Professional Documents
Culture Documents
Textbook Mobile Networks and Management Jiankun Hu Ebook All Chapter PDF
Textbook Mobile Networks and Management Jiankun Hu Ebook All Chapter PDF
Jiankun Hu
Visit to download the full and correct content document:
https://textbookfull.com/product/mobile-networks-and-management-jiankun-hu/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
https://textbookfull.com/product/data-and-energy-integrated-
communication-networks-jie-hu/
https://textbookfull.com/product/5g-green-mobile-communication-
networks-xiaohu-ge/
https://textbookfull.com/product/principles-of-chinese-
management-1st-edition-haibo-hu/
https://textbookfull.com/product/from-gsm-to-lte-advanced-pro-
and-5g-an-introduction-to-mobile-networks-and-mobile-
broadband-4th-edition-martin-sauter/
Evolutionary Computing and Mobile Sustainable Networks
Proceedings of ICECMSN 2020 V. Suma
https://textbookfull.com/product/evolutionary-computing-and-
mobile-sustainable-networks-proceedings-of-icecmsn-2020-v-suma/
https://textbookfull.com/product/mobile-networks-and-
management-5th-international-conference-monami-2013-cork-ireland-
september-23-25-2013-revised-selected-papers-1st-edition-minji-
kim/
https://textbookfull.com/product/mobile-networks-and-
management-8th-international-conference-monami-2016-abu-dhabi-
united-arab-emirates-october-23-24-2016-proceedings-1st-edition-
ramon-aguero/
https://textbookfull.com/product/machine-learning-and-cognitive-
computing-for-mobile-communications-and-wireless-networks-1st-
edition-krishna-kant-singh/
https://textbookfull.com/product/technologies-and-eco-innovation-
towards-sustainability-ii-eco-design-assessment-and-management-
allen-h-hu/
Jiankun Hu · Ibrahim Khalil
Zahir Tari · Sheng Wen (Eds.)
235
Mobile Networks
and Management
9th International Conference, MONAMI 2017
Melbourne, Australia, December 13–15, 2017
Proceedings
123
Lecture Notes of the Institute
for Computer Sciences, Social Informatics
and Telecommunications Engineering 235
Editorial Board
Ozgur Akan
Middle East Technical University, Ankara, Turkey
Paolo Bellavista
University of Bologna, Bologna, Italy
Jiannong Cao
Hong Kong Polytechnic University, Hong Kong, Hong Kong
Geoffrey Coulson
Lancaster University, Lancaster, UK
Falko Dressler
University of Erlangen, Erlangen, Germany
Domenico Ferrari
Università Cattolica Piacenza, Piacenza, Italy
Mario Gerla
UCLA, Los Angeles, USA
Hisashi Kobayashi
Princeton University, Princeton, USA
Sergio Palazzo
University of Catania, Catania, Italy
Sartaj Sahni
University of Florida, Florida, USA
Xuemin Sherman Shen
University of Waterloo, Waterloo, Canada
Mircea Stan
University of Virginia, Charlottesville, USA
Jia Xiaohua
City University of Hong Kong, Kowloon, Hong Kong
Albert Y. Zomaya
University of Sydney, Sydney, Australia
More information about this series at http://www.springer.com/series/8197
Jiankun Hu Ibrahim Khalil
•
Mobile Networks
and Management
9th International Conference, MONAMI 2017
Melbourne, Australia, December 13–15, 2017
Proceedings
123
Editors
Jiankun Hu Zahir Tari
SEIT, UNSW Canberra RMIT University
Canberra Melbourne, VIC
Australia Australia
Ibrahim Khalil Sheng Wen
RMIT University Swinburne University of Technology
Melbourne, VIC Hawthorne, Melbourne, VIC
Australia Australia
© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer International Publishing AG
part of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
We are delighted to introduce the proceedings of the 9th European Alliance for
Innovation (EAI) International Conference on Mobile Networks and Management
(MONAMI 2017). This conference has brought together researchers, developers, and
practitioners from around the world who are leveraging and developing mobile com-
puting, wireless networking, and management.
The technical program of MONAMI 2017 consisted of 30 full papers in the main
conference track. The conference had a special session on Trust, Privacy and Security
in Internet of Things (IoT) and Cloud. Aside from the high-quality technical paper
presentations, the technical program also featured five keynote speeches. These key-
note speakers were Prof. Wanlei Zhou from Deakin University Australia, Prof. Jie Lu
from University of Technology Sydney Australia, Prof. Joe Dong from UNSW Sydney
Australia, Prof. Yongsheng Gao from Griffith University Australia, and Dr. Shui Yu
from Deakin University.
Coordination with the steering chairs, Dr. Imrich Chlamtac from Create-Net, Italy,
and Prof. Jiankun Hu and Conference General Co-Chair Prof. Yang Xiang from
Swinburne University Australia from UNSW Canberra, Australia, was essential for the
success of the conference. We sincerely appreciate their constant support and guidance.
It was also a great pleasure to work with such an excellent Organizing Committee who
worked hard in organizing and supporting the conference. In particular, the Technical
Program Committee, led by our TPC co-chairs, Dr. Zahir Tari and Dr. Ibrahim Khalil
from RMIT Australia, completed the peer-review process for the technical papers and
compiled a high-quality technical program. We are also grateful to other chairs
including the publication chair, Dr. Sheng Wen from Swinburne University of Tech-
nology, the PC members for their support, and all the authors who submitted their
papers to the MONAMI 2017 conference and workshops.
We strongly believe that MONAMI 2017 provided a good forum for all researchers,
developers, and practitioners to discuss all scientific and technological aspects that are
relevant to mobile networks and management. We also expect that future MONAMI
conferences will continue being successful and stimulating, as indicated by the con-
tributions presented in this volume.
Steering Committee
Imrich Chlamtac (Chair) EAI/CREATE-NET, Italy
Jiankun Hu UNSW Canberra, Australia
Organizing Committee
General Co-chairs
Jiankun Hu UNSW Canberra, Australia
Xiang Yang Deakin University, Melbourne, Australia
Web Chair
Naveen Chilamkurti Latrobe University, Melbourne, Australia
Workshops Chair
Ibrahim Khalil RMIT University, Melbourne, Australia
Publications Chair
Sheng Wen Deakin University, Australia
Panels Chair
Jiankun Hu UNSW, Australia
Tutorials Chair
Ibrahim Khalil RMIT University, Melbourne, Australia
VIII Conference Organization
Demos Chair
Ibrahim Khalil RMIT University, Melbourne, Australia
Local Chair
Ibrahim Khalil RMIT University, Melbourne, Australia
Honorary Co-chairs
Albert Zomaya The University of Sydney, Australia
Mohammed Atiquzzaman University of Oklahoma, USA
Hsiao-Hua Chen National Cheng Kung University, Tainan
Ernesto Damiani Khalifa University, UAE
Conference Manager
Alzbeta Mackova EAI (European Alliance for Innovation)
The Public Verifiability of Public Key Encryption with Keyword Search . . . . 299
Binrui Zhu, Jiameng Sun, Jing Qin, and Jixin Ma
1 Introduction
With the modern advancements of wireless communications, wireless networks
have seen an explosion in data traffic over the past decade [6]. This rapid demand
for more data is largely attributed to video and multimedia streaming, where it is
expected that three-fourths of data traffic will be consumed by video [6]. To com-
pound this further, it is expected that the next generation of wireless networks
will encapsulate the new paradigm of the internet of things (IoT). This concept
c ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018
J. Hu et al. (Eds.): MONAMI 2017, LNICST 235, pp. 1–13, 2018.
https://doi.org/10.1007/978-3-319-90775-8_1
2 B. Quinton and N. Aboutorab
moves to further integrate more and more devices into communication networks,
where it is foreseen that the IoT will add a further 50 billion heterogeneous wire-
less devices by 2020 [6]. Consequently, this growing demand puts further pressure
on data networks, where the offloading of the servers becomes an increasingly
important problem.
This ever-growing demand for real-time data, where users expect to maintain
their quality-of-experience (QoE) has led to much research to address the data
networks backhaul problem. Multiple areas of research have shown promising
methods to deal with this problem, one such option is to distribute the data
closer to the users with improved redundancy [2,7,8,11]. The idea of distributing
resources to the edge of a network is known as “fog” networking [5]. Motivated
by very high temporal correlation among the “popular” content demanded by
end-users, it is expected that the proactive (i.e. without users requests) diffusing
of such popular content from its storage and transmission clouds behind the
backhaul, and caching it in a “fog” of low-cost storage units close to the end-
users to serve the requests to download this content could largely improve the
network performance and service quality. Using this approach not only the users’
requests can be immediately and efficiently addressed, but also the access to the
backhaul could be significantly offloaded [10,12].
In addition to distributing the data, with the rapid increase in the number
of wireless devices, there are more and more devices in each others proximity.
Such “geographically close” wireless devices form an autonomous local network
over which the users can communicate and exchange files without contacting
the backhaul servers. Such scenario may occur for instance when co-workers are
using their tablets to share and update files stored in the cloud (e.g. Dropbox),
or when users, in the subway or a mall, are interested in watching the same
popular video. Under such scenario, the benefits of communicating over a local
network can be utilized not only to reduce the users’ download time but also
offload the backhaul of the data network (i.e., minimizing the download from it).
Furthermore, network coding (NC), initially introduced in [1], can help in
offloading of the backhaul servers in the considered distributed cooperative data
network scenario by maximizing the number of served users in one transmission,
thus maximizing the backhaul offloading. Although NC was originally imple-
mented at the network layer, more attractive application was found at the data
link layer where there is coded combinations of files to improve throughput.
Multiple areas of study have focused on various types of network coding, where
this paper will focus on opportunistic network coding (ONC) [14], in particular
instantly decodable network coding (IDNC) [13]. This technique has recently
gained much attention due to its instant decodability (as the name suggests)
by using a simple XOR operation that results in reducing the computational
complexity of the decoding at the end users. It also provides a significant benefit
to real-time communications, where studies in [3,7,13] show through a heuristic
algorithm that utilizing IDNC results in shown significant performance improve-
ments over uncoded transmissions in both centralized point-to-multipoint (PMP)
and decentralized network settings.
Offloading of Fog Data Networks 3
2 System Model
Fig. 1. An illustration depicting our system model for a distributed storage network,
showing two servers, six users, and two proximity based wireless networks, where users
can conduct cooperative D2D communications.
must satisfy U(si ) ∩ U = U. The model shows a distributed setting where the
users are in coverage of multiple servers. Also in the model, multiple proximity
networks (possibly Wi-Fi or LAN) are shown.
The proximity regions are defined
as the proximity set P = p1 , . . . , pNp with Np proximity-enabled D2D com-
munication networks. The proximity networks contain a subset of the users in U,
defined as U(pi ), that is the users in the coverage area of the proximity-enabled
network pi . It is assumed that there is no overlap of the users in each proximity
set, that is, the users in each proximity network that are “geographically close”
can communicate locally but not outside this network.
In our model, we assume the users have received some files in the initial
transmission phase1 . That is, a user ui has partially downloaded some of the files
from a transmitted frame which constitutes the users Has set Hui . Furthermore,
the remaining files wanted by user ui in the frame form the user’s Wants set,
denoted as Wui . Similarly, the servers will store a subset of the files in F, however
the union of all files at the servers should contain the complete set of F (with
possible repetition). Here, a server’s Has set is defined as Hsi . It is assumed
that the servers will maintain a global knowledge of the system state during
the initial transmissions, that is the users will respond with positive/negative
feedback depending if they receive their files successfully or not. At completion
of this phase the system will move into the recovery transmission phase.
IDNC can now be utilized to exploit users’ side data to optimize the trans-
missions in the current time epoch. In the recovery transmission, we assume an
erasure free channel, where the different users and servers will operate on orthog-
onal channels. It is also assumed that the servers have an unlimited capacity,
1
This first phase of the transmission is known as the initial transmission phase. During
the initial transmission the servers will attempt to serve all files to the users in the
network. However, some users will have received only a portion of the files requested
due to channel erasure.
Offloading of Fog Data Networks 5
Fig. 2. Conventional IDNC approach from [2], where there is a total of three downloads
required for the optimal solution.
Fig. 3. A D2D-enabled approach showing the potential for offloading servers, where
only one download is required.
such that after all cooperative D2D communications all requests remaining will
be served by the server in the current time epoch. Here, the main goal is to opti-
mize the selection of the files for network coding for the users and the servers,
where priority is given to the cooperative D2D communications, such that we
reduce the amount of downloads required from the servers.
3 Problem Formulation
Solution 1. In this solution (depicted in Fig. 2), we employ the method used in
[2], this solution utilizes a conventional IDNC approach without D2D-enabled
cooperation. One possible optimal solution using this method is:
– s1 transmits f1 to u1 .
– u1 transmits f2 ⊕ f3 to u2 and u3 .
– u4 transmits f1 to u5 and u5 transmits f4 to u1 .
In the second solution, it can be seen that we only need one download from one of
the servers, while no download is required from the other server. This approach
shows that even in a small network setting, there is a download reduction of two
thirds of the previous solution, freeing up valuable servers’ resources.
Although much work has focused on implementing IDNC in various network
models, the graph-theoretical modellings used in these cases are limited in a
cooperative full-duplex environment. In previous approaches, the graph models
incorporated assume that there is a clear differentiation between the sender and
the receiver. In our setting, we remove this restriction and allow users the ability
of full-duplex communications, therefore the existing IDNC graph models are
not appropriate and there is a need for a new model.
Generate Vertex Set. Vertices are generated from a server and user perspec-
tive under the two conditions:
2
A clique is a sub-set of the graph, where every distinct pair of vertices in the induced
subgraph are pairwise adjacent. A maximal clique is one that cannot be a subset of
a larger clique [4].
8 B. Quinton and N. Aboutorab
Fig. 4. A visualization of the IDNC induced subgraph methodology proposed for sys-
tem model shown in Fig. 1. The figure shows coding opportunities represented by edges,
transmission conflicts represented in subgraphs Kj and limitation of one transmission
per user represented in subgraph L3 (that is, u3 may only transmit to either u2 or u1
in the current time epoch).
subject to Ii ⊆ G, Ii ∩ Ψ = ∅
∃ul ∈ U where F(Ii ) ⊆ Hul
where is the disjoint set union operator and the first constraint ensures that
the independent set Ii is selected only from the vertices that belong to the
graph G. This is to ensure the selected coded file combinations that the users
serve reduces the chromatic number of the servers graph (optimal solution),
as the chromatic number of the remaining graph Ψ is equal to the number of
downloads required from the servers. The second constraint shown in (1) ensures
that for all files selected in the independent set Ii , denoted by F(Ii ), there exists
a user that posses the files and can XOR them. If this is not satisfied then the
coded transmission cannot be sent and the conditions in (1) are not met.
Solving for the optimal solution that has been presented requires that we
determine the chromatic number of a graph (equivalent to finding all maximal
cliques). It is well known that determining the chromatic number of a graph is
proven to be NP-hard [9]. This is further compounded by the fact that we not
only need to calculate the chromatic number of one graph, but we need to find
the selection of independent sets such that we minimize the chromatic number of
the remaining subgraph. Hence, the optimal solution is not applicable for online
and real-time communications. Therefore, we will propose a heuristic scheme in
the following section to solve sub-optimally.
wijk = δi j k (2)
vi j k ∈N (vijk )
4
The degree of a vertex (δ) in a graph is equal to the number of incident edges to
that vertex [4].
10 B. Quinton and N. Aboutorab
1: repeat
2: ∀vijk ∈ Gs : Compute wijk using (2)
∗
3: vijk ⇐= argmaxvijk wijk , ∀vijk ∈ Gs
∗
4: add vijk to Γ
∗
5: Gs ← Gs ∪ vijk
6: until Gs = ∅
7: return the clique listing Γ
where N (vijk ) is the set of adjacent vertices to vijk . Therefore, each vertex
in graph G will have a large weighting if it has a large number of adjacent
vertices, which themselves have a large number of adjacent vertices. The search
will then select the vertex with the largest weighting, or between those with the
same largest weight with equal probability. The algorithm then removes all non-
adjacent vertices to vijk , and then checks if the vertex vijk belongs to a subgraph
Kj or Li and will remove all other vertices that are a member of either subgraph.
Secondly, the algorithm will then update all weights in G before selecting the
next (if any) adjacent vertex in graph G that forms a clique with all previously
selected vertices. The algorithm then continues to iterate these steps until no
more vertices can be added to the clique. Finally, once a maximal clique listing is
found and removed, we iterate the whole procedure until no more vertices are left
in the graph G. The steps of algorithm described is summarized in Algorithm 1.
At this stage, the algorithm has removed all possible D2D cooperations avail-
able in hopes to minimize the amount of downloads from the servers. Therefore,
we now need to serve the remaining vertices in graph Ψ that were not served
locally from D2D cooperation. We now conduct the exact same procedure on
the remaining vertices in graph Ψ , where each maximal clique represents one
download from a server and continue until all vertices are removed from the
graph. Once all vertices have been removed from the graph Ψ the system will
have reached absorption, that is, all users will have received the file in their
Wants sets.
5 Simulation Results
In this section, we present our simulation results for the proposed algorithm
in a cooperative D2D setting in comparison with a uncooperative decentralized
conflict free IDNC approach that was incorporated in [2]. In both cases, the aim
of the approaches is to reduce the number downloads from the servers.
Offloading of Fog Data Networks 11
20
Uncooperative IDNC
18 Cooperative IDNC
14
12
10
0
10 20 30 40 50 60 70 80 90 100
Number of Users
Fig. 5. The average number of downloads required from the servers as a function of
the number of users.
In the simulations, each user is interested in receiving one file and has two
files already received and stored in the Has set (when fixed), where the recovery
downloads are to be completed in one time epoch. We assume each users’ Has
and Wants sets to be determined probabilistically, with uniform distribution
over all files in the library. In all simulations, there are two servers available
with total coverage of all users in the network, while in the cooperative model
we consider a dual network where the users are split evenly between the two
proximity networks p1 and p2 (similar to Fig. 1).
Firstly in Fig. 5, we show the average number of downloads required from the
servers for a fixed number of files in the transmission frame of Nf = 20, as a func-
tion of the number of users Nu . The result shows that for the algorithm imple-
mented for a cooperative D2D-enabled setting, as the number of users increase
the average number of server downloads tends to monotonically decrease. Intu-
itively, this result is expected as more users in the network will result in a greater
likelihood that the users can serve themselves independently from the servers, as
the collective Has set of the users in the network will cover the files in the frame
F. Additionally, it can be seen that in comparison to a conventional uncoopera-
tive conflict-free IDNC approach, as the network size increases there is significant
improvement, where we see an improvement of approximately 550% with only
20 devices in the network setting. Furthermore, approximately no downloads
from the servers are required as the number of users approach 60 in this network
setting, that is, 30 users in each D2D-enabled network.
Now if we consider fixing the number of users to 20, while varying the amount
of files per transmission frame, we can see the results in Fig. 6 for cooperative
versus uncooperative IDNC transmission schemes. In both cases, it can be seen
as the number of files increase, both schemes show a similar increase on the
12 B. Quinton and N. Aboutorab
20
Uncooperative IDNC
18 Cooperative IDNC
14
12
10
0
10 20 30 40 50 60 70 80 90 100
Number of Files
Fig. 6. The average number of downloads required from the servers as a function of
the number of files.
number of downloads required from the servers. Although the two schemes tend
to converge if we consider an asymptotic limit, the cooperative scheme still shows
reasonable improvement of approximately 50% for up to 100 files. Again, this
result is expected as increasing the number of files in a frame reduces the poten-
tial to leverage a coded transmission. Additionally, as the number of files increase
the likelihood of a users ability to diffuse the wanted packets is diminished. Nev-
ertheless, the cooperative approach still shows significant ability to reduce the
number of downloads required from the network servers.
6 Conclusion
References
1. Ahlswede, R., Cai, N., Li, S.-Y.R., Yeung, R.W.: Network information flow. IEEE
Trans. Inf. Theory 46(4), 1204–1216 (2000)
2. Al-Habob, A.A., Sorour, S., Aboutorab, N., Sadeghi, P.: Conflict free network
coding for distributed storage networks. In: 2015 IEEE International Conference
on Communications (ICC), pp. 5517–5522. IEEE (2015)
3. Baran, P.: On distributed communications networks. IEEE Trans. Commun. Syst.
12(1), 1–9 (1964)
4. Bondy, J.A., Murty, U.S.R.: Graph Theory with Applications, vol. 290. Macmillan,
London (1976)
5. Bonomi, F., Milito, R., Zhu, J., Addepalli, S.: Fog computing and its role in the
internet of things. In: Proceedings of 1st Edition of the MCC Workshop on Mobile
Cloud Computing, pp. 13–16. ACM (2012)
6. Cisco: Cisco visual networking index: global mobile data traffic forecast update.
Technical report, February 2016
7. Dimakis, A.G., Godfrey, P.B., Wu, Y., Wainwright, M.J., Ramchandran, K.: Net-
work coding for distributed storage systems. IEEE Trans. Inf. Theory 56(9), 4539–
4551 (2010)
8. Dimakis, A.G., Ramachandran, K., Wu, Y., Suh, C.: A survey on network codes
for distributed storage. Proc. IEEE 99(3), 476–489 (2011)
9. Edwards, C.S., Elphick, C.H.: Lower bounds for the clique and the chromatic
numbers of a graph. Discret. Appl. Math. 5(1), 51–64 (1983)
10. Golrezaei, N., Molisch, A., Dimakis, A.G., Caire, G.: Femtocaching and device-
to-device collaboration: a new architecture for wireless video distribution. IEEE
Commun. Mag. 51(4), 142–149 (2013)
11. Papailiopoulos, D.S., Luo, J., Dimakis, A.G., Huang, C., Li, J.: Simple regenerating
codes: network coding for cloud storage. In: 2012 Proceedings of IEEE INFOCOM,
pp. 2801–2805. IEEE (2012)
12. Shanmugam, K., Golrezaei, N., Dimakis, A.G., Molisch, A., Caire, G.: Fem-
toCaching: wireless content delivery through distributed caching helpers. IEEE
Trans. Inf. Theory 59(12), 8402–8413 (2013)
13. Sorour, S., Valaee, S.: On minimizing broadcast completion delay for instantly
decodable network coding. In: 2010 IEEE International Conference on Communi-
cations (ICC), pp. 1–5. IEEE (2010)
14. Sorour, S., Valaee, S.: An adaptive network coded retransmission scheme for single-
hop wireless multicast broadcast services. IEEE/ACM Trans. Netw. (TON) 19(3),
869–878 (2011)
Persistent vs Service IDs in Android:
Session Fingerprinting from Apps
1 Introduction
Mobile devices, especially smartphones have become an indispensable part of our
daily lives, as a big part of our communications and daily activities is processed
and monitored by them. One of the main reasons for their wide adoption is that
they have a plethora of embedded sensors that allow them to understand their
context and adapt accordingly. For instance, through luminosity and proximity
sensors as well as accelerometers, mobile phones may adapt the UI to fit better to
user expectations. Moreover, thanks to GPS, mobile devices are location aware
enabling them to render content according to the spacial restrictions significantly
improving the user recommendations.
Data mining and data profiling can be used in order to collect valuable infor-
mation about a particular user or group of users, in order to generate a profile
[12], which can be further used by companies to gain profit. As stated in [14],
this kind of information, namely user profiling, is valuable also for advertisers
who want to target ads to their users and in return, advertisers may pay more
money to their hosting applications’ developers. Building user profiles requires,
as the authors state, sensitive privileges in terms of permissions, such as Internet
access, location, or even retrieving installed applications in a user’s device [14].
To this end, we may infer that collecting and successfully fusing user data from
more than one service can create even better and more complete user profiles,
which will consequently translate in higher monetization. Looking back in 2009,
it was quite clear that:
c ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018
J. Hu et al. (Eds.): MONAMI 2017, LNICST 235, pp. 14–29, 2018.
https://doi.org/10.1007/978-3-319-90775-8_2
Session Fingerprinting from Apps 15
The above has been realised by tech giants. For instance quoting a statement
from Google’s current privacy policy [10]:
“We may combine personal information from one service with information,
including personal information, from other Google services - for example
to make it easier to share things with people you know. Depending on your
account settings, your activity on other sites and apps may be associated
with your personal information in order to improve Google’s services and
the ads delivered by Google”
In order to “enable” data fusion from different sources and services, one could
argue that unique identifiers should be either implicitly or explicitly present. In
particular, a value describing a quantity of some valuable variable might be
useless if it is not accompanied by a unique identifier that would allow us to
track its source. Contrariwise, identifiers coming from different services that
are matched, may act as a “bridge” between these services to combine their
corresponding datasets and integrate them.
During the last decade relevant surveys have revealed that the majority of
both iOS and Android apps were transmitting the phone’s unique ID and the
user’s location to advertisers. These findings are confirmed by “The Haystack
Project” [11] which revealed that nearly 70% of all Android apps leak personal
data to third-party services such as analytics services and ad networks [20].
All this wealth of information apart from the benign usage for the user benefit
has been a constant target by companies who wish to monetize it, mainly through
targeted advertisement. The recent advances in big data and data mining have
enabled the extraction of information from theoretically diverse data, leading to
the revealing of a lot of sensitive data through data fusion. To this end, many
fingerprinting techniques have been introduced in order to link data flows to
specific individuals. Apparently, since Android in currently the prevailing mobile
platform, most companies are targeting it with their apps, under the freemium
model, harvesting user data to monetize them. There is even a common saying
in the privacy community suggesting that “If you’re not paying for the product,
you are the product”. To this end, If users are not paying for an app, they are
usually selling their profiles (with or without their knowledge/consent) to an ad
network, which will use their unique identifiers to track and target them.
In view of the above and targeting at improving the OSes privacy, the new
coming Android O, makes a number of privacy-related changes to the platform,
many of which are related to how unique identifiers are handled by the system [9],
and in particular aiming to help provide user control over the use of identifiers
[2]. One of the most important improvements concern “limiting the use of device-
scoped identifiers that are not resettable”.
16 E. Alepis and C. Patsakis
2 Related Work
Unique identifiers have been used for a long time and facilitate many tasks in
modern database systems as they allow us to perform record linkage between
Session Fingerprinting from Apps 17
different entities and extract the necessary information and thus knowledge from
the corresponding database tables. The most typical example of a unique iden-
tifier is the Social Security Number, which allows us to distinguish two people
from each other. However, in the digital era, unique identifiers can be considered
hardware identifiers like the MAC address of the network card, or a set of prop-
erties such as browser fingerprints which consist among others of the browser
version, OS, fonts, and browser plugins.
In the Android ecosystem there is a plethora of unique identifiers which have
so far been extensively exploited by advertisement companies to track users and
their interests as ad libraries have become more and more greedy and rogue
[3,18] while apps may deliberately leak information to the ads [4,21] harness-
ing arbitrary amounts of users’ sensitive information directly or indirectly [5].
A key role in this procedure is the use of unique identifiers [17]. Acknowledg-
ing this situation, Google initially introduced some recommendation guidelines
for the proper use of unique identifiers in Android [1]. Then, Google gradually
started requesting more permissions from the apps to allow them access to these
identifiers. For instance, a typical unique identifier for mobile phones are IMEI
and IMSI, however, after Marshmallow, the user has to grant the dangerous
READ PHONE STATE permission to an app to access them. While many users may
ignore app permissions [8], for many others it works as an obstacle, forcing many
companies to comply with the rule.
Despite the ads, apps may collaborate in order to perform malicious acts
which independently would not be allowed to perform. Orthacker et al. [15]
study this problem from the aspect of permissions. In this regard, the malicious
apps which are installed in the victim’s device may result in “possessing” and
correspondingly using dangerous permissions that other normal apps do not. The
concept is that the user would not allow camera and microphone permission to a
single app. However, since the permissions are requested by two apps which are
seemingly independent, the permissions are “spread” so the user grants them,
yet an adversary controls both of them getting access to the desired resource.
Contrary to Orthacker et al. we do not aim to resources, but access to information
that the user would not share to one specific app to prevent his profiling.
In Nougat, the current stable version of Android, Google prohibited unpriv-
ileged access to even more hardware identifiers, such as the MAC address of
the WiFi card, by restricting access to /proc. While the latter measure creates
many issues with applications targeting towards security and privacy services
as Google has not provided any permission so far to access this information,
undoubtedly, it leaves little space to adversaries to exploit.
REFERENCES.
[1] Pliny: lxxxiii., 11, N.c.v.
[2] Stockhusen: De Litharg. Fumo, etc. Goslar, 1656.
[3] Tronchin: De Colica Pictonum. 1758.
[4] John Hunter: Observations of Diseases of the Army in Jamaica.
London, 1788.
[5] Meillère, G.: Le Saturnisme. Paris, 1903.
[6] Bisserie: Bull. Soc. Pharmacol. May, 1900.
[7] Houston: Local Government Board Annual Report, 1901-02,
supplement, vol. ii.
CHAPTER II
ÆTIOLOGY
Lead poisoning of industrial origin rarely occurs in the acute form.
Practically all cases coming under the notice of either appointed surgeons,
certifying surgeons, or even in the wards of general hospitals, are of the
subacute or chronic type. There is no reason to suppose that lead
compounds are used more frequently by the workers in lead industries as
abortifacients than by other persons.
The compounds of lead which are responsible for poisoning in industrial
processes are for the most part the hydrated carbonate, or white lead, and
the oxides of lead, whilst a comparatively small number of cases owe their
origin to compounds, such as chromates and chlorides.
The poisonous nature of any lead compound from an industrial point of
view is proportional to (1) the size of the ultimate particles of the substance
manufactured, and therefore the ease with which such particles are capable
of dissemination in the air; and (2) the solubility of the particles in the normal
fluids of the body, such as the saliva, pharyngeal and tracheal and bronchial
mucus, etc., and the fluids of the stomach and intestine. An instance of the
variation in size of the particles of lead compounds used industrially is the
difference between ground lead silicate (fritted lead) used in the potteries,
and the size of the particles of ordinary white or “raw” lead. By micrometric
measurements one of us [K. W. G.[1]] found the average size of the particles
of fritt to be ten times that of the white lead particles. Further, direct
experiment made with equal masses of the two compounds in such a manner
that the rate of settling of the dust arising could be directly compared in a
beam of parallel light showed presence of dust in the white lead chamber
fifteen minutes after the fritt chamber was entirely clear. It is found as a matter
of practice that where dust is especially created, and where it is difficult to
remove such dust by exhaust fans, the greatest incidence of lead poisoning
occurs. The association of dusty processes and incidence of lead poisoning is
discussed in relation to the various trades in Chapters XV. to XVII. Fume and
vapour given off from the molten metal or compounds, such as chlorides
(tinning), are only a special case of dust.
The channels through which lead or its compounds may gain entrance to
the animal body are theoretically three in number:
1. Respiratory tract.
2. Gastro-intestinal.
3. Cutaneous.
For many years most authorities have held that industrial poisoning by
means of compounds of lead takes place directly through the alimentary
canal, and that the poison is conveyed to the mouth mainly by unwashed
hands, by food contaminated with lead dust, and by lead dust suspended in
the air becoming deposited upon the mucous membrane of the mouth and
pharynx, and then swallowed. As evidence that lead dust is swallowed, the
classical symptom of colic in lead poisoning has been adduced, on the
supposition, in the absence of any experimental proof, that the lead
swallowed acted as an irritant on the gastro-intestinal canal, thus causing
colic, and, on absorption from the canal, setting up other general symptoms.
Much of the early treatment of lead poisoning is based upon this assumption,
and the administration of sulphuric acid lemonade and the exhibition of
sulphate of magnesia and other similar compounds as treatment is further
evidence of the view that the poisoning was considered primarily intestinal.
One of the chief objections to this view, apart from the experimental
evidence, is that in those trades where metallic lead is handled, particularly
lead rolling, very few hygienic precautions have ever been taken in regard to
washing before meals, smoking, etc. Although in these trades the hands
become coated with a lead compound (oleate), and the workers frequently
eat their food with unwashed hands, thus affording every opportunity for the
ingestion of lead, the incidence of poisoning is by no means as high or so
pronounced in these occupations as in those giving rise to lead dust, such as
the white lead industry, where special precautions are taken, and where the
incidence of poisoning is always related to the dust breathed.
Respiratory Tract.—In a report on the incidence of lead poisoning in the
manufacture of paints and colours, one of us [T. M. L.[2]] in 1902 laid stress on
the marked incidence of poisoning in the specially dusty lead processes.
Following on that report special attention was given to the removal of dust by
means of exhaust ventilation. With the introduction of precautionary
measures, the incidence of poisoning underwent a marked decrease, this
decrease being most definite in those industries where efficient exhaust
ventilation could be maintained (see p. 47). Experience shows that cases of
poisoning in any given trade or manufacturing process are always referable to
the operations which cause the greatest amount of dust, and where,
therefore, the opportunity of inhaling lead dust is greatest.
The investigations of Duckering[3], referred to on p. 203, show the amount
of dust present in the air in certain dangerous processes. His results clinch
the deductions made from general observation, that dusty processes are
those especially related to incidence of industrial poisoning. Ætiologically,
therefore, the relationship of dust-contaminated air and poisoning is
undeniable, and in not a few instances on record persons residing at a
distance from a lead factory have developed poisoning, although not
employed in any occupation involving contact with lead, aerial infection
through dust remaining the only explanation. The actual channel through
which the lead dust suspended in the air gains entrance to the body is,
therefore, of especial importance; one of two channels is open—gastro-
intestinal and respiratory.
The investigations of one of us (K. W. G.) on the experimental production of
lead poisoning in animals has shown conclusively that the dust inhaled was
far more dangerous, and produced symptoms far earlier than did the direct
ingestion of a very much larger quantity of the same compound by way of the
mouth and gastro-intestinal canal. There is no doubt whatever that the chief
agent in causing lead poisoning is dust or fume suspended in the air. That a
certain amount finds its way into the stomach direct is not denied, but from
experimental evidence we consider the lung rather than the stomach to be
the chief channel through which absorption takes place (see p. 81).
The following table gives a specific instance of the incidence of lead
poisoning in a white lead factory, and demonstrates clearly the ætiological
importance of dust. The increase in reported cases, as well as in symptoms
of lead absorption not sufficiently severe to prevent the individual from
following his usual occupation, was associated with the rebuilding of a portion
of the factory in which the packing of dry white lead had been carried on for a
large number of years. The alterations necessitated the removal of several
floors, all of which were thoroughly impregnated with lead dust. Before the
alterations were undertaken it was recognized that considerable danger
would arise; stringent precautions were therefore taken, and the hands
engaged in the alterations kept under special observation. Notwithstanding
this there was an increase in the number of reported cases, which were all
mild cases of colic; all recovered, and were able to return to their work in a
short time.
Table I.—Lead Poisoning in a White Lead Factory.
The figures refer to the weekly examination of the whole of the men. For example, if a man
was returned as suffering from anæmia on three occasions, he appears as three cases in
Column 7.
Year Total Total Cases Cases Cases of Cases of Cases of Blue Line
Number Cases of in Dusty in Other Suspen- Anæmia Tremor
of Poisoning Processes Processes sion
Exami-
nations
(1) (2) (3) (4) (5) (6) (7) (8) (9)
1905 5,464 9 8 1 20 78[B] 249[B] 311[B]
1906 [A] 5,096 18 16 2 9 256 215 532
1907 4,303 4 3 1 6 62 81 38
1908 3,965 4 3 1 5 40 25 11
[A] Structural alterations in progress, including cutting up “lead floor,” saturated with
white lead dust.
[B] These numbers for the half-year only, the inspection being taken over in June,
1905.
In the second digestion, in which the analysis of the contents showed the
patient to be suffering from the condition known as “hyperhydrochloridia,” the
results were—
Lead sulphate 0·046 per cent.
White lead 0·042 „
Litharge 0·340 „
A very large number of experiments have also been performed for the
purpose of determining the solubility of raw lead glaze, and white lead, in
artificial digestions, the digestions having been made up in such a way that
they resembled as far as possible in every particular the ordinary stomach
contents. The type of digestion used was as follows:
Dry breadcrumbs 140 grammes.
Hydrochloric acid 5 c.c.
Lactic acid 0·1 c.c.
Acetic acid 0·1 c.c.
Pepsin 1·2 grammes.
Milk 1,200 c.c.
Digestions were performed with this mixture, and in every case the digest
was divided into two portions; each portion was retained at body temperature,
with agitation for a couple of hours, and at the end of that time one portion
was submitted to analysis. The second portion was neutralized, sodium
carbonate and pancreatic ferment added, and digestion carried on for another
two and a half hours at body temperature. At the end of this time the
pancreatic digest was examined.
Thirty-five digestions were performed. When 1 gramme of white lead was
used—that is, 0·01 per cent., containing 0·75 per cent. of lead oxide—the
quantity of lead found as lead oxide in the acid digest varied from 2 to 3 per
cent., whilst the amount found in the pancreatic digest varied from 4 to 6·5
per cent. of the added salt. On increasing the amount to 12 grammes—that
is, 1 per cent.—the quantity returned in the digest only increased from 1·5 to
2 per cent. In other words, in the addition of larger quantities of material the
ratio of solubility did not rise in proportion to the quantity added. Where a
direct pancreatic digestion was performed without the preliminary digest of
the gastric contents, the amount of lead present in the digest was only about
0·2 per cent. of the quantity added; indeed, it was very much smaller than the
amount dissolved out after preliminary acid digestion—that is, if the normal
sequence of digestion is followed, the solubility progresses after the gastric
digest has been neutralized and pancreatic ferment has been added,
whereas very slow action indeed occurs as the result of action of the
pancreatic digest alone. Some experiments described by Thomason[12],
although carried out without special regard to the physiological question of
the progressive nature of digestion, distinctly confirm the point raised. Thus,
in a digest of gastric juice, milk, and bread, 5·0 per cent. of lead was
dissolved, whereas when pancreatic juice alone was used only 0·4 per cent.
was found to be dissolved, a remarkable confirmation of the point under
discussion.
The difficulty of estimating lead present in these gastric digestions is a very
real one, as, owing to the precipitation of lead by various fluids of an
albuminoid nature, it is difficult to determine the amount of lead present in a
given quantity of digest; moreover, in making such a digest, much of the
material may become entangled among the clot of the milk in a purely
mechanical fashion, and, in attempting to separate the fluid from the other
portion of the digest, filtration no doubt removes any lead which has been
rendered soluble first of all, and reprecipitated as an albuminate. An
albuminate of lead may be formed with great ease in the following way: A 5
per cent. solution of albumin in normal saline is taken, 0·02 per cent. of
hydrochloric acid is added, and 10 per cent. solution of lead chloride added
as long as a precipitate is formed. The precipitate is then filtered off, and
washed in a dialyser with acidulated water until no further trace of lead is
found in the washings. A portion of this substance taken up in distilled water
forms a solution of an opalescent nature, which readily passes through the
filter and gives the reaction of protein with Millon’s reagent, and the lead
reaction by means of caustic potash and sulphuretted hydrogen, but very
large quantities of mineral acid are required to produce any colour with
hydrogen sulphide. Lead which gains access to the stomach, either dissolved
in water or swallowed as fine dust, becomes in all probability converted first
into a soluble substance, chloride, acetate, or lactate, which compound is
then precipitated either by the mucin present in the stomach, or by the protein
constituents of the food, or by the partially digested food (peptonate of lead
may be formed in the same way as the albuminate described above). In this
form, or as an albuminate or other organic compound, it passes the pylorus,
and becomes reprecipitated and redigested through the action of the
pancreatic juice. A consideration of the action of artificial gastric juices and
the properly combined experiments of gastric and pancreatic digestions
suggest that the form in which lead becomes absorbed is not a chloride, but
an organic compound first formed and gradually decomposed during the
normal process of digestion, and absorbed in this manner from the intestine
along with the ordinary constituents of food. Dixon Mann[13] has shown that
about two-thirds of the lead administered by the mouth is discharged in the
fæces, and that the remaining one-third is also slowly but only partially
eliminated. This point is of very considerable importance in relation to
industrial poisoning of presumably gastro-intestinal origin, and consideration
of the experiments quoted suggests that the digestion of albuminate or
peptonate may to some extent be the basis which determines the excretion of
so much of the lead via the fæces. This alteration of solubility has no doubt a
bearing on the immunity exhibited by many animals when fed with lead, and
probably explains the fact that many of the experimental animals fed with lead
over long periods exhibited no symptoms of poisoning (see p. 85), whereas
control animals, given a far smaller quantity of lead by other means and
through the lung, rapidly developed symptoms of poisoning. A diversity of
opinion exists as to the effect of pepsin upon the solubility of lead. Oliver[14]
considers that the pepsin has a retarding influence on the solubility of lead in
the gastric juice, and Thomason’s experiments also support this view,
although it is difficult to see why the action of pepsin alone should be of such
extreme importance. There is also the complicating fact that other added
substances in the food may mask any direct pepsin factor that may be
present. Albumose and peptone rather than pepsin are to be regarded as the
more important substance physiologically in their reaction with lead, and it is
interesting to note that Schicksal[15] found that by exposing lead in the form of
white lead in a 1 per mille solution of hydrochloric acid in the presence of
peptone produced a greater solvent effect on white lead than did the diluted
acid alone, and the same effect was also seen on metallic lead.
Table II.—Schicksal’s Table.
Amount
dissolved
returned
as
Metallic
Solution. Substance. Time. Lead.
(a) 1·0 per cent. peptone 100 White lead, 10 3 days at 37°
- 0·1471 grm.
0·1 per cent. HCl c.c. grms. C.
REFERENCES.
[1] Goadby, K. W.: A Note on Experimental Lead Poisoning. Journal of Hygiene, vol.
ix., No. 1, April, 1909.
[2] Legge, T. M.: Report on the Manufacture of Paints and Colours containing Lead
(Cd. 2466). 1905.
[3] Duckering, G. E.: Journal of Hygiene, vol. viii., No. 4, September.
[4] Meillère, G.: Le Saturnisme, chap. iv. Paris, 1903.
[5] Armit, H. W.: Journal of Hygiene, vol. viii., No. 5, November, 1908.
[6] Tanquerel des Planches: Traité des Maladies de Plomb, ou Saturnines. Paris,
1839.
[7] Stanski: Loc. cit.
[8] Gautier: Intoxication Saturnine, etc. Académie de Médecine, viii., November,
1883.
[9] Thresh, J. C.: The Lancet, p. 1033, October 7, 1905.
[10] Ibid., January 5, 1909.
[11] Thomason: Report of the Departmental Committee on Lead Manufacture:
Earthenware, China, vol. ii., appendices, p. 61. 1910.
[12] Ibid.
[13] Dixon Mann: Forensic Medicine and Toxicology, p. 495. 1908.
[14] Oliver, Sir T.: Lead Poisoning (Goulstonian lectures). 1891.
[15] Schicksal: Die Bekämpfung der Bleigefahr in der Industrie, p. 38. 1908.
[16] Steinberg: International Congress of Industrial Hygiene. Brussels, 1910.
[17] Cloetta: Dixon Mann’s Forensic Medicine and Toxicology, p. 463.
[18] Little: The Lancet, March 3, 1906.
[19] Canuet, T.: Thèse, Paris, 1825, No. 202. Essai sur le Plomb.
[20] Drouet: Thèse, Paris, 1875. Recherches Experimentales sur le Rôle de
l’Absorption Cutanée dans la Paralysie Saturnine.
[21] Manouvrier, A.: Thèse, Paris, 1873, No. 471. Intoxication par Absorption
Cutanée.
CHAPTER III
SUSCEPTIBILITY AND IMMUNITY
A large number of poisonous substances, among which lead may
be included, are not equally poisonous in the same dose for all
persons. It is customary to speak of those persons who show a
diminished resistance, or whose tissues show little power of resisting
the poisonous effects of such substances, as susceptible. On the
other hand, it is possible, but not scientifically correct, to speak of
immunity to such poisonous substances. Persons, particularly, who
resist lead poisoning to a greater degree than their fellows are better
spoken of as tolerant of the poisonous effects than as being partially
immune.
The degree of resistance exhibited by any given population
towards the poisonous influence of lead shows considerable
variation. Thus, in a community using a water-supply contaminated
with lead, only a small proportion of the persons drinking the water
becomes poisoned. There are, of course, other factors than that of
individual idiosyncrasy which may determine the effect of the poison,
as, for example, the drawing of the water first thing in the morning
which has been standing in a particular pipe. But even if all
disturbing factors are eliminated in water-borne lead poisoning,
differing degrees of susceptibility are always to be observed among
the persons using the water.
Lead does not differ, therefore, from any other drugs to which
persons show marked idiosyncrasies. Thus, very small doses of
arsenic may produce symptoms of colic in susceptible persons; a
limited number of individuals are highly susceptible to some drugs,
such as cannabis indica, while others are able to ingest large doses
without exhibiting any sign of poisoning; and it is well known that
even in susceptible persons the quantity of a particular drug which
first produces symptoms of poisoning may be gradually increased, if
the dosage be continued over long periods in quantities insufficient