang, Y. Bi, Q. Yu, X. Shen, D. Shan, F. Bai, TV white space enabled connected
vehicle networks: challenges and solutions. IEEE Netw. 31(3), 6–13 (2017)
24. H. Zhou, N. Cheng, Q. Yu, X. Shen, D. Shan, F. Bai, Toward multi-radio vehicular data piping
for dynamic DSRC/TVWS spectrum sharing. IEEE J. Sel. Areas Commun. 34(10), 2575–2588
(2016)
25. K. Ishizu, K. Hasegawa, K. Mizutani, H. Sawada, K. Yanagisawa, T. Keat-Beng, T. Matsumura,
S. Sasaki, M. Asano, H. Murakami et al., Field experiment of long-distance broadband
communications in TV white space using IEEE 802.22 and IEEE 802.11 AF, in IEEE
International Symposium on Wireless Personal Multimedia Communications (WPMC) (IEEE,
Piscataway, 2014), pp. 468–473
26. R. Almesaeed, N.F. Abdullah, A. Doufexi, A.R. Nix, Performance evaluation of 802.11
standards operating in TVWS and higher frequencies under realistic conditions, in IEEE 80th
Vehicular Technology Conference (VTC2014-Fall) (2014), pp. 1–5
27. TR 21.914, 3GPP TR 21.914 version 14.0.0 Release 14’
28. TR 21.915, 3GPP TR 21.915 version 1.1.0 Release 15
29. 3GPP TR 22.886, Study on enhancement of 3GPP support for 5G V2X services
30. 3GPP TR 37.885, Study on evaluation methodology of new Vehicle-to-Everything V2X use
cases for LTE and NR
31. S. Chen, J. Hu, Y. Shi, L. Zhao, LTE-V: A TD-LTE-based V2X solution for future vehicular
network. IEEE Internet Things J. 3(6), 997–1005 (2016)
32. 5G Automotive Association, The case for cellular V2X for safety and cooperative driving, in
5GAA Whitepaper (2016)
33. S. Vasudevan, K. Papagiannaki, C. Diot, J. Kurose, D. Towsley, Facilitating access point
selection in IEEE 802.11 wireless networks, in Proceedings of the 5th ACM SIGCOMM
Conference on Internet Measurement (2005), pp. 26–26
34. Hotspot 2.0 technical task group, Wi-Fi Alliance Technical Committee
35. W. Xu, C. Hua, A. Huang, A game theoretical approach for load balancing user association in
802.11 wireless networks, in Proc. IEEE Globecom’10 (2010), pp. 1–5

36. Hotspot 2.0 specification and passpoint project. http://www.wi-fi.org/discover-wi-fi/wi-fi-
certified-passpoint
37. W. Xu, H. Zhou, Y. Bi, N. Cheng, X. Shen, L. Thanayankizil, F. Bai, Exploiting hotspot-2.0
for traffic offloading in mobile networks. IEEE Netw. 99, 1–7 (2018)
38. L. Florio, K. Wierenga, Eduroam, providing mobility for roaming users, in Proceedings of the
EUNIS 2005 Conference, Manchester (2005)
39. Eduroam: secure, world-wide roaming access service for international research and education
community. https://www.eduroam.org/
40. Y.-T. Chen, Achieve user authentication and seamless connectivity on WiFi and WiMAX
interworked wireless city, in IFIP International Conference on Wireless and Optical Com-
munications Networks (2007), pp. 1–5
41. K. Han, S.D. Potluri, K.G. Shin, On authentication in a connected vehicle: secure integration
of mobile devices with vehicular networks, in 2013 ACM/IEEE International Conference on
Cyber-Physical Systems (ICCPS) (2013), pp. 160–169
42. A. Fu, G. Zhang, Z. Zhu, Y. Zhang, Fast and secure handover authentication scheme based on
ticket for WiMAX and WiFi heterogeneous networks. Wirel. Pers. Commun. 79(2), 1277–1299
(2014)
43. A. Bohák, L. Buttyán, L. Dóra, An authentication scheme for fast handover between wifi access
points, in Proc. of ACM Wireless Internet Conference (WICON) (2007)
44. H. Moustafa, G. Bourdon, Y. Gourhant, Providing authentication and access control in vehic-
ular network environment, in IFIP International Information Security Conference (Springer,
Berlin, 2006), pp. 62–73

Chapter 2
Internet Access Modeling for Vehicular
Connection
In this chapter, we focus on the analytical modeling of the Internet access procedure
for vehicles. Specifically, the Markov chain model is applied to describe the
management frame exchange between the vehicle and the roadside access point.
Due to the non-negligible overhead, the access delay is analyzed, which can
determine the overall data throughput that can be achieved by the drive-by vehicle.
Such access delay and throughput performance is crucial for future IoV network
protocol design. We have demonstrated the accuracy of our analysis via both
simulation and experimental verification methods.
2.1 Background and Motivation
Due to the ever growing IoV big data, it is expected to support data-rich and
bandwidth consuming Internet applications, e.g., in-vehicle infotainment, remote
driving, etc. These applications are often deployed onboard to envision immersive
experience for both the drivers and the passengers [1, 2]. For example, Intelligent
Transportation (ITS) system can collect and disseminate the vehicles’ internal and
external conditions via the Vehicle-to-Infrastructure (V2I) connections to improve
the road efficiency and driving safety level [3, 4]. Besides, with the Internet access,
a myriad of infotainment applications, such as video streaming, web page surfing,
etc., are becoming indispensable for passengers. Furthermore, some data-craving
applications, such as High Definition (HD) map, autonomous driving, etc., is
expected to be realized via the high bandwidth connection. It is predicted that
the global vehicular data traffic will reach 300 Zettabytes by 2020 [5], which
can cause a great pressure to current Internet access technologies for vehicles [6].
Cellular networks are initially adopted for Internet access for vehicles. However,
the costs of downloading/uploading all traffic to cellular networks are usually not
affordable for vehicle users. In addition, cellular network capacity will be drained
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021
W. Xu et al., Internet Access in Vehicular Networks,
https://doi.org/10.1007/978-3-030-88991-3_2
13

14 2 Internet Access Modeling for Vehicular Connection
up in dense condition where lots of vehicles are requesting heavy data tasks [7]. To
overcome the drawbacks of cellular Internet access, different wireless technologies
have been proposed to provide alternate choices. Zhou et al. used the TV white
space (TVWS) spectrum enabled infostation to disseminate the multimedia content
[8]. However, the adoption of TVWS is restricted by the geo-location where the
regulation and policy of using TVWS spectrum varies place by place. Wu et al.
adopted the heterogeneous small cells to offload the cellular traffic, however, it
requires frequent vertical and horizontal handoff for vehicle users due to their
high mobility [9]. Ligo et al. utilized the Dedicated Short Range Communications
(DSRC) to offload the vehicular traffic to Internet [10]. However, the link rate is
limited as the DSRC bandwidth is only half of 802.11a. Luo et al. investigated the
inter-vehicle performance based on the visible light communication (VLC) [11],
which is greatly affected by the day light noise and line-of-sight condition and the
network deployment is difficult.
WiFi can overcome the restriction of the above radio technologies. First, WiFi
has been widely used for Internet access around the world for years. It is predicted
that by the year of 2021, 73% of the global Internet traffic will be served by
WiFi networks [12]. WiFi devices are universally compatible and the unlicensed
spectrum is used that are not restricted over all regions of the world. Compared with
licensed spectrum based technologies, e.g., LTE-V2X, WiFi device does not need a
permission, and different generation WiFi can communicate with each other without
upgrading their devices. Secondly, WiFi has significant link throughput. The latest
802.11ac protocol can achieve the peak link rate around 1 Gbps [13], which provides
enough capacity for vehicles in a WiFi cell even in dense condition. Furthermore,
unlike deploying TVWS station or consuming in cellular networks, the economic
cost of operating WiFi networks are relatively low. A roadside WiFi network can
be agilely setup using commercial off-the-shelf devices and open source software,
which is much cheaper than building infrastructures for macro-cells, e.g., LTE-V2X
base stations [14].
There have been extensive research works on roadside WiFi Internet access for
vehicles. Ott et al. first proposed the concept of ‘Drive-thru Internet’ that utilize a
802.11b hotspot to provide temporal Internet access for drive-by vehicles [15]. The
conducted road test showed that considerable data traffic can be transmitted between
the roadside hotspot and the vehicle, which demonstrated that it is feasible to
provide Internet access for vehicles by WiFi technologies. Mahajan et al. measured
the end-to-end connectivity between moving vehicles and the roadside WiFi Access
Points (APs). And the throughput performance between the AP and the moving
vehicles are also investigated on different regions [16]. Cheng et al. adopted the
queueing model to analyze the traffic offloading performance for vehicles using
the intermittent roadside WiFi networks. The relationship between the offloading
effectiveness and the average service delay of the data tasks are analyzed [17].
Similarly, Zhou et al. proposed a cluster based scheme to conduct cooperation
between multiple roadside APs to deliver content to vehicles [18]. However, existing
works seldom consider the access procedure that a vehicle user has to accomplish

2.1 Background and Motivation 15
before he/she can actually access to Internet via a roadside hotspot [19]. The access
procedure generally contains the three steps as mention in Sect. 1.2.
(1) Network Detection: Before access to the roadside WiFi network, the vehicle
should perceive the existence of the network via beacon frames exchange or
other query protocols (e.g., the Access Network Query Protocol (ANQP) of
Hotspot 2.0) [20].
(2) User Authentication: It is required to setup reliable and secure wireless connec-
tions between the AP and its users.
(3) Network Parameters Assignment: To communicate with other entities on
Internet, it is required that the associated WiFi users to have an IP address.
Dynamic Host Configuration Protocol (DHCP) protocol is often used to assign
the IP address for WiFi users dynamically. In some conditions, the vehicle
might need to configure extra network settings.1 Such steps require a number of
management frames to be exchanged between the vehicle, hotspot and remote
servers.
In practice, the above steps are necessary to setup effective and reliable Internet
connection for vehicles. However, most previous research works and conducted
experiments have neglected these steps. In [15], Ott et al. used an open WiFi hotspot
without verifying user credential before allowing network access, i.e., no user
authentication step is included. And a static IP address was assigned to the vehicle
who could access to the network immediately when the vehicle drives over. The
experiment in [21] did not include the authentication step, while the measurement
result showed that the DHCP latency could be several seconds. Mahajan et al.
also did not consider the overhead of authentication and IP address acquisition in
their measurements in [16]. The analytical works from [17, 18, 22, 23] assume that
Internet can be accessed as soon as the vehicle drives into the cell coverage areas,
where the impacts of the access procedure were omitted.
A vehicle cannot access to Internet via the roadside WiFi AP until the access
procedure is accomplished. Since the sojourn time of a vehicle within the WiFi
coverage area is limited, a fast access procedure will leave more time for down-
loading/uploading Internet data, and vice versa. The accomplishment of the access
procedure can be deferred in the following conditions. First, when the number of
other WiFi clients, e.g., other vehicles, are associated to the same AP increases,
the finish of the procedure will be deferred as the exchange of all the man-
agement frames requires more time when contending channel resource with its
peers. Secondly, when the channel quality degrades, each management frame may
require more re-transmission attempts as the packet error rate increase. Besides,
the parameters of the IEEE 802.11 distributed coordination function (DCF), e.g.,
minimum window size, back off stage number, may also change the DCF process for
transmitting each management frame [24], which lead to different consequence of a
transmission attempt, and thus affect the accomplishment of the access procedure.
1 For example, apply VLAN settings to divide the broadcast domain of several sub-nets.

Furthermore, different authentication methods require the vehicle to exchange
different set of management frames with the AP and authentication server, and thus
the steps of user authentication are different.
In this chapter, we investigate the performance of drive-thru Internet considering
the accomplishment of the practical access procedure. Particularly, our objective
is to find the relationship between the access delay and throughput via a roadside
AP, and environmental and protocol execution conditions such as the channel
conditions, contention level, network protocol configurations, etc. by considering
the overhead of the Internet access procedure. The overhead of network detection
via query or beaconing, and network parameters assignment were not considered
in status quo literature. Shin et al. utilized a selective channel scanning method
to shorten the network detection delay and thus the access overhead can be
reduced [25]. However, the authentication step which takes the majority part of the
access delay was not considered. In fact, to eliminate the impacts of the access
procedure, which are difficult to evaluate, most experiment and measurement of the
WiFi connection between vehicles and APs in literature works applied the open
association [15, 26], static IP address setting [16, 27] and without authentication
step [21, 28]. Lu et al. argued that the time for access procedure cannot be neglected
due to high mobility of vehicles, which can take up to ten or more seconds [1]. Our
purpose is to analyze that in a practical scenario where the three steps of the access
procedure cannot be waived, what is the Internet access performance of the IoV
connection.
2.2 Delay Analysis of Vehicular Internet Access
Limited existing works focus the time duration that a vehicle user needs to take
before the user can access the service of an on-road WiFi AP and actually connect
to the Internet [29]. This time duration, referred to as ‘access delay’, is required
mainly to perform the authentication and Internet Protocol (IP) address assignment.
In [15], the conducted drive-thru experiments employ an open access scheme and
a static IP address, which allow a vehicle user to automatically associate and
access the AP service, without any consideration of the access delay. Yet, except
for experimental testing or research purposes, the access delay is unavoidable to
perform the authentication procedure, which is essential for WiFi network users and
operators.
We consider the WPA2 and Hotspot 2.0 authentication methods that can be auto-
matically accessed without any manual interaction. In addition to the time duration
required to complete the authentication procedure, another duration is needed for a
vehicle user to obtain an IP address, e.g., via Dynamic Host Configuration Protocol
(DHCP) protocol. The sum of the durations required for authentication and IP
assignment constitutes the access delay, which can last for a few seconds [21]. In
such a case, a vehicle user can have a limited time to utilize the Internet resources
before the vehicle moves out of the coverage area of a WiFi AP, especially with a

2.2 Delay Analysis of Vehicular Internet Access 17
high vehicle moving speed. Hence, a ‘quickWiFi’ scheme was proposed to reduce
the access delay by tuning related WiFi parameters and optimizing the AP scanning
strategy for clients [26].
The access delay can be affected in several ways. First, if the AP is serving
a large number of users, the access delay will increase for a new user due to
a high level of channel contention using the IEEE 802.11 standard distributed
coordination function (DCF). Second, a poor wireless propagation channel can
result in a high frame error rate, which further increases the access delay, due to
re-transmission of management frames that are not successfully delivered. Third,
different authentication protocols require different sequences of management frame
exchanges between the AP and a new user, leading to a different access delay
associated with each authentication methods. To the best of our knowledge, the
effects of the number of contending WiFi users, the wireless channel conditions, and
the employed authentication method on the access delay have not been analyzed.
We investigate how these factors affect the access delay. We propose a Markov
chain-based analytical model that can be applied for any authentication method,
in order to calculate the average access delay, given the time-varying channel
conditions and number of contending WiFi users in a vehicular environment. The
accuracy of the proposed analytical model is studied via MATLAB simulations and
experimental testing. The experimental testing is conducted using commercial off-
the-shelf (COTS) WiFi products supporting the IEEE 802.11n standard, together
with an advanced channel emulator that emulates the wireless channel conditions
between the vehicles and a WiFi AP in an expressway scenario. The analytical,
simulation, and experimental testing results of the average access delay are obtained
for the WPA2-PSK and WPA2-802.1X authentication methods, under various
wireless channel conditions and for various numbers of contending WiFi users.
2.2.1 System Model
We consider a single WiFi AP that provides Internet connectivity for vehicles
on the road. When a vehicle enters the communication range of the AP, before
connecting to the Internet, the vehicle exchanges a sequence of management frames
with the AP in order to perform the necessary procedures for authentication and
IP address allocation. The management frame exchanges between the vehicle and
the AP depend on the WiFi network access standard, e.g., WPA2 [30] and Hotspot
2.0 [31], and the authentication mechanism, e.g., IEEE 802.1X [32] and extensible
authentication protocol (EAP) [33, 34].
For instance, Figs. 2.1 and 2.2 respectively show the sequence of management
frames exchanged between a vehicle and the AP for the WPA2-PSK and WPA2-
802.1X authentication methods. The generation of some management frames may
require communication between the AP and a remote server through a core
network. For instance, as shown in Fig. 2.2, the AP needs to connect to a remote
authentication, authorization, and accounting (AAA) server before replying to some

Vehicle AP
Authentication request
Authentication reply
Association request
Association response
EAPoL: EAP over local
area network
DHCP request
DHCP acknowledgment
Time Time
EAPoL four-way handshake
52
52
144
140
113
135
209
342
362
Frame length
in bytes 1
2
3
4
5
6
8
10
9
7
Frame index
113
Fig. 2.1 Management frames exchanged between a vehicle and an AP based on the WPA2-PSK
mode for authentication (Nf = 10)
frames from a vehicle. We focus on a single vehicle, referred to as tagged vehicle,
that just enters the communication range of the AP and attempts to connect to the
Internet via the AP. To perform this Internet connection, the management frames
exchanged between the tagged vehicle and the AP, as shown in Figs. 2.1 and 2.2,
are indexed from 1 to Nf , and the length of the ith management frame is denoted
by li, i = 1, .., Nf . The frame length indicates the length of the data field of the
physical layer (PHY) protocol data unit (PPDU), which consists of the encoded
MAC layer protocol data unit (MPDU) and other fields that are included by the
PHY and transmitted over-the-air using the same bit rate as the MPDU, such as the
service field and tail bits added by the IEEE 802.11 orthogonal frequency division
multiplexing (OFDM) PHY standard [35].
In addition to the tagged vehicle, there exist a number of neighbor vehicles that
are already connected to the Internet via the AP and uploading data to the AP.
It is assumed that, each neighbor vehicle always has a data frame to upload to
the AP, from the instant that the tagged vehicle enters the communication range
of the AP until all the Nf management frames are successfully exchanged. Each
data frame uploaded by a neighbor vehicle has a fixed length denoted by l, and is
transmitted at a constant PHY bit rate denoted by r. All the nodes (i.e., the neighbor
vehicles, the tagged vehicle, and the AP) are within the communication range of
each other and employ the IEEE 802.11 DCF to access the channel [35], with a
minimum contention window size denoted by w, and a number of back-off stages
indexed from 0 to m − 1, where m denotes the total number of back-off stages
in the absence of request-to-send/clear-to-send (RTS/CTS) handshaking. At each
back-off stage, the tagged vehicle and the AP employs a PHY bit rate, denoted by
rib, i = 1, . . . , Nf and b = 0, . . . , m − 1, for the next transmission attempt of the
ith management frame that is being exchanged. For the same management frame
index, i, the values of rib ∀b are determined based on a certain rate switching

Vehicle AP
Authentication request
Authentication reply
Association request
Association response
DHCP request
DHCP acknowledgment
Time Time
EAPoL four-way handshake
52
52
108
88
135
135
179
113
342
362
Time
AAA server
EAP: request identity
EAP: response identity
Access request
Access challenge
EAP-PEAP: request
TLSv1 client Hello Access request
Access challenge
TLSv1: server Hello
EAP-PEAP: response
Access request
Access challenge
Server Hello done
Client key exchange Access request
Access challenge
Change cipher spec
EAP-PEAP: response Access request
Access challenge
TLSv1 application data
Access request
Access challenge
Access request
Access challenge
Access request
Access challenge
Access request
Access challenge
EAP success
22
61
61
61
61
61
109
109
77
23
27
24
129
1024
24
126
162
83
24
1
2
3
4
24
26
27
29
28
23
22
21
20
15
16
19
18
17
5
6
7
8
10
11
12
13
14
9
25
Frame length
in bytes
PEAP: protected
EAP
TLSv1: transport
layer security
version 1.0
Frame index
Fig. 2.2 Management frames exchanged between a vehicle and an AP based on the WPA2-802.1X
mode for authentication (Nf = 29)

algorithm, while for the same back-off stage index, b, the value of rib depends
on whether the tagged vehicle or the AP is the source of the ith management
frame. If a management/data frame is successfully received, an acknowledgment
(ACK) frame of length a is transmitted using the same PHY bit rate as that for
the management/data frame transmission. On the contrary, if a management/data
frame is not successfully delivered to its destination, the frame is referred to as a
‘lost’ frame. The ACK timeout duration that the source of a lost frame needs to wait
for, before invoking the DCF back-off procedure, is neglected [35]. A lost frame
is retransmitted by its source node until it is successfully delivered, without any
maximum retry limit.
When the tagged vehicle or the AP attempts to transmit the ith management
frame, i = 1, . . . , Nf , the total number of nodes that are contending to access
the channel is constant and denoted by ni, which consists of all the neighbor
vehicles plus one node (i.e., either the AP or the tagged vehicle, depending on
which one is the source of the ith management frame). For the ni contending
nodes, i = 1, . . . , Nf , let τi denote the probability that a node transmits a frame
in a randomly selected slot duration,2 αi the probability that a transmitted frame
is lost due to a transmission collision, βi the probability that a transmitted frame
is lost due to a poor channel condition (0 < βi < 1), and δi the probability
that a transmitted frame is lost due to a transmission collision or poor channel,
i.e., δi = 1 − (1 − αi)(1 − βi). It is assumed that the value of each of αi, βi,
and (consequently) δi, i = 1, . . . , Nf , is the same for any frame transmitted by
any of the ni contending nodes, and remains constant until the ith management
frame is successfully exchanged between the tagged vehicle and the AP. Also,
the success events of different delivery trials of the same management/data frame
are independent. If a transmission collision happens among management and data
frames, none of the contending nodes can successfully receive any of the colliding
frames. On the contrary, if no transmission collision happens for a transmitted
frame, but the frame is lost due to a poor channel condition, the back-off procedure
of each node that successfully received the frame is invoked immediately at the end
of transmission of the lost frame, i.e., the additional wait time that consists of short
interframe space (SIFS) and ACK transmission durations is neglected [35].
In the following, the notation E(Y) denotes the expected value of a random
variable Y, E(Y|Z = z) the conditional expected value of Y given the event that
another random variable Z takes the value z, and max(a, b) the maximum of the
two values a and b.
2 The slot duration is defined as the duration between two consecutive variations in the back-off
counter or back-off stage of a contending node [36].

2.2.2 Access Delay Analysis
The objective of this section is to derive the average access delay that is required
for the tagged vehicle and the AP to complete the authentication and IP allocation
procedures by exchanging the necessary Nf management frames. First, we define a
time step as the sum of the durations required by the source of a management frame
to:
(a) generate the frame,
(b) complete the DCF back-off procedure and start the over-the-air transmission of
the frame, and
(c) either successfully transmit the frame and receive the corresponding ACK frame
or unsuccessfully transmit the frame and wait until the channel is sensed idle
(the earlier of the two events).
Based on the definition, the access delay from the time instant that the first
management frame is being generated until all the Nf management frames are
successfully exchanged can be partitioned into a sequence of time steps. At the start
of each time step, a management frame is required to be (re)transmitted either by
the tagged vehicle or by the AP. Let Xn be the index of the management frame that
should be exchanged between the tagged vehicle and the AP at the start of the nth
time step. Based on the system model, Xn is a discrete-time Markov chain that takes
integer values from 1 to Nf . Additionally, the value of Nf + 1 is added to the state
space of Xn to represent the event that all the Nf frames are successfully exchanged
between the tagged vehicle and the AP.3 Hence, when Xn = i, i = 1, . . . , Nf , the
Markov chain either transits to state i + 1 or remains at its current state, based on
whether or not the transmission of the ith frame is successful at the end of the nth
time step, as illustrated in Fig. 2.3. Therefore, in order to calculate the average access
delay, the main idea is to find the average duration that the Markov chain Xn needs
in order to transit from state 1 to state Nf + 1 for the first time. The remainder of
this section shows how this average duration can be obtained.
For Markov chain Xn, let pij denote the one-step transition probability from state
i to state j, where
Fig. 2.3 Illustration of the Markov chain and one-step transition probabilities for states 1 to Nf +1
3 When Xn = Nf + 1, the kth time step, k ≥ n, can take any positive value.

pij =
⎧
⎪
⎪
⎪
⎪
⎨
⎪
⎪
⎪
⎪
⎩
δi, i = j = 1, . . . , Nf
1, i = j = Nf + 1
1 − δi, i = j − 1 = 1, . . . , Nf
0, elsewhere.
(2.1)
In (2.1), the value of δi can be obtained by extending Bianchi’s DCF model [36]
to account for the frame loss due to channel conditions.4 That is, for each i =
1, . . . , Nf , the value of δi is calculated by solving the system of Eqs. (4.1a)–(2.19c)
in variables τi, αi, and δi:
τi =
2(1 − 2δi)
(1 − 2δi)(w + 1) + δiw(1 − (2δi)m−1)
(2.2a)
αi = 1 − (1 − τi)ni−1
(2.2b)
δi = 1 − (1 − αi)(1 − βi). (2.2c)
To show that there exists a unique value for each of τi, αi, and δi, from (2.19c)
and (2.19b) we have
τi = 1 −
1 − δi
1 − βi
1
ni−1
. (2.3)
Therefore, using (2.19a) and (2.3), we can prove the existence and uniqueness of the
solution for the system of the three equations (2.19a)–(2.19c) following a similar
approach as in [36]. Given the one-step transition probabilities in (2.1), the first
passage time probabilities can be obtained using
f (1)
ij = pij (2.4a)
f
(n)
ij =
Nf +1

k=1
k=j
pikf
(n−1)
kj , n 1 (2.4b)
where f (n)
ij denotes the n-step first passage time probability from state i to state j.
Note that, for the Markov chain in Fig. 2.3, ∞
n=1 f
(n)
ij = 1 iff j i or j = i =
Nf + 1, provided that δi = 1 ∀i. Now, let Dij denote the first passage delay from
state i to state j, i.e., the delay that the Markov chain requires to transit to state
4 When the tagged vehicle or the AP attempts to transmit the ith management frame, i =
1, . . . , Nf , each of the ni contending nodes always has a frame to transmit, i.e., in a traffic
saturation conditions [36], until the ith frame is successfully delivered.

j for the first time, given that the Markov chain is currently at state i, where i =
1, . . . , Nf , j = 1, . . . , Nf +1, and j i. By using the law of total expectation and
the first passage time probabilities from (2.4a)–(2.4b), and by noting that f
(n)
ij = 0
only if n ≥ j − i (Fig. 2.3), the expected value of Dij is given by
E(Dij ) =
∞

n=j−i
E(D(n)
ij )f (n)
ij ,
i, j ∈ {1, . . . , Nf + 1} and i j
(2.5)
where D
(n)
ij denotes the n-step first passage delay from state i to state j, i.e., the
delay that the Markov chain requires to transit to state j for the first time in n time
steps, given that the Markov chain is currently at state i. Consequently, the average
access delay can be directly obtained from (2.5), by setting i = 1 and j = Nf + 1.
However, in order to evaluate (2.5) for specific i and j values, the expected value
E(D
(n)
ij ) should be calculated ∀n ∈ N+ such that n ≥ j − i. For n ≥ j − i and
n = 1, the value of E(D
(n)
ij ) can be obtained in a recursive way as follows. Let
random variable K
(n)
ij denote the index of the first state to which the Markov chain
transits from state i, given that the Markov chain transits from state i to state j for
the first time in n steps, where i, j = 1, . . . , Nf , i j, and n ≥ max(j − i, 2). For
these i, j, and n values, let set Ω
(n)
ij = {k : pik = 0 and j − n + 1 ≤ k j} denote
all possible values of random variable K
(n)
ij , which is given by
Ω(n)
ij =
⎧
⎪
⎪
⎨
⎪
⎪
⎩
{i}, j = i + 1
{i + 1}, j = i + n
{i, i + 1}, elsewhere.
(2.6)
Hence, the expected value E(D
(n)
ij ) can be calculated by using
E(D
(n)
ij ) =

k∈Ω
(n)
ij
E(D
(n)
ij |K
(n)
ij = k)
pikf
(n−1)
kj
f
(n)
ij
=

k∈Ω
(n)
ij

E(D(1)
ik ) + E(D(n−1)
kj )
pikf
(n−1)
kj
f
(n)
ij
,
i, j ∈{1, . . . , Nf + 1}, i j, and n ≥ max(j − i, 2).
(2.7)
In order to evaluate E(D
(n)
ij ), it is required to find the values of E(D
(1)
ik ), ∀i ∈
{1, . . . , Nf } and k ∈ {i, i + 1}. First, we have

E(D
(1)
ik ) = E(Ui) + E(Vi) + DIFS + E(Rik),
i ∈ {1, . . . , Nf } and k ∈ {i, i + 1}
(2.8)
where Ui is the processing time at the start of a time step required to generate the
ith management frame, including the duration needed for communication through
the core network (if exists); Vi is the time spent until the channel is sensed idle and
the back-off procedure is invoked by the source of the ith management frame; DIFS
is the duration of a DCF interframe space [35]; and Rik is the remainder of a time
step, excluding the Ui, Vi, and DIFS durations, when the ith management frame is
either successfully (k = i + 1) or unsuccessfully (k = i) delivered. The processing
time, Ui, of the ith management frame is nonzero only before the first transmission
attempt of the frame (i.e., when the source of the frame is at back-off stage 0). When
Ui = 0, we have Vi = 0 in consequence, since each time step starts at a moment
the channel already starts to become idle.5 In order to calculate E(Ui), E(Vi), and
E(Rik), k ∈ {i, i + 1}, for a specific value of i ∈ {1, . . . , Nf }, let random variable
Bi denote the back-off stage of the source node that attempts to transmit the ith
management frame at the start of a time step. The probability distribution function
of Bi is given by
PBi (b) =
⎧
⎪
⎨
⎪
⎩
δb
i (1 − δi), b = 0, . . . , m − 2
1 −
m−2
q=0
δ
q
i (1 − δi), b = m − 1.
(2.9)
Hence,
E(Ui) =
m−1

b=0
E(Ui|Bi = b)PBi (b)
= E(Ui|Bi = 0)PBi (0)
(2.10)
E(Vi) =
m−1

b=0
E(Vi|Bi = b)PBi (b)
= E(Vi|Bi = 0)PBi (0)
(2.11)
E(Rik) =
m−1

b=0
E(Rik|Bi = b)PBi (b),
i ∈ {1, . . . , Nf } and k ∈ {i, i + 1}.
(2.12)
5 An exception is the first time step when i = 1, for which the value of V1 is neglected.

In (2.10), the value of E(Ui|Bi = 0) can be found for a given probability density
function of Ui, while in (2.11), the value of E(Vi|Bi = 0) can be approximated as
the duration of a successful over-the-air delivery of a data frame, i.e.,
E(Vi|Bi = 0) = h +
l
r
+ SIFS +
a
r
(2.13)
where h is the transmission duration of PHY information other than the PPDU data
field, e.g., PHY convergence procedure (PLCP) preamble and signal fields of the
IEEE 802.11 OFDM PHY [35]. In (2.12), the conditional expectation E(Rik|Bi =
b) can be calculated using (2.14a)–(2.14b) as follows:
E(Rii+1|Bi = b) = E(Cb)E(Si) + yib (2.14a)
E(Rii|Bi = b) = E(Cb)E(Si) + zib
i ∈ {1, . . . , Nf } and b ∈ {0, . . . , m − 1}
(2.14b)
where Cb denotes the value of the back-off counter of the source node at back-
off stage b, Si the duration required to decrease the back-off counter of the source
node by 1 when attempting to transmit the ith management frame, and yib (zib) the
remainder of a time step after the over-the-air transmission of the ith management
frame starts, when the transmission is successful (unsuccessful) and the source node
is at the bth back-off stage. Since at the bth back-off stage, the value of the back-off
counter is equally likely selected from 0 to 2bw − 1 [35], the expected value E(Cb)
is given by
E(Cb) =
2bw − 1
2
, b ∈ {0, . . . , m − 1}. (2.15)
The values of yib and zib can be calculated (by neglecting the propagation delay)
using
yib = h +
li
rib
+ SIFS +
a
rib
(2.16a)
zib = h +
βi(1 − αi)
δi
li
rib
+
αi
δi
max(
li
rib
,
l
r
) (2.16b)
i ∈ {1, . . . , Nf } and b ∈ {0, . . . , m − 1}.
Note that, in (2.16b), the values of βi(1−αi)
δi
and αi
δi
respectively equal the proba-
bility that a failure of delivering the ith management frame is due to a poor channel
condition only (i.e., no transmission collision) or involves a transmission collision
with a data frame. Finally, the value of E(Si) can be obtained using (2.17a)–(2.17c),
given by

E(Si) = (1 − ζi)σ + ζi

h +
l
r
+ DIFS

+ νi

SIFS +
a
r

(2.17a)
ζi = 1 − (1 − τi)ni−1
(2.17b)
νi = (1 − βi)(ni − 1)τi(1 − τi)ni−2
(2.17c)
i ∈ {1, . . . , Nf }
where σ is the idle slot duration, ζi and νi respectively denote the probability of a
transmission and the probability of a successful transmission in a slot duration from
the ni − 1 nodes that are contending with the source node of the ith management
frame. By using (2.1)–(2.4) and (2.6)–(2.17), the expected value of the first passage
delay, E(Dij ), from a state i to another state j can be obtained from (2.5). By setting
i = 1 and j = Nf +1, the value of E(D1Nf +1) represents the average access delay.
2.2.3 Delay Analysis and Simulation
This section provides numerical results based on the mathematical analysis in
Sect. 2.2.2 to investigate the access delay performance with respect to the number of
contending nodes, the wireless channel conditions, and the associated authentication
mechanisms. The first authentication mechanism under consideration is based on
the WPA2-802.1X mode, which is used for enterprise networks and requires an
authentication server [32]; while the second authentication mechanism is based
on the WPA2-PSK mode, which is mainly employed for home and small office
networks and does not require an authentication server [33]. The two authentication
mechanisms result in two different sequences of management frame exchanges
between the AP and the tagged vehicle, as well as different values of the additional
delay introduced for some management frames due to possible communication
between the AP and an authentication server through the core network. The numer-
ical results are generated based on the IEEE 802.11n standard, which (together
with the authentication mechanism) defines the sequence of management frames
that should be exchanged for the tagged vehicle to connect to the Internet through
the AP. When delivering the management frames, the values of each of βi and ni ∀i
(Sect. 2.2.2) are set to fixed values, denoted by β and n, respectively. Similarly, for
the ith management frame, the values of rib ∀b are set to a fixed value, denoted by
ri, i = 1, . . . , Nf , where each ri is set to the bit rate employed by the source of the
ith management frame at back-off stage 0, as obtained from the experimental testing
in Sect. 2.2.4. The experiment in Sect. 2.2.4 also provides the average processing
delay for each management frame, E(Ui|Bi = 0) ∀i in (2.10). This section also

includes computer simulations using MATLAB, in order to study the accuracy of the
mathematical analysis presented in Sect. 2.2.2. We simulate the exchange of the Nf
management frames between the tagged vehicle and the AP, for the WPA2-PSK and
WPA2-802.1X authentication modes, based on the IEEE 802.11 DCF for channel
access by all nodes. For each combination of n and β values in the simulations, the
average access delay required to exchange the Nf frames is estimated by using
200 samples (i.e., 200 repetitions of successful delivery of all the Nf frames),
which result in acceptable 95% confidence interval for all the n and β values under
consideration for each authentication mode. The parameter values used to obtain the
analytical, simulation, and experimental results are summarized in Table 2.1.
Figure 2.4 shows the access delay performance when the WPA2-802.1X mode
is used for authentication. As shown in Fig. 2.4a, the average access delay increases
almost linearly with the number of contending nodes, n, for a given wireless channel
represented by the probability, β, that a frame is lost due to a poor channel condition.
The rate of average access delay increase with n is higher when the β value
increases. For instance, in Fig. 2.4, the rate of increase of the curve corresponding
to β = 0.6 is approximately double that of the curve corresponding to β = 0.1.
The effect of β on the average access delay is illustrated in Fig. 2.4b for different
n values. When the value of n is small (n ≤ 5), increasing β up to 0.5 does not
result in a significant increase in the average access delay. The reason is that, if a
management frame is lost due to channel conditions, the additional delay required
to regain access of the channel and retransmit the fame is not significant when n
is small, due to a low channel contention level. On the contrary, when the n value
increases, the effect of β on the average access delay becomes more noticeable, as
shown in Fig. 2.4b. When β approaches 1, the average access delay tends to ∞,
as expected, since no management frame can be successfully delivered. There is
a good match between the analytical and simulation results, which indicates the
accuracy of the analytical model presented in Sect. 2.2.2. The same behavior of
the average access delay illustrated in Fig. 2.4 for the WPA2-802.1X standard is
observed for larger n values (up to 150) and when the WPA2-PSK mode is used
for authentication. However, when WPA2-PSK is employed, the average access
delay is considerably lower than that of the WPA2-802.1X mode, due to a smaller
number of management frames required to achieve the Internet access (Figs. 2.1
and 2.2). Figure 2.5 compares the average access delay for the WPA2-802.1X and
the WPA2-PSK modes for different n and β values. The average access delay
and its rate of increase with respect to n are higher for the WPA2-802.1X mode
as compared with the WPA2-PSK for all the n and β values shown. Results in
this section help to understand the behavior of the average access delay under
various channel conditions, with different number of contending nodes, and using
the different authentication methods, which is useful to select or develop a suitable
WiFi network access scheme for a vehicular environment.

Table
2.1
Parameter
values
used
to
generate
the
analytical,
simulation,
and
experimental
results
Param
Value
Param
Value
Param
Value
Param
Value
w
16
DIFS
SIFS
+
2σ
N
f
for
WPA2-802.1X
29
frames
a
32
bytes
m
7
Preamble
length
16
µs
N
f
for
WPA2-PSK
10
frames
l
1574
bytes
σ
9
µs
PLCP
header
length
4
µs
r
i
(ith
frame
transmitted
by
the
AP)
24
Mbps
l
i
for
WPA2-802.1X
Fig.
2.2
SIFS
16
µs
h
Preamble
length
+
PLCP
header
length
r
i
(ith
frame
transmitted
by
the
tagged
vehicle)
6
Mbps
l
i
for
WPA2-PSK
Fig.
2.1
r
24
Mbps
E(U
i
)
for
WPA2-802.1X
Varies
from
45
µs
to
64
ms
for
i
=
1,
.
.
.
,
N
f
E(U
i
)
for
WPA2-PSK
Varies
from
87
µs
to
70
ms
for
i
=
1,
.
.
.
,
N
f
–
–

0 5 10 15 20 25 30
0
1
2
3
4
5
6
7
n
Average
access
delay
(s)
Analysis
Sim, β =0.1
Sim, β =0.4
Sim, β =0.5
Sim, β =0.6
Sim, β =0.7
Sim, β =0.8
Sim, β =0.9
0.4
β
0.9
0.8
0.7
0.6
0.5
0.1
(a)
0.3 0.4 0.5 0.6 0.7 0.8 0.9
0
1
2
3
4
5
6
7
β
Average
access
delay
(s)
Analysis
Sim, n =1
Sim, n =5
Sim, n =10
Sim, n =20
Sim, n =30
n
30
20
10
5
1
(b)
Fig. 2.4 Analytical and simulations (Sim) results of the access delay when the WPA2-802.1X
standard is employed for authentication. (a) Average access delay versus n. (b) Average access
delay versus β

Other documents randomly have
different content

should be choaked with the great snows, or devoured by the wild
beasts, which come together in great troops. What he means by trees
that grow in a square figure I cannot tell, but I suppose he intends
only that they did use to erect their Tents between 4 trees which grew
so, that each of them might be the corner prop, of the four square
shed, but this sort is to us quite unknow. Tacitus saies the Fenni used
to dwell among a company of boughs, and perhaps that gave occasion
to our Author to talk thus. He hath also got a 4th
sort which he could
have no where else but from Zieglerus, for Zieglerus had called them
Amaxobios, from whence Olaus Magnus saies they dwelt in Waines
and Carts; and therefore Olaus induced by this word of Zieglerus,
thought the Laplanders had bin such. But this is quite false, for
Waggons and Carts were utterly unknown to the Laplanders, for whom
it was impossible to use them, by reason of the slipperiness of the Ice,
and depth of their snows. Neither was it indeed in that sense that
Zieglerus calls them Amaxiobios, but because they wandred up and
down like the Amaxobii, who are a known Nation of the Scythians.
There remains therefore only these two sorts of sheds, which I have
mentioned, for the 5th
, which Paulus Jovius reckons, was either upon
sudden occasions, or used only by those that were under the dominion
of the Moscovites: the words of this Author are, “These People lie in
caves filled with dried leaves, or in trunks of trees made hollow either
by fire or age.” But in both our forementioned sorts, things are so
ordered that every Tent had two doors, one, a foredoor, and the other,
a backward; the former bigger and more ordinarily used, the latter
less, through which they use to bring in their provisions, and especially
the prey they took in hunting, also Birds, Beasts, Fishes, which it was
unlawfull for to bring in at the foredoor. These are the two doors with
the use of both, especially the back-door, through which it was
unlawfull for any woman to pass, because, as I said before, women
were forbidden to go into the back part of the Tent, the reason of
which I think to be partly this, because in that part they placed Thor
and sacrificed to him, and partly this, because it was esteemed an ill
omen for a hunter to meet a woman. And hither may we refer what
Zieglerus saies of that door, that it was unlawfull for the Woman to go
out of the door of the Tent that day her husband was gone a hunting,

which cannot be understood of any door but the back-door, the use of
which was not only that day but alwaies forbid women. The Laplanders
have no Chambers, but only certain spaces, which they determine and
bound by loggs and posts laid along on the ground, of which we shall
next speak. The whole space of ground within the Tent was so
ordered, that in the middle there might be a hearth, surrounded with
stones, in which there was a continual fire, except at midnight; behind
the hearth, toward the back part of the tent, they place three loggs,
with which they bound that space, of which we but now spoke. In the
middle of this space is the little door, at which only men must enter,
which they call Posse; right over against that is the common door,
which they call Ox; but that space we told you was bounded with
these three loggs, they call, lops; this place therefore is only proper to
men, and it is unlawfull for any woman to pass those loggs, and go
into it. Sam. Rheen saies about the kettle hanging over the fire, they
place the 3 blocks, upon which, with a hatchet, they divide their flesh,
fish, or other things they intend to make ready. He saies here indeed
the space is called Posse, but understands chiefly the space of the
door, for that was properly called Posse; the other space being called
Lops. The common door they used to make towards the South, and
the other towards the North. The space on both sides, and the sides
themselves they called Loide; here they made their bed chambers, the
husband with his wife and children lying on one side, and the servants
on the other. Olaus Petri saies only the daughters lay on the side of
the husband and wife, I believe, that their Parents might have them
alwaies nigh them, and so take greater care to secure their honesty,
whilst the sons in the mean time lay with the servants: but now the
spaces that remain towards the doors they call Kitta, and are ordained
for the use of the women, for in the space nigh the common door they
are brought to bed. But that you may the better understand all this, I
will here insert a description of the Area. A is the little door they call
posse, B and C is called lopps, as is the place where the men lay up
their hunting instruments. D and E are called loide, whereof one is the
appartment of the Master of the Family and his wife, the other of the
servants. F. G. is kitta, were the women are conversant. H. is the
hearth, I. the door called ox; those three logs upon which they divide

their flesh are the two that lay along towards I. and the 3d
crosswaies
distinguishes from other parts the mens appartment, or posse.
The 3d
thing we are to note in these sheds, is that they strew
their floors with branches of Birch trees, least by the rain they should
be wetted, and they use no other kind of pavement; only upon the
boughs, for cleanliness sake, they lay skins of Rain-deers, on which

they sit and lie. And these are the dwelling houses of the Laplanders,
besides which they have also Store-houses in which they keep their
commodities, especially flesh, fish, and such other provisions; these
they call Nalla, and make thus: they cut the upper part of a tree off, so
that the body remain four or five ells from the ground high, upon this
trunk they place two rafters in the figure of an X, or St
Andrews Cross,
and upon these they build their repository, making a door to it, and
covering it with boards. There is one thing peculiar to these Store-
houses, which is, that the door is not in the side, but bottom of them,
so that when the Laplander is come down, the door falls too, like a
trap-door, and all things are safe. To these they go up by ladders
which they make of the trunks of trees, in which they cut great
notches like stairs. Now the reason why they place them so high, is
because of the Bears and other wild beasts, who oftentimes pull them
down, and to the great dammage of the Master eat all his provision;
they used also to cut off the bark of the tree, and anoint the stock, so
that neither mice nor wild beasts could be able to climb up for
slipperiness. And perhaps these are the houses Olaus magnus meant,
when he said, they placed their houses upon trees for fear of wild
beasts. But that you may the better conceive these Store-houses also,
I shall here give you the Figure of them.

Internet Access in Vehicular Networks Wenchao Xu

CHAP. XVII.
Of the Garments of the Laplanders.
Among the Laplanders the men and women wear different kinds of
Garments, which they alter according to the Weather, and place: for
they wear one sort of clothes in the Winter, and another sort in the
Summer, one kind at home, and another abroad. Let us first consider
the Garments of the men: These in the Summer have trouses, or
brougs, reaching down to their feet, close to their body, upon which
they wear a gown, or rather a coat with sleeves, which comes down to
the middle leg, which they tie fast with a girdle. And in this respect it
was that Zieglerus in his time wrote, that they used close Garments
fitted to their body, least they should hinder their work. He calls them
close because of their trouses, and fitted to their body because of their
being girded. These they wear next their bare skin, without such
linnen shirts as the Europeans use, they having no flax in their
Country. These Garments are of course home-spun woollen cloth
called Waldmar, of a white or gray color, such as the wool is of before
it is dyed. The wool they have from Swedland, and buy it of the
Merchants called Birkarli, but the richer sort wear a finer cloth, and not
of the same color, but sometimes green or blew, and sometimes red,
only black they abominate. Tho sometimes in dirty works, and at home
they wear the meanest clothes, yet abroad, and especially upon
Festivals and Holydaies, they love to go very neat. Their girdles are
made of leather, which the richer sort adorn with silver studs, and
poorer with tin. These studs stick out like buttons in a semicircular
figure. At this girdle they hang a knife and sheath, and a kind of
square bag, tho something longer then broad, also a leathern purse,
and then a case with needles and thred in it. Their knives they have
from Norway, the sheath is of the skin of the Rain-deers, sewed
together with tin wire, and in other parts with the same adornments,

at the end of which they use to hang rings: the bag is also made of
the skin of the Raindeers, with the hair on it, on the outside of which
they also place another skin, equall to the bag, and make it fast by
three knots, and this skin they cover again with red cloth or of some
other color, adorned also with wire. In this bag they keep a stone to
strike fire, not of flint, but christall, as I will shew hereafter. Also a
steel, with some brimestone to light a fire where ever they come: as
also Tobacco and other odd things. The leathern purse is also made of
the same skin in an oval figure like a pear, in which they keep their
mony, and other more choice things, and at this also they hang rings.
Their needle case is of a peculiar sort, they have a single cloth with
four sides, but the upper part is much narrower then the lower, so that
it is like an oblong triangle cut off at the vertical angle, and to make it
stronger they bind about the edges with leather, and so stick their
needles into it, this they put into a bag of the same shape, adorned
with red, or some other colored cloth, and wire, drawn together by a
leathern string, by which they hang it to their girdle. Besides these,
they have Alchymy chains, with a great company of rings of the same,
these they hang about all their body, the bag they hang before, nigh
their navel, all the rest they fling behind them. And these are the
Garments and ornaments of the body: their head they cover with a
cap, over which the richer sort wear a case of Fox, Beaver, or Badgers
skin, they are very like our night-caps, it is made of red or other
colored cloth, or of the Hares fur, first twisted into a thred, and then
knit almost like our stockins; or lastly of the skin of the bird called
Loom, with the feathers on it: sometimes they so order it, that keeping
also the head and wings of the bird, they make not an unbecoming
cover for the head. Olaus Magnus in his 4. Book, Cap. 3. saies they
make their caps of the skins of Geese, Ducks, Cocks, which, as well as
other birds, are there in great abundance. But he doth not here mean
common Cocks, but the Urogalli, or Heath-Cocks; however he gives us
the picture in his 17 Book, Cap. 26. They have ordinary gloves, but
shoes of a peculiar make, they are made of the skin of the Rain-deer
with the hair on, out of one piece, only where they tread they sew
both ends together, so that the haires of one part may lie forward, and
the other backward, least if they lay all one way they should be too
slippery: but neither is there any more leather on the bottom then on

other parts, as it is in our shoes, only there is a hole at the top in
which they put in their feet: the toe bends upwards, and ends as it
were in a point. Upon the seame they place some narrow pieces of
red, or other colored cloth: these shoes they wear on their bare feet,
and bind them twice or thrice about the bottom with a thong, and
least they should be too loose, they fill them up with a sort of long
Hay, which they boil and keep for that purpose.
But now let us come to the garments they do not so ordinarily
wear, but only on some occasions, which both for the men and women
are made alike, and all of leather, to secure them from the gnats. But
in the Winter time the men have breeches to defend them from the
weather, and coats which they call Mudd. These Mudd are not all alike,
but some better, some worse; the best are of the skins of young wild
Rain-deers, just when they have cast their first coat, in the place of
which comes a black one, which is about the Feast of St
James, and
these are very soft and delicate. Their feet they defend with boots of
the same skins, and their hands with gloves or mittens of the same,
and their heads with a cap, which reaches down and covers part of
their shoulders also, leaving only a space for them to see through. All
these Garments they wear next their skin without any linnen
underneath, and tie them round with a girdle, only their boots and
gloves they stuff with hay, and sometimes in the Winter with wool.
And this is that which Johannes Tornæus saies of their cloathing, that
their garment is made of the Rain-deer, the skin of the beast supplying
them with coats, breeches, gloves, sandals, shoes, c. the hair being
alwaies on the outside, so that they seem to be all hairy. And hence
we may understand Zieglerus, when he saies their Winter garments
were made of the skins of Bears and Sea-Calves, which they tied in a
knot at the top of their heads, leaving nothing to be seen but their
eyes, so that they seemed to be in a sack, only that it was made
according to the shape of their members; and hence, saies he, I
beleive they came to be supposed all hairy like beasts, some reporting
this out of ignorance, and some delighting to tell of strange wonders
they saw abroad. And truly it is not without reason that he gathers the
fable of hairy men to be raised from their hairy Garments, which sort
of monsters whether there be in other Countries I cannot tell, but I

find the Cyclops’s with one eye in their forehead by Adamus Bremensis
to be placed here upon the same account, because they had only a
hole in their cap through which they looked, all the rest of their body
seeming hairy, and therefore this hole they feigned to be an eye. But
whereas he saies the skins were of Bears and Sea-Calves, he is a little
mistaken, for these skins were not so common among the Laplanders,
and are by them designed quite for another use. However these
Garments they used after their fashion to adorn with pieces of red, or
other colored cloth, and embroider them with wire, in flowers, stars,
c. as I will hereafter declare more at large.
But I come to the habit of the women, which also was of one sort
in the Summer, and of another in the Winter. In the Summer they wear
coats which cover their breasts, arms, and all their body, about the
middle they are gathered, and so hang down, these they call Volpi.
These gowns they also wear next their skin, for the use of smocks is
no more known among women then the use of shirts among men: and
they horribly imposed upon Lomenius Comes, that made him beleive
otherwise. Lomenius saies thus, they have smocks, not made of
linnen, but of the entrails of beasts, which they first spin into thread,
and afterwards wear them: but all this is quite false. The entrals
indeed they do spin into thred, but of that they make neither cloth nor
smocks, but use it to sew their skins; but women of the common sort
wear course cloth, and the better sort finer, as it is with the men,
which for the most part is English cloth, richly wrought. They have also
a girdle, but different from that of the men, for it is much larger, and
sometimes three fingers broad, and then also it is adorned not with
studs, but plates of a fingers length, or more, which are engraved with
divers shapes of Birds, Flowers, c. and these they fasten upon a
leathern fillet so nigh one another, that the girdle is almost covered
with them. These plates are most commonly made of tin, from whence
Sam. Rheen calls them tin girdles, but those for the better sort are
made of silver. Upon these girdles they hang many Alchymy chains,
upon one of which they hang a knife and sheath, upon another a
pouch or purse, upon another a needle case, and upon all a great
company of Alchymy rings, according to the fashion of the men: These
things they do not hang by their sides, as women among us use, but

before them. The weight of the trinkets they carry about them, doth
commonly weigh twenty pound, a pretty heavy burden, and such as a
man would wonder they should be able to bear: but they are very
much delighted with it, especially with the number of the rings, the
gingling of which is very gratefull to their ear, and as they think no
small commendation to their beauty. Wexionius makes the chains and
rings to be tin, which I beleive is hardly true, commonly I am sure they
were made of Alchymy, and if they had bin of tin they had neither bin
durable, nor would they have made a noise. They have also another
ornament for their breast, which they call Kracha, it is made of red, or
some other colored cloth. And first it goes about their neck, and then
on both sides comes down upon their breast, and a little below their
breast ends in a narrow point. This cloth, especially before, and
sometimes about the neck, they adorn with studs, engraved with
divers forms, as also with bracelets, which the richer have of silver and
gold, the poorer of tin and Alchymy. After this manner, in short as he
uses, Johannes Tornæus describes them, the women do so deck
themselves with gold and silver that their breasts shine like sheilds,
but those that cannot reach silver, use copper and Alchymy. Now these
studs they use to have not only about their neck, but upon their
gowns where they draw them together, and lace them; and not only in
single but double and triple rows. They cover their heads with a low
kind of kercheif, plain at top, round, and of red color, some of the
richer sort on extraordinary times add also a strip of linnen for
ornament, as at their Fairs, Weddings, and Feasts. Upon their legs
they wear stockins, which reach no lower then their ankles, but that
only in the Summer. Their shoes are like the mens, and so also bound
to their feet with thongs. The womens habit in the Winter is almost
the same with the mens, for they have the Muddas made of the skins
of Rain-deers, and at that time wear breeches too, by reason of the
deep Snows, storms, and badness of the waies: nay and cover their
head with the same caps men do, which sort of caps they wear also
sometimes in the Summer to defend them from the gnats: these caps
they tie about their heads, and the lower part, which would otherwise
fall about their shoulders, they make to stand out like the brims of our
hats. And these are the garments as well of Virgins as married women,
for both use the same attire, neither is there any sign in their habit

whereby to distinguish them. Besides these garments wherewith they
clothe themselves in the day, they have also other which they use a
nights, such as are called night-cloathes, for they have no feather
beds: and without all doubt Olaus Magnus is mistaken who in his 4
Book saies they had. Their night garments were of 2 sorts, such as
they lay upon, or such as they did cover themselves with, which also
differ according to the Summer and Winter Seasons. Those they lie
upon are Rain-deers skins, 2 or 3 of which they fling upon some birch
leaves, which they use instead of matts, without beds, upon the
ground, that they may lie softer, so that they lie upon the skins
without sheets, of the use of which they are quite ignorant. They cover
themselves in the Summer with blankets, which they call raaner or
ryer, and with these blankets they cover not only their whole body, but
also their heads too, to avoid the gnats, with which they are extremly
infested in the night time. But that they may breath with more
freedom, and not be inconvenienced with the weight of these
blankets, they sometimes hang them up over their head with ropes
fastned to the top of their Hut. These are their Summer coverlets: but
in the Winter they first throw about them the skins of Sheep or Rain-
deer, and on them the blankets now mentioned. And there is one thing
more worth our notice, that they lie under these both Winter and
Summer stark naked, and make no use of linnen. And so much for the
Garments of the Laplanders. I shall add the Figures of both Sexes
habited after their manner. The woman hath a child in her arms, in a
Laplandish Cradle.

CHAP. XVIII.
Of the Diet of the Laplanders.
Having discoursed of their Garments, I proceed to speak of their
Diet. Their food is not the same amongst them all, but different
according to the places they inhabit. The Mountaineers live almost
wholly on their Rain-deers, that furnish them with Milk, Cheese, and
Flesh: tho sometimes they buy from the neighboring parts of
Norway Sheep, Goats, and Oxen, which they milk in the Summer,
and kill in the Winter, because they have neither Pasture nor Stable
room for them to keep them long. And for this reason they buy but
very few of them, and feed almost altogether on their Rain-deer,
which they have in great abundance. The flesh of these they feed on
in the Winter, and that alwaies boiled, but in the Summer their diet is
Milk, Cheese, and dried flesh. Their dainties most in esteem with
them are the tongue and marrow of their Rain-deers, and with these
they are want to entertain their Priests. One odd kind of dish these
of the Mountains have, and that is the blood of their Rain-deers
boiled in water to the consistence of a hasty pudding. The others
that dwell in the Woods feed partly on Fish, and partly on Birds and
Beasts, and that too both Summer and Winter, but more frequently
on Fish. The flesh of Beares they prefer before all other, and with
that they feast their dearest friends.
They have also some kind of Sawces of Black-berries, Straw-
berries, and other peculiar ones of their own, as also wild Angelica,
and the inner rine of the Pine-tree. The use of Bread and Salt is
almost unknown to them, and when they have any of the later, they
use it very sparingly. Instead of bread they eat dried fish, which by
grinding they reduce to a kind of meal, and instead of Salt the
inward rine of the Pine-tree, prepared after an odd kind of manner.

They pull the bark off first, and then they take the inward rine, and
divide it into thin skins like parchment, making it very clean; these
they dry in the Sun, and then tearing it into small pieces they put it
up in boxes made of the barks of trees: these they bury under
ground, and cover them with sand. When they have bin dried about
a day, they kindle a great fire over the hole where they put their
boxes, and by that means the rines acquire a red color, and a very
pleasant tast. On Fridaies they eat no flesh, but feed either on fish,
or milk, having retained this custom from their Roman Catholic
Priests. They boil all their fresh flesh, but not very much; that their
broth may be the better and fuller of gravy: and sometimes they put
also fish into the same kettle. Their milk they either boil with some
quantity of water, it being of it self to thick, or else they let it stand
in the cold, to freeze into a kind of Cheese, that it may be kept
longer for use. Their fish they eat sometimes fresh as soon as they
catch them; sometimes they dry them in the Sun, and being
hardned by the wind and air, they may be kept severall years.
Their sweet meats, which serve them instead of Apples, Nuts,
and the like, are preparations made of severall sorts of Berries.
When their Straw-berries begin to be ripe, they gather them, and
boil them in their own juice, without the addition of water, with a
slow fire, till they are very soft: then they sprinkle them over with a
little salt, and putting them into a vessell made of birch-bark, they
bury it in the ground: and in the Autumn and Winter when they have
occasion for them, they take them out as fresh as if they had bin
newly gathered: and these stand them in good stead when no other
Berries are to be had. Sometimes whilst they are fresh they put
them to the flesh of Fish, and make an odd kind of dish, after this
manner. Having boiled the Fish they first bone them, and then add
Straw-berries to them, and beat them together in a wooden pestle
to a mash, and so eat it with spoons. And this dish they make also
with all other kinds of Berries. Another Kickshaw that pleaseth them
very much, they make of Angelica. They take the staulks before it
seed, and scraping of the outward skin, they put the rest upon coals,
and so eat it broiled. They have also another way of preparing it,

and that is to boil them in whay for a whole day till they look as red
as blood. But this sort of meat is very bitter of it self, but by custom
becomes plesant enough to them, especially since they are
perswaded ’tis a great preservative of health. They likewise boil
sorrell in milk; as also the rine of the Pine-tree, which, as was said
before, being prepared, serves them instead of salt.
I come next to speak of their drink, which is ordinarily nothing
but water; Lomenius calls it dissolved Ice: but certainly he is
mistaken, for having such plenty of Rivers and Lakes, for all the Ice
they can hardly want water. And to prevent its freezing, they have
alwaies some hanging over the fire in a kettle; out of which
everyone with a spoon takes what he pleases, and so drinks it hot,
especially in the Winter time. Besides common water, they often
drink the broth I spoke of, made of flesh and fish, which they call
Læbma, and also whay, if you will beleive Olaus. These are their
usuall drinks; for Ale and Beer is utterly unknown to them. That
which they drink for plesure, is spirit of Wine and Brandy, with a little
of which you may win their very souls. This they buy from Norway at
their Fair times, and use it especially at their solemn Feasts and
Weddings. I had almost forgot Tobacco, of which they are very great
admirers, and traffic for it as one of their cheif commodities.
In the next place let us see the manner of their eating. Their
dining room in the Winter time is that part of the Hut where the man
and his wife and daughters use to be, and is on the right hand as
you go in at the foregate: but in Summer without doors upon the
green grass. Sometimes too they are want to sit about the kettle in
the middle of the Hut. They use not much ceremony about their
places, but every one takes it as he comes first. They seat
themselves upon a skin spread on the ground cross-leg’d in a round
ring; and the meat is set before them in the middle, upon a log or
stump instead of a table; and severall have not that, but lay their
meat upon the skin which they sit on. Having taken the flesh out of
the kettle, the common sort put it upon a woollen table cloth called
Waldmar, the richer on a linnen; as for trenchers and dishes they are
quite unknown to them. But if any liquid thing be to be served up,

they put it in a kind of trey made of birch. Sometimes without any
other ceremony every one takes his share out of the kettle, and puts
it upon his gloves, or his cap. Their drink they take up in a wooden
Ladle, which serves instead of plate. And it is farther observable that
they are abominable gluttons when they can get meat enough; and
yet hardy too to endure the most pinching hunger when they are
forc’t to it. When their meal is ended they first give God thanks, and
then they mutually exhort one another to Faith and Charity, taking
each other by the right hand, which is a symbol of their unity and
brotherhood. Samuel Rheen tells us they lift up their hands first, and
then say Grace after this manner, All thanks be given to God, who
hath provided this meat for our sustenance. This is their Grace in
Pithilapmark. In Tornelapmark their Grace is a little different; they
say in their own Tongue, Piaomaos Immel lægos kitomatz piergao
odest adde misg mosea wicken ieggan taide ko mig læx iegnaston,
that is, Good God praised be thou for this meat: make that which we
have at this time eaten give strength to our bodies. And so much for
their Diet, and manner of eating.

CHAP. XIX.
Of the Hunting of the Laplanders.
Having spoke of those things that relate to their Meat, Drink,
Cloathing, and other necessities, the subject of our next discourse
will be their employments, which are either rare and more solemn,
or daily and more usual: these latter too are of two sorts, either
common to both Sexes, or peculiar to one. Of those that are proper
to men Hunting is the cheif, for in this Countrey that exercise is
lawfull to none but men: Olaus Magnus is of the contrary opinion,
and saies Lib. 4. Cap. 12. that there is here such a multitude of
Beasts, that the men alone, without the help of women, dare not go
out to hunt; and therefore they are as active in this sport, if not
more than men. I beleive he had not this from any good tradition, or
his own knowledg, but rather followed the authority of some ancient
Writers, as Procopius Lib. 2. Gothic. or Tacitus de mor. Ger. for
whatsoever they say concerning the Fenni and Scritfinni, is so far
from being true of the Laplanders, that they do not permit women
so much as to touch their hunting weapons or beasts brought home,
and debar them all passage at that door thro which they go to that
sport, as will be shewed hereafter. They observe in hunting many
things with great superstition, as not to go out upon ominous daies,
such as St
Marks (whom they call Cantepaive) St
Clements and S.
Catherines, because they believe on these daies some misfortune
will happen to their weapons, and that they shall have no good
success all the year after. They think they cannot prosper, unless
they have first consulted their Gods by their Drum, which they use
before their going out, and have therefore severall beasts pictured
upon it. This is chiefly before the hunting a Bear. The third
observation is that they will not go out at the usual door, but at one

in the backside of the house called Posse, I suppose it is to avoid
women, the meeting of whom is an ill omen to huntsmen, and
therefore they are forbidden to come on that side of the house
where this door is, as Ol. Matthias assured me while I was writing
this, who was very well acquainted with this Country. Zieglerus saies
the same, tho something obscurely, that a woman is not to go thro
this door that day her husband is hunting: but it is not only that day,
but at no time else. All these things are by way of preparation. The
hunting it self is various according to the time of year, and severall
sizes of beasts. In the Summer they hunt on foot with Dogs, which
are very good in these parts, not only for their scent, but that they
dare set upon any thing, being still tied up to make them more
fierce. In the Winter they themselves run down the game, sliding
over the snow in a kind of scates, which I shall describe more fully in
another place. Little beasts they chase with bow and arrows, the
greater with spears and guns; tho sometimes they use other arts.
That sort of beast they call Hermelines, they take in traps as we do
Mice, which are so contrived of wood that the touching of any part
makes them fall; sometimes in pits and holes covered with snow, to
hide the deceit, as also with Dogs that will gripe them to death.
Squirrels they shoot with blunt darts, that they may not do an injury
to their skins, which they very much esteem. After this manner also
they take Ermines. Other beasts, as Foxes, Beavers, they kill with
Javelins spiked with iron: but if they meet with a beast that hath a
pretious skin, they are so expert at their weapons, as to direct the
blow where it will do it least harm. Foxes are frequently tempted
with baits upon the snow strowed upon twigs over deep pits, or
caught in gins laid in their usuall haunts, or else poisoned with a sort
of moss, which is peculiar for this use, but is seldome made use of
where there are abundance of field Mice, which are the Foxes
generall food. They fasten snares to boughs of trees to catch Hares
in, and some of the above mentioned beasts: and if any one find any
thing fast in these, he is obliged to give notice to the owner. I come
now to the larger beasts, of which Wolves are most commonly
caught in holes, but sometimes shot with bullets: these are their
game frequently because they have the greatest plenty of them, and

suffer the most dammages by them: and for their greater
destruction, Sithes are often hiden under the snow to cut off their
legs. After this manner too Leopards and Gulo’s are destroyed, which
is now a daies almost left off, because the Countrey is so well
furnished with guns, with which they also kill Elkes when they can
find them. But with greatest care and diligence they hunt Rain-deers
and Bears, the former with all kind of weapons. At their rutting time
in Autumn, about S. Matthews day, they entice them to their tame
does, behind which the Huntsman lies to shoot them. And in the
Spring, when the Snow is deep, the men themselves slide after
them, and easily take them, or sometimes drive them into traps with
Dogs: or lastly they set up hurdles on both sides of a way, and chase
them in between them, so that at last they must necessarily fall into
holes made for that purpose at the end of the work. The hunting of
the Bear follows, which, because it is done with the most ceremonies
and superstitions, will require the more care and accurateness in the
relating of it.
First of all, their business is to find out where the Bear makes
his den against Winter. He that finds it is said hafwa ringet bioern,
i.e. to encompass the Bear. He usually after this goes to all his
friends and acquaintance with much joy, to invite them to the
hunting as to a solemn and magnificent feast, for, as is before said,
this beasts flesh is a great delicacy. But they never meet before
March or April, till they can use their sliding shooes: at which time
he chooses the best drummer among them, and by his beating
consults whether the hunting will be prosperous or no, which done
they all march into the field in battel array after him that invited
them as Captain, who must use no other weapon then a club, on
whose handle is hung an Alchymy ring. Next him goes the drummer,
then he that is to give the first blow, and after all the rest as their
office requires, one to boil the flesh, another to divide it, a third to
gather sticks and provide other necessaries: so they strictly observe
that one should not incroach upon anothers office. When in this
order they are come to the den, they set upon the Bear valiantly,
and kill him with spears and guns, and presently sing in token of

victory thus, Kittulis pourra, Kittulis ii skada tekamis soubbi iælla
zaiiti, that is, they thank the Bear for coming, and doing them no
harm in not breaking their weapons, in the singing of which their
Captain is the cheif Musician. After celebration of their victory, they
drag the Bear out, beating him with staves, whence they have a
Proverb, slao bioern med riis, that is, the Bear is beat, which
signifies he is killed. Then putting him upon a sledge, they draw him
with Rain-deers to the Hut where he is to be boiled, singing Ii paha
talki oggio, ii paha talka pharonis, that is, they beseech the Bear that
he would not raise tempests against them, or any way hurt them
that killed him. This they say by way of jest, unless we will suppose
them (as some of them really do) to imagine the killing of some kind
of wild beast portends ill to the hunter. Samuel Rheen speaks of a
different song from this we have mentioned, much to this purpose,
that they thank God for making beasts for their service, and giving
them strength and courage to encounter and overcome so strong
and cruel a creature, and therefore I beleive they may join them
together and sing both. That Rain-deer that brings home the Bear is
not to be used by Women for a year, and some say, by any body
else. If there be materialls, near the place where the Bear is kill’d,
they usually build up a hovel there to boil him in, or if not, carry him
to a place that is more convenient, where all their Wives stay to
expect them, and as soon as the men come nigh them they sing
Læibi ia tuoli susco, that is they ask their wives to chew the bark of
the Alder Tree and spit it in their faces. They use this rather then
any other Tree, because when ’tis bruised between their teeth, it
grows red, and will dy any thing, and the men being sprinkled with
this, as if it were the Bears blood, seem to have gone through some
notable exploit not without danger and trouble. Then their wives
aiming with one eye through an Alchymy Ring spit upon them:
Samuel Rheens opinion differs only in this, that but one woman spits
in the Captains face. This ceremony is not done in the Hut where the
Bear is kill’d, but at the back door: for they build two Tents, one for
the men where the Bear is to be drest, and the other for the women
in which they make the feast: where as soon as the men come in,
the Women sing Kittulis pouro tookoris, that is, they thank their

Internet Access in Vehicular Networks Wenchao Xu

More Related Content

Similar to Internet Access in Vehicular Networks Wenchao Xu (20)

Recently uploaded (20)

Internet Access in Vehicular Networks Wenchao Xu