PDF Fog Radio Access Networks F RAN Architectures Technologies and Applications Mugen Peng Download
PDF Fog Radio Access Networks F RAN Architectures Technologies and Applications Mugen Peng Download
PDF Fog Radio Access Networks F RAN Architectures Technologies and Applications Mugen Peng Download
OR CLICK LINK
https://textbookfull.com/product/fog-radio-access-
networks-f-ran-architectures-technologies-and-
applications-mugen-peng/
Read with Our Free App Audiobook Free Format PFD EBook, Ebooks dowload PDF
with Andible trial, Real book, online, KINDLE , Download[PDF] and Read and Read
Read book Format PDF Ebook, Dowload online, Read book Format PDF Ebook,
[PDF] and Real ONLINE Dowload [PDF] and Real ONLINE
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
https://textbookfull.com/product/5g-radio-access-networks-
centralized-ran-cloud-ran-and-virtualization-of-small-cells-1st-
edition-trestian/
https://textbookfull.com/product/spectrum-access-and-management-
for-cognitive-radio-networks-1st-edition-mohammad-a-matin-eds/
Photonic Applications For Radio Systems Networks Fabio
Cavaliere
https://textbookfull.com/product/photonic-applications-for-radio-
systems-networks-fabio-cavaliere/
https://textbookfull.com/product/elastic-optical-networks-
architectures-technologies-and-control-1st-edition-victor-lopez/
https://textbookfull.com/product/privacy-enhancing-fog-computing-
and-its-applications-xiaodong-lin/
https://textbookfull.com/product/media-convergence-and-the-
development-strategies-of-radio-and-television-in-china-1st-
edition-peng-duan/
https://textbookfull.com/product/media-convergence-and-the-
development-strategies-of-radio-and-television-in-china-1st-
edition-peng-duan-2/
Wireless Networks
Mugen Peng
Zhongyuan Zhao
Yaohua Sun
Fog Radio
Access
Networks
(F-RAN)
Architectures, Technologies, and
Applications
Wireless Networks
Series Editor
Xuemin Sherman Shen
University of Waterloo
Waterloo, ON, Canada
The purpose of Springer’s new Wireless Networks book series is to establish
the state of the art and set the course for future research and development in
wireless communication networks. The scope of this series includes not only all
aspects of wireless networks (including cellular networks, WiFi, sensor networks,
and vehicular networks), but related areas such as cloud computing and big data.
The series serves as a central source of references for wireless networks research
and development. It aims to publish thorough and cohesive overviews on specific
topics in wireless networks, as well as works that are larger in scope than survey
articles and that contain more detailed background information. The series also
provides coverage of advanced and timely topics worthy of monographs, contributed
volumes, textbooks and handbooks.
Yaohua Sun
School of Information and Communication
Engineering
Beijing University of Posts and
Telecommunications
Beijing, China
This Springer imprint is published by the registered company Springer Nature Switzerland AG.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
In the fifth generation (5G) radio access networks and beyond, a paradigm of
fog computing-based radio access network (F-RAN) has emerged to meet the
requirements of the explosively increasing high-speed applications and the massive
number of Internet-of-Things (IoT) devices. Inherited from both heterogeneous
networks (HetNets) and cloud computing-based radio access networks (C-RANs),
F-RANs take full advantages of cloud computing, fog computing, and heterogenous
cooperative processing. F-RANs can coordinate the severe interference via the adap-
tive distributed and centralized processing techniques and provide great flexibility
to satisfy quality-of-service requirements of various intelligent applications and
services, such as the ultralow-latency of about 1.0 ms required by self-driving based
Internet of vehicles, and up to 99.999% reliability, and even 106 connections/km2
density as required by intelligent manufacturing.
Since F-RAN first combines fog computing, cloud computing, and HetNets
together and has been regarded as the key evolution path to the 5G beyond
system, it has drawn a large number of attention from both academia and industry.
As integrating with artificial intelligence (AI), non-orthogonal multiple access
(NOMA), and other advanced emerging technologies, F-RANs have entered a new
researching and development era to solve the challenges that 5G beyond system
meets. Actually, it is well known that F-RAN is potentially an evolutionary path
to the sixth generation (6G) mobile system. This is a cutting-edge technique of
multiple disciplines, including AI, wireless networks, radio signal processing, fog
computing, and cloud computing. The versions and classical application scenarios
defined in 5G are hardly fulfilled in 2020, while they will be fully provided in 6G.
In particular, 6G can meet requirements of enhanced mobile broadband, massive
machine-type communications, and ultra-reliable and low-latency communications
for rich IoT services, which will penetrate various industry and business applica-
tions. In terms of these rich IoT services, F-RAN can provide a unified framework
for massive access of heterogonous IoT devices, which simplifies the control
and management mechanisms. With respect to AI applied in F-RANs, sufficient
computation resources in fog nodes can be provided to execute machine learning
v
vi Preface
and deep learning algorithms, which means that both the cost and the training
efficiencies can be significantly improved.
As the F-RAN moves from the theoretical research to real world applications
industry and academia are working together towards the protocols defined in
standards and algorithms of all air interface layers in products, so as to enable
spectral-, energy-, and cost-efficient F-RANs to be widely used as a key solution
for 5G beyond and even 6G systems.
This book is firstly intended to present a comprehensive overview framework
of recent advances in F-RANs, from both the academia and industry perspectives.
In particular, this book covers the architecture, performance analysis, physical-
layer design, resource allocation, computation offload, and field trials. The recent
academic research results of F-RANs, such as the analytical results of theo-
retical performance limits and the optimization theory-based resource allocation
algorithms, have been introduced. Meanwhile, to promote the implementation of F-
RANs, the latest standardization procedure and testbed design have been discussed
as well. Finally, this book will be concluded by summarizing the existing open issues
and future trends of F-RANs.
We sincerely hope that this book will serve as a powerful reference for engineers
and students in the majors of electronic, computations, and communications, which
results in motivating a large number of students and researchers to tackle these
numerous open issues and challenges highlighted in F-RANs for 5G beyond and
even 6G systems.
First, we would like to express our sincere gratitude to all contributors, without
whom this book would have never been published. Our contributors are as follows:
Mugen Peng, Zhongyuan Zhao, Yaohua Sun, Shi Yan, Chenxi Liu, Bin Cao, Xiqing
Liu, Hongyu Xiang, Kecheng Zhang, Binghong Liu, Zhendong Mao, Wenbin Wu,
Tian Dang, Ling Qi, Yangcheng Zhou, Yuan Ai, Xinran Zhang, Xian Zhang, Bonan
Yin, and Wenyun Chen. Meanwhile, we would like to thank Springer Press staff for
their continuous encouragement and support.
Second, we would like to thank our colleagues and friends all over the world who
promoted fog computing into cellular networks and Internet-of-Things (IoT) for 5G
and 6G developments. This book was supported in part by the State Major Science
and Technology Special Project (Grant No. 2018ZX03001025-004), the National
Natural Science Foundation of China under No. 61925101, 61921003, 61831002,
and 61901044, the Beijing Natural Science Foundation under No. JQ18016, and
the National Program for Special Support of Eminent Professionals. Without the
support of these funding, F-RANs are hard to formulate and the corresponding
research works could be hardly achieved within a short time. Meanwhile, I would
like to thank Prof. H. Vincent Poor and Dr. Chonggang Wang because they helped
open the window of C-RANs when Prof. Mugen Peng was an academic visiting
fellow at Princeton University.
Third, we also thank all of those who have played a part in the preparation of
this book. Our coordinator, Hemalatha Velarasu, provided useful guidelines and
monitored the whole process carefully. Our Senior Editor, Mary E. James, arranged
the book review and fed back the review comments in a very short time. Also, special
thanks are given to Production Editors, Brian Halm and Brinda Megasyamalan, as
well as Assistant Editor, Zoe Kennedy.
Most importantly, we appreciate Series Editor, Prof. Xuemin (Sherman) Shen,
for his great support and invaluable suggestions.
Last but not least, we would also like to sincerely thank the support of our family
members for supporting our research work when we were away from home.
vii
Contents
ix
x Contents
10 Future Trends and Open Issues in Fog Radio Access Networks . . . . . . 203
10.1 Future Trends of F-RANs: Federated Learning-Based Paradigms . 206
10.1.1 The Conventional Federated Learning Paradigm . . . . . . . . . 207
10.1.2 A Hierarchical Federated Learning Paradigm . . . . . . . . . . . . . 208
10.1.3 Potential Applications of Federated Learning in F-RANs 209
10.2 Fundamentals of Federated Learning in F-RANs. . . . . . . . . . . . . . . . . . . 209
10.2.1 Loss Compensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
10.2.2 Model Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
10.2.3 Performance of Loss Compensation and Model
Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
10.3 Key Enabling Techniques of Future Evolved F-RANs . . . . . . . . . . . . . 214
10.3.1 Hierarchical Cloud and Fog Computing . . . . . . . . . . . . . . . . . . . 214
10.3.2 Advanced Transmission Techniques . . . . . . . . . . . . . . . . . . . . . . . 214
10.3.3 Resource Management and User Scheduling . . . . . . . . . . . . . 215
10.3.4 Intelligent NFV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
10.4 Open Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
10.4.1 Massive Multiple Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
10.4.2 New Theory and Techniques of Deep Learning . . . . . . . . . . . 217
10.4.3 Security and Privacy Issue of Local Model Feedback . . . . 217
10.5 Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Chapter 1
Brief Introduction of Fog Radio Access
Networks
In Sect. 1.1, we will firstly review the evolution path of radio access networks
(RANs) from heterogeneous networks (HetNets) and cloud radio access networks
(C-RANs) to the fifth generation (5G) and even the sixth generation (6G) mobile
communication systems, and various new emerging services will be introduced,
which show the necessity of presenting fog computing. Then, in Sect. 1.2, the key
features of fog computing and the relationship between fog computing and mobile
edge computing (MEC) will be summarized. Finally, the fog radio access network
(F-RAN) for 5G and 6G will be presented in Sect. 1.3, and the applications as well
as standardization activities with respect to F-RANs will be introduced in Sect. 1.4.
Fig. 1.1 The evolution of BS structure in cellular systems (Peng et al. 2015b)
as relay stations (RSs), distributed antennas, and access points (APs) for picocells,
femtocells, and small cells. These new small BSs are either operator-deployed or
consumer-deployed, and they have taken big changes to the traditional cellular
architecture, which forms a mix of small cells underlying the macrocells. To address
these challenges and the corresponding changes, heterogeneous network (HetNet)
comprising of marco base stations (MBSs) and small BSs has been presented as an
emerging network paradigm evolution in 4G/5G systems (Vision 2013).
the frequency band can be reused. In particular, the severe inter-tier interference
is hardly coordinated due to the random deployment of LPNs.
As shown in Fig. 1.2c, in the hybrid overlay and underlay strategy, LPNs partially
reuse the spectral resources of HPNs and thus results in the underlay structure.
While the other portion of spectral bands are separately and orthogonally reserved
for HPNs and LPNs, respectively. Considering the balance between performance
gains and implementing complexity, the hybrid strategy is a good solution.
Since the frequency band is scarce, and the underlay HetNet strategy is the most
promising to improve SE and EE, it has already attracted significant attentions
and been defined in 3GPP standards. However, for the successful rollouts, it still
comes with own challenges. As shown in Fig. 1.3, interference coordination and
cancelation (ICC), radio resource allocation optimization (RRAO), cooperative
radio resource management (CRRM), and self-organizing network (SON) are four
key techniques. In terms of ICC, different types of interference should be tackled
in the physical layer, including the inter-tier, inter-cell, and intra-cell interference.
RRAO aims to assign the scarce physical radio blocks (PRBs) to different users
for maximizing SE or EE with low complexity in aspects of the multi-dimensional
resources. Since the resource assignment is strictly related to the radio channel status
and the adopted ICC in physical layer, RRAO often is based on the cross-layer
design mode. In addition, RRAO is responsible for scheduling the PRBs in one
cell, while CRRM mainly tackles the radio resource management among multiple
adjacent cells, which is often relevant to mobility management and soft frequency
reuse. Finally, to reduce the configuration and optimization cost, SON is proposed
to enable AI into the HetNet organization.
1.1 History and Evolution of RANs 5
common public radio interface (CPRI) protocol, and it is often capacity and delay
constrained (Peng et al. 2015a).
C-RANs have been advocated by both mobile operators and equipment vendors
due to the potential significant benefits through introducing cloud computing into
cellular networks, but they also come with their own challenges. The biggest
problem is the constraints of capacity and transmit delay from fronthaul links.
Meanwhile, the processing complexity and the corresponding delay in the BBU
pool degrade the performance gains from LSCP and CRRA. Meanwhile, due to the
non-ideal channel status information for all access links among RRHs and active
UEs in the BBU pool, it is hard to make interference mitigate (Park et al. 2013).
The evolution of C-RAN is necessary, but C-RAN’s advantages should be kept in
consideration.
1.1.3 5G Overview
Unlike 2G, 3G, and 4G, 5G is expected to fundamentally transform the role that
telecommunications technology plays in society, which can enable further economic
growth and pervasive digitalization of a hyper-connected society. Not only people
can be connected to 5G via smartphones whenever needed, but also devices,
machines, and even things can create the communicate society through 5G. As
a result, 5G formulates the internet of everything and connects people, devices,
machines, things, data, applications, transport systems, and cities via wireless
networks, which can support a wide variety of applications and services (Chih-Lin
et al. 2014).
There are three typical usage scenarios well defined in 3GPP for 5G, which
includes:
• Enhanced mobile broadband (eMBB): This usage is to deal with hugely increased
data rates, high user density, and high traffic capacity for hot spots scenarios as
well as seamless coverage and high mobility scenarios with still improved used
data rates.
• Massive machine-type communications (MMTC): This usage requires low power
consumption and low data rates for the connected devices. It is mainly with IoT
service.
• Ultra-reliable and low-latency communications (URLLC): This usage caters for
safety-critical and mission-critical applications, such as the massive connections
in the industry IoT.
As shown in Fig. 1.6, 5G is expected to provide 20 times the peak data rate, 10
times lower latency, and 3 times more spectral efficiency than 4G. 5G can transport
a huge amount of data with ultra-high bit rate, reliably connect an extremely
large number of devices, and process high volumes of data with minimal delay.
5G is expected to deliver significantly increased operational performance, i.e.,
increased spectral efficiency, higher data rates, low-latency, as well as superior user
8 1 Brief Introduction of Fog Radio Access Networks
Fig. 1.6 Performance requirements comparisons between 4G and 5G (Chih-Lin et al. 2014)
experience. Meanwhile, 5G is cater not only for eMBB, but also cater for massive
deployment of IoTs, which can offer acceptable levels of energy consumption,
terminal cost, network deployment, configuration, and operation cost.
The increased capacity and peak data rates require more frequency spectrum and
vastly more spectral efficient techniques in 5G than those in 4G. The spectrum bands
allocated for 5G can be divided into three main categories: sub-1 GHz, 1–6 GHz,
and above 6 GHz. Since the propagation properties of the signal benefits to create
large coverage areas and deep in-building penetration, sub-1 GHz bands are often
used to support traditional voice, real-time emerging services, and special services
in high mobility, which can extend the coverage from urban to suburban and rural
areas. The 1–6 GHz bands offer a reasonable mixture of coverage and capacity,
while spectrum bands above 6 GHz provide significant capacity thanks to the very
large bandwidth.
The additional spectrum mainly comes from frequency bands above 24 GHz,
which poses a huge number of challenges from the intrinsic propagation character-
istics of millimeter waves. To enhance transmit performance in the physical layer
of 5G, new radio (NR) has been defined in 3GPP, which is mainly based on flexible
1.1 History and Evolution of RANs 9
multiple access and coding techniques. Until now, there are two main frequency
bands that have been well defined, i.e., sub-6 GHz and the mmWave range (24–100
GHz) (Series 2015).
Unlike the traditional cellular system that requires both RAN and core network
work in the same generation to be deployed, 5G is expected to integrate elements
of different generations with different configurations. Standalone (SA) mode is
defined for using only one RAN, while non-standalone (NSA) mode is defined
for combining multiple RANs. For NSA, the 5G NR or the evolved 4G LTE radio
cells and the core network can be operated alone, which suggests that the NR or
evolved LTE radio cells can be used for both control and user planes. While SA is
a simple solution for operators, to make user of service continuity, it is deployed
as an independent network through normal inter-generation handover between
4G and 5G.
The initial phase of NSA in 5G mainly focuses on eMBB, which provides high
bit rate complemented by moderate latency improvements and supports several
classical use cases, such as AR/VR, UltraHD, 360-◦ streaming video. MMTC
has been already developed as a part of NB-IoT technologies. A huge number of
MMTC devices connect to the 5G BS, which makes it infeasible to allocate a priori
resources to individual MMTC devices. As a result, the corresponding radio access
mechanisms should be enhanced in 5G. The usage scenario of URLLC supports
low-latency transmissions with small payloads and high reliability, whose bit rate is
relatively low, and the main challenge is to ensure the transmit error rate should be
typically lower than 10−5 .
To satisfy with requirements of the aforementioned three usage scenarios,
especially allow them to coexist with a unified network architecture, network slicing
is urgent, which jointly allocates the resources of communication, cloud computing,
edge computing, fog computing, cloud storage, edge storage, and its aim is to
guarantee the isolation under the required performance levels.
1.1.4 6G Prospect
With the open of the scale-up commercial deployment of 5G, more and more
researchers and related organizations began to consider the evolution of 5G. At
the 2018 Mobile World Congress, an official of the Federal Communications
Commission looked ahead to 6G in public. Not only the USA, China also has
launched 6G related work in March 2018. Consider that wireless communication
systems upgrade to a new generation every 10 years, it can be foreseen that there is
a certain consensus on starting 6G related research since 2020.
The goal of 6G is to meet the needs of the informatization society ten years
later, so the 6G vision should focus on the needs that cannot be satisfied by 5G.
The 6G vision requires massive connectivity, reliability, real-time, and throughput
requirements, which are new and huge challenges to the existing 5G. As shown
in Table 1.1, 6G is expected to upgrade and improve to achieve 10–100 times
higher peak data rate, system capacity, spectrum efficiency, moving speed than 5G.
Meanwhile, it will achieve lower delay, wider and deeper coverage, which enables to
serve the interconnection of everything, fully support the development of intelligent
life and industries. There are several key characteristics that have been general
consent as follows (Chen et al. 2020):
• 6G is expected to be ubiquitous and integrated with broader and deeper coverage
than 5G, including many kinds of communications, such as terrestrial land
communication, space communication, air communication, sea and underwater
communication. With AI driven intent-based networking and network slicing
technologies, 6G can serve in various application scenarios, such as airspace, sky,
land, and sea. In a word, 6G can realize a global ubiquitous mobile broadband
communication system.
• 6G is expected to work on a higher frequency band and a wider bandwidth
to achieve higher peak bit rate and average network capacity than 5G, such as
mmWave communication, TeraHertz communication, visible light communica-
tion, and so on. Compared with 5G, 6G can provide a data rate up to 10–100
1.1 History and Evolution of RANs 11
times, supporting the peak data rate with 1 Tbps and the user experienced data
rate with 10 Gbps. In addition, 6G can achieve a flexible frequency sharing goal,
which can further enhance the frequency reuse efficiency.
• 6G is expected to be a personalized intelligent and visualized network. Based
on SDN, NFV, SDR, cloud computing, fog computing, and AI techniques,
6G will realize the intent-based, software-defined, flexible, and virtualized net-
working, which depends on the communication, computing, and communication
cooperation. Meanwhile, fog computing will be promising to make cloud and
edge computing adaptive to the application and networking status. As a result,
the traditional centralized 5G will be evolved into the advanced phase of
communication, computing, and communication cooperation. 6G should be data
centralized, content centralized, user centralized, and fully service centralized.
• 6G is expected to have an endogenous security, and the function security is
integrated designed. By introducing blockchain based trust and safety communi-
cation mechanisms, 6G will have the capability of self-awareness, self-analyzed,
self-optimized, self-healing, and self-protect. Both the real-time dynamic analy-
sis and the adaptive risk evaluation will be incorporated, which help realize the
space cybersecurity.
• 6G is expected to merge communication, computation, sensing, and navigation
functions together. To make sure the seamless coverage in sea and mountain,
6G will use satellite communication and make it cooperatively work with
the terrestrial land communication. Meanwhile, the satellite navigation and
positioning systems and even the radar sensing systems will be incorporated.
Based on the open and software-defined networking architecture, 6G can make
networking fast and self-intelligent development.
• 6G is expected to enhance its intelligence via collecting and deeply analyzing
these massive configurations and running data. Meanwhile, 6G can realize
everything intelligence and group collective intelligence, i.e., swarm intelligence.
As shown in Fig. 1.7, from the viewpoint of 3GPP standard organization,
according to the standard scheduling, 3GPP Release 16 mainly for NR techniques
will plan to be finalized in the early 2020, then research on 5G beyond systems
will begin from Release 17 toward Release 19. According to the scheduling, the
key technique research for 6G may be followed from Release 20 about in 2023. In
the ITU standard organization, 5G standards are expected to be formally issued at
the end of 2020, then the corresponding technology research on 6G may be started,
in which the 6G vision and technology trend will be first considered. Meanwhile,
from the viewpoint of industry, the 6G relative research has been started since 2018.
The visions, performance requirements, and key technologies have been discussed
by academics and industries all over the world since 2019. It is expected that these
works will further undergo during 2024–2026, then the standard related works for
6G will be scheduled after 2026, and the final 6G standards will be finalized toward
2030.
The challenges of 6G include system coverage, peak capacity, average user data
rate, transmit delay, movement speed, SE, and EE. 6G will develop the new air
12 1 Brief Introduction of Fog Radio Access Networks
C. Hair Wavy.
VII. 9. Indo-Afghan.
VIII. North African.
10. Arab or Semite.
11. Berber (N. Africa).
IX. Melanochroid.
12. Littoral (W. Mediterranean).
13. Ibero-insular (Spain, S. Italy).
14. Western European.
15. Adriatic (N. Italy, Balkans).
F. Hair straight.
XIII. American.
21. South American.
22. North American.
23. Central American.
24. Patagonian.
XIV. 25. Eskimo.
XV. 26. Lapp.
XVI. Eurasian.
27. Ugrian (E. Russia).
28. Turco-Tartar (S.W. Siberia).
XVII. 29. Mongol (E. Asia).
In spite of its apparent complexity, this classification coincides quite closely with the
classification which is followed in this book. Inspection reveals that Deniker’s grand division A is
Negroid, C and D Caucasian, F Mongoloid. Of his two remaining grand divisions, B is
intermediate between A and C, that is, between Negroid and Caucasian, and consists of peoples
which are either, like the East Africans, the probable result of a historical mixture of Negroids and
Caucasians, or which, like the Australians, share the traits of both, and are therefore admitted to
have a doubtful status. The other grand division, E, is transitional between Caucasian D and
Mongoloid F, and the peoples of which it consists are those whom we too have recognized as
difficult to assign positively to either stock. In short, Deniker’s classification is much the more
refined, ours the simpler; but essentially they corroborate one another.
A. Ulotrichi or Woolly-haired.
1. Lophocomi or Tuft-haired: Papua, Hottentot-Bushmen.
2. Eriocomi or Fleecy-haired: African Negroes.
B. Lissotrichi or Straight-haired.
3. Euthycomi or Stiff-haired: Australian, Malay, Mongolian, Arctic, American.
4. Euplocomi or Wavy-haired: Dravidian (S. India), Nubian, (Sudan), “Mediterranean”
(Europe, N. Africa, etc.).
The distinction here made between the Tuft and Fleecy-haired groups is unsound. It rests on a
false observation: that a few races, like the Bushmen, had their head-hair growing out of the scalp
only in spots or tufts. With the elimination of this group, its members would fall into the Fleecy or
Woolly-haired one, which would thus comprise all admitted Negroids; whereas the two remaining
groups, the Stiff and Wavy-haired, obviously correspond to the Mongoloid and Caucasian. The
only remaining peculiarity of the classification—and in this point also it is unquestionably wrong—
is the inclusion of the Australians in the Stiff or Straight-haired group. But even this error reflects
an element of truth: it emphasizes the fact that in spite of their black skins, broad noses, and
protruding jaws, the Australians are not straight-out Negroids.
The underlying feature of this classification, after allowing for its errors, is that mankind consists
of two rather than three main branches: the Ulotrichi or Negroids, as opposed to the Lissotrichi or
combined Mongoloids and Caucasians. This basic idea has been advocated by others. Boas, for
instance, reckons Mongoloids and Caucasians as at bottom only subtypes of a single stock with
which the Negroids and Australians are to be contrasted.
Somewhat different in plan is Huxley’s scheme, which recognizes four main races, or five
including a transitional one. These are (1) Australioids, including Dravidians and Egyptians; (2)
Negroids, with the Bushmen and the Oceanic Papuans, Melanesians, Tasmanians, and Negritos
as two subvarieties; (3) Mongoloids, as customarily accepted; (4) Xanthochroi, about equivalent
to Nordics and Alpines; (5) Melanochroi, nearly the same as the Mediterraneans, but supposed by
Huxley to be hybrid or intermediate between the Xanthochroi and Australioids. This classification
in effect emphasizes the connection between Australoids and Caucasians, with the Negroids as a
distinctive group on one side and the Mongoloids on the other.
Haeckel’s classification is basically similar, in that besides the usual three primary stocks—
which he elevates into species—he recognizes a separate group comprising the Australians,
Dravidians, and Vedda-like Indo-Australians.
38. Disease
Pathology might seem to promise more than normal physiology.
So far as mortality goes, there are enormous differences between
races. And the mortality is often largely the result of particular
diseases. Measles, for instance, has often been a deadly epidemic
to uncivilized peoples, and smallpox has in some regions at times
taken toll of a quarter of the population in a year or two. Yet it is
short-sighted to infer from such cases any racial predisposition or
lack of resistance. The peoples in question have been free for
generations, perhaps for their entire history, from these diseases,
and have therefore not maintained or acquired immunity. Their
difference from us is thus essentially in experience, not hereditary or
racial. This is confirmed by the fact that after a generation or two the
same epidemics that at first were so deadly to Polynesians or
American Indians sink to almost the same level of mild virulence as
they show among ourselves.
Then, too, immediate environment plays a part. The savage often
has no idea of contagion, and still less of guarding against it; he
thinks in terms of magic instead of physiology—and succumbs. How
far heavy mortality is the result of lack of resistance or of
fundamentally vicious treatment, is often hard to say. If we tried to
cure smallpox by subjecting patients to a steam-bath and then
having them plunge into a wintry river, we should perhaps look upon
the disease as a very nearly fatal one to the Caucasian race.
1890-92 1900-02
Lawyers 199 159
Physicians 102 121
Clergymen 81 91
Chimneysweeps 532 287
Brewers 190 239
Metal workers 120 137
Gardeners 88 93
All occupations 118 145
If allowance is made for the facts that the negro population of the
United States is poorer and less educated than the white; that it lives
mainly in lower latitudes; and that it tends to be rural rather than
urban, the comparative cancer death rates for the country of negro
56 and white 77 would appear to be accounted for, without bringing
race into consideration.
In short, what at first glance, or to a partisan pleader, would seem
to be a notable race difference in cancer liability, turns out so
overwhelmingly due to environmental and social causes as to leave
it doubtful whether racial heredity enters as a factor at all. This is not
an assertion that race has nothing whatever to do with the disease; it
is an assertion that in the present state of knowledge an inherent or
permanent connection between race and cancer incidence has not
been demonstrated. If there is such a connection, it is evidently a
slight one, heavily overlaid by non-racial influences; and it may be
wholly lacking.
The case would be still less certain for most other diseases, in
which environmental factors are more directly and obviously
influential. Racial medical science is not impossible; in fact it should
have an important future as a study; but its foundations are not yet
laid.