=Paper=
{{Paper
|id=Vol-3642/paper11
|storemode=property
|title=Reviewing Recent Literature on IoT-Based System-of-Systems: A Bibliometric Analysis
|pdfUrl=https://ceur-ws.org/Vol-3642/paper11.pdf
|volume=Vol-3642
|authors=Aymen Abdelmoumen,Zakaria Benzadri,Ismael Bouassida Rodriguez
|dblpUrl=https://dblp.org/rec/conf/tacc/AbdelmoumenBR23
}}
==Reviewing Recent Literature on IoT-Based System-of-Systems: A Bibliometric Analysis
==
https://ceur-ws.org/Vol-3642/paper11.pdf
Reviewing Recent Literature on IoT-Based
System-of-Systems: A Bibliometric Analysis
Aymen Abdelmoumen1,* , Zakaria Benzadri1 and Ismael Bouassida Rodriguez2
1
University of Constantine 2 – Abdelhamid Mehri, LIRE Laboratory, Ali Mendjeli B.P. 67A, Constantine, 25016, Algeria
2
ReDCAD Laboratory, ENIS, University of Sfax, Tunisia
Abstract
Emerging System-of-Systems incorporating Internet-of-Things devices is the topic of ever-increasing
interest. Ranging from miniature devices (e.g., smart watch) to metropolis-wide infrastructures (e.g.,
smart cities), its influence is affecting our daily lives on all levels. This paper presents a bibliometric
analysis focused on the integration of systems-of-systems (SoS) in Internet-of-Things (IoT) platforms.
IEEExplore was the source of all the data retrieved for this study. By analyzing the relevant literature,
this paper aims to provide insight into the current state-of-the-art on the integration of IoT and SoS
interchangeably. Python was the main bibliometric tool used for conducting statistical analysis and
visualizing the collected data. The results revealed strong correlations between groups of authors as
well as co-occurrence of IoT-related terms. The data are displayed in sorted tabular form, graphs and
networks for comprehensiveness and conciseness. Subsequent findings may help contribute to a better
understanding of the field and inform future research directions.
Keywords
Bibliometric Analysis, Systematic Review, IEEExplore, Internet-of-Things, Cyber-physical Systems,
System-of-Systems, System-of-Systems-Engineering
1. Introduction
The most intuitively compatible application for systems-of-systems (SoS) is their integration
within Internet of Things (IoT) platforms. This convergence facilitates the seamless integration
of the adaptable and intelligent aspects of IoT devices with the high-level coordination and
collective behaviors inherent in Systems-of-Systems (SoS). As a result, it has the potential
to enhance overall efficiency and performance, optimize decision-making processes, elevate
situational awareness, and streamline resource management [1].
However, it’s worth noting that both SoS and IoT exhibit distinctive characteristics that can
present intricate design and practical challenges, which may potentially hinder the full realiza-
tion of their combined benefits [2]. Several well-established challenges encountered by system
designers include the need to describe and manage the constant evolution of heterogeneous
systems, as well as the identification of unforeseen behaviors arising from the interactions
TACC 2023: Tunisian-Algerian Joint Conference on Applied Computing, November 6 - 8, Sousse, Tunisia
*
Corresponding author.
$ aymen.abdelmoumen@univ-constantine2.dz (A. Abdelmoumen); zakaria.benzadri@univ-constantine2.dz
(Z. Benzadri); bouassida@redcad.org (I. Bouassida Rodriguez)
0000-0001-8177-9885 (A. Abdelmoumen); 0000-0002-5605-7415 (I. Bouassida Rodriguez)
© 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
Workshop
Proceedings
http://ceur-ws.org
ISSN 1613-0073
CEUR Workshop Proceedings (CEUR-WS.org)
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
among these systems [3]. These challenges encompass issues of interoperability, scalability,
heterogeneity, complexity, and security.
Numerous approaches have been explored to tackle their inherently intricate characteris-
tics [2] [3]. These efforts have predominantly encompassed hybrid design strategies [4] [5],
pragmatic solutions [6], diverse tools and frameworks [7] [8], all aimed at bridging the di-
vide between theory and practice. These solutions have found their way into a wide array of
real-world application domains, spanning various stages of the engineering process, such as
conceptual modeling [9] [10], practical design [11] [12], simulation [13] [14], post-deployment
quality assessment and validation [15] [16] [17].
In this paper, we have embraced a highly systematic approach to delve into recent advance-
ments stemming from practices and modeling techniques concerning Systems-of-Systems and
the Internet-of-Things. The method we employed involves a "bibliometric analysis," akin to
a survey, which entails the quantitative evaluation of the scholarly excellence of journals or
authors through statistical methods and quality metrics such as citation impact, citation counts,
journal rankings, h-index, and other pertinent indicators within the research domain [18]
[19]. Furthermore, this analysis entails the careful selection of a suitable empirical foundation,
encompassing a set of journals, authors, or publications relevant to the study.
This approach adeptly accommodates comprehensive data gathering and interpretation, ensur-
ing the provision of insightful perspectives on the subject under investigation. Simultaneously,
it evaluates impacts while mitigating potential subjective biases [20].
The structure of this paper is delineated as follows. Commencing with Section 2, we outline
our preliminary plan for the research, motivating the reasoning behind our choice of tools and
data sources that guide this study. This is followed by Section 3, which offers a detailed exposition
of the adopted analysis methodology, offering readers a comprehensive understanding of our
procedural approach. In Section 4, we present our findings, displaying the results we gathered
from the specified bibliometric indicators. These results are thoroughly discussed in section 5,
acknowledging any shortcomings the study suffered from. Conclusively, section 6 summarizes
the findings and highlights potential directions and perspectives for future research.
2. Research Design
In this section, we present a summarized rundown on the strategies undertaken for realizing
the outlined objectives. The adopted approach is a hybrid formulation combining conventional
bibliometric methodology [21] and innovating our own. The integral process is partitioned into
2 distinct yet complementary phases; 1. ‘Research Design’ and 2. ‘Realization’, accentuating
the strategic alignment between the conceptual and practical stages. This section will host the
‘design’ phase in detail. As for the ‘practical’, it will be explained in section 3. Figure 1 briefly
illustrates the steps of the approach:
Each stage of this plan contributes to a coherent and impactful research study:
1. Defining the Research Questions: This initial step involves formulating clear and
concise research questions that guide the entire bibliometric analysis [22].
2. Selecting Data Sources: In this phase, an appropriate data source that aligns with the
context of the research is chosen.
Figure 1: Research Design and Realization
3. Choosing Bibliometric Indicators: These indicators are identified and underlined early
in the process as to answer to each of the analysis’ objectives.
4. Collecting Data: this step comprises of carefully choosing the relevant keywords and
retrieving back the resulting sets of data.
5. Planning Data Filters: Establishing a clear plan for preprocessing data is key step that
involves cleaning, formatting, and organizing the collected data.
6. Applying the Analysis: the cornerstone of our approach is to derive insights from the
collected data. This involves selecting appropriate statistical methods, data visualization
techniques, and algorithms tailored to the research objectives and the chosen bibliometric
indicators. (This step is separate from reporting any results [23]).
2.1. Research Questions
In this section of the article, we will be giving direction and purpose to our bibliometric review
by enumerating specific research questions. The answer to these queries may either help
solidify our pre-established assumptions about the research, or identify and revise any potential
misconceptions. The subsequent analysis will be founded on answering 3 main research
questions using a variety of quantitative methods:
• RQ1: How well does the literature fare regarding SoS and IoT in recency?
• RQ2: What are the most resourceful entities known to partake extensively in the enrich-
ment of the IoT-based SoS state-of-the-art?
• RQ3: What fields of research are dominantly associated with IoT/SoS in recent years?
• RQ4: What other fields is the research on IoT shifting towards?
2.2. Data Sources
In regards to the topic of this study, IEEE is considered a credible and reputable publisher. IEEE
journals are top ranked in the Thomson Reuters Journal Citation Reports® (JCR). Based on
Scopus data, as of June 23rd, 2023, IEEE Systems Journal upholds an impact score of 5.6 and a
h-index of 98 [24]. it has thus been elected as our main source of documents for bibliometric
data. Its advanced command search for scientific documents is very selective and results are
largely relevant to our topic, as IEEE mainly publishes Computer and Electronics-related articles.
Additionally, it accurately sorts and filters results by simply ticking the corresponding boxes on
the IEEExplore webpage [25].
2.3. Bibliometric Indicators
The scope of this study, as the title suggests, extends to a timeframe that is deemed relatively
novel for computer science research and especially relevant to IoT platforms, SoS and SoS
Engineering. Therefore, it is characterized by recent and well received advancements within the
domain. We intend to achieve that by ensuring that both ‘activeness’ and ‘impact’ are equally
presented. Hence, we have selected pertinent bibliometric measurements including: 1. citation
counts, 2. contribution frequency and 3. co-occurrence relationships of several parameters,
depending on the available data. With the finalization of the research design, our next logical
step would be the execution of the meticulously designed plan in this next section labeled
‘methodology’.
3. Methodology
This section serves as a strict elaboration on the ‘realization’ phase previously mentioned in the
schema (figure 1). We outlined the steps undertaken to collect, process and organize the data,
utilizing Python as our primary tool throughout the procedure.
3.1. Data Collection
To initiate the data collection process as a first step, we have collected recent publications
related to IoT, CPS, SoS, and SoS Engineering. Our search began in July 10th, 2023 on the
“IEEExplore” indexing database. The chosen terms are deemed relevant and align with the
objectives of our research. The following command was executed via the advanced command tab:
(system*of?systems OR SoS OR system*of?systems?engineering OR SoSE)
AND (internet?of?things OR IoT OR cyber?physical?systems OR CPS)
In addition to this, only IEEE Journal publications from 2018 until 2023 were considered.
These filters were applied for novelty and pertinence in accordance to the scope of this paper.
The search was done at exactly: 10:53 am. Due to the large volume of results (7032 paper),
they were sorted by ‘Most Cited by Papers’ and exported to a limit of 2000, and were then
retained in a .csv file. Another search with identical parameters was initiated at 11:05 am.
However, this time, the resulting papers were exported based on “Relevance”. Likewise, the
first 2000 documents were retained in .csv format. The documents’ metadata include: Title,
Authors, Author Affiliations, Publication Title (Journal), Publication Year, DOI, PDF Link, Article
Citation Count, Reference Count, Author Keywords, IEEE Terms, INSPEC Controlled Terms
and Non-Controlled Terms. Two queries were performed for the purpose of optimizing data
quality, as only papers that are extensively cited by other papers and also highly relevant to the
specified search query are included in the results.
3.2. Data Processing
The two sets of data were downloaded, then merged and saved in a separate file, subsequently
undergoing a check for duplicates. The processing was done through a python script. Duplicates
were retained on the basis of “Document Title” and “DOI”; resulting in the removal of 2598
distinct entry, meaning that duplicate entries sum up to 1402. By retaining one record of each
duplicate, the processed dataset then contained information on a maximum of 701 papers.
Figure 2 summarizes the data collection process and the preprocessing done to the dataset
for the upcoming analysis:
Figure 2: Summuary of Data Collection and Preprocessing of IEEExplore results
3.3. Data Analysis
Python was instrumental in conducting the analysis by taking collected data as input and
employing various libraries and functions to derive meaningful statistics and extract relevant
bibliometric indicators. Furthermore, its capabilities to output graphical data allowing us to
visualize the results and facilitating its comprehension and interpretation. We employed a
number of analytics regarding authorship, geographic distribution, institutional contributions,
keyword occurrence and co-mapping. Of which we cite the following:
• Publications: the initial step was to present superficial data on the search. 2 key analyses
were done; (i.) publication frequency over time, and (ii.) impactful papers marked by a
relatively high number of citations.
• Authors: regarding this analysis, we have employed a variety of algorithms for retriev-
ing significant bibliometric data. First off, we were able to sort the authors with most
contributions overall, as well as most contributions as a first author. We have also mapped
them to the number of citations of their respective contributions. For visual comparison,
we drew a bar chart for each author. The bars correspond respectively to the number
of citations as a first author and the overall number of contributions. Additionnally, we
have drawn a series of co-authorship graphs for most influencial authors only.
• Keywords: As for keyword occurrence. A number of measures were taken. Firstly, the
initial dataset provides 4 types of key terms:
1. Author-defined keywords
2. IEEE assigned terms
3. INSPEC Non-Controlled terms
4. INSPEC Controlled terms
"Author keywords" were discarded for highly common cases of redundancy. There
is no clear standard as to how some keywords are spelled and therefore we were left
with different names for identical concepts (e.g. internetofthings, internet-of-things, IoT. . . )
"IEEE terms" however, are distinct, and that helped us identify them individually for
further analysis. For this instance, we applied a similar process to draw a multitude
of co-occurrence graphs. The graphs only contains keywords cited in great numbers.
"INSPEC Controlled" and "Non-Controlled Terms" were also discarded as, in
programming terms, additional processing was required in order to identify the topic of
given records. Besides, their IEEE assigned counterparts were already available.
• Countries: this subsection sets the spotlight on the contributability and influence of each
country mentioned in the results of the search by countries on the basis of (i.) number of
papers published and (ii.) number of citations per published article. Both of which were
projected on a pie chart.
• Institutes: in a much similar manner, this analysis simply highlights the most active
institutes and the most cited ones throughout the search.
4. Findings
4.1. General analysis
Activity in Recent Years: the graph in figure 3 presents the number of articles published for
each year in the span of these past 5 years (including ‘2023’, the year in which the search was
done) concerning the 701 publications in the dataset.
Figure 3: Number of Articles Over the Years
As shown in the graph (figure 3), it turns out the number of publications regarding our topic
is in decline as compared reflexively. This might be due to the nature of the search which
selectively excluded all papers that are yet to be published (early-access), and as all these papers
are journal articles, the publishing process can take a considerable amount of time.
Influential Publications: among 701 records in the dataset, Table 1 showcases the first 10
most cited articles that are relevant to the topic mentioned in the command query:
Table 1
Citations per Publication
Rank Reference Citations
1 [26] 654
2 [27] 651
3 [28] 631
4 [29] 630
5 [30] 467
6 [31] 448
7 [32] 433
8 [33] 419
9 [34] 373
10 [35] 371
4.2. Authorship analysis
Most Contributing Authors: Table 2 highlights the most active authors in terms of contribu-
tion count. The middle-right column indicates the number of times the author is mentioned in
published articles regardless of how they contributed in it. On the right-most column, it is shown
how many of those mentions correspond to a main authorship. 2396 individual researcher who
have contributed in at least one article is present in this dataset. We considered ‘most first
authorships’ as a criterion for sorting these results:
Table 2
Contributions per Author
Rank Author’s name Contributions as First
1 Y. Zhang 18 5
2 G. Fortino 6 4
3 M. Abdel-Basset 5 4
4 Y. Li 12 3
- X. Liu 12 3
6 J. Wang 11 3
7 G. Yang 8 3
- J. Li 8 3
9 Y. Xu 7 3
- X. Li 7 3
Most Cited Authors: Moreover, after having mapped all authors to their contributions, we
extracted the number of citations for each one in correspondence to the number of citations of
their respective papers. Here are the 10 most cited scholars:
Table 3
Contributions per Author
Rank Author’s name Citations Contributions
1 Y. Zhang 867 18
2 K. K. R. Choo 818 14
3 T. Taleb 753 3
4 N. Moustafa 728 7
5 B. Sikdar 689 2
6 V. Chamola 676 2
7 R. H. Glitho 665 2
- C. Mouradian 665 2
9 R. Bera 654 1
- L. Chettri 654 1
To give these numbers more significance, we employed the ‘matplotlib’ library in python
to generate a bar chart containing 2 super-imposed bars for each author; the blue one repre-
senting their overall number of contributions, and the green one representing only the ones as
a first author. This helps illustrate a clear comparative view of the most influential authors and
the number of contributions as first author or otherwise. The results were sorted by number of
citations and the first 50 were selected for display in figure 4:
Figure 4: Overall and First Author Contributions of Top 20 Most Cited Authors
Co-authorship: The ’networkx’ python library allowed for the systematic representation
of authors as nodes and co-author relationships as edges in the graph. By leveraging the
library’s functions, 3 networks were meticulously crafted to reflect the connections between the
most influencial authors and their co-authors; end results displayed in Figure 5. The separation
of each author apart was done so to enhance to the clarity of the analysis. By isolating each
significant author, the resulting networks provide a focused and manageable view of their
co-author reltationships, thereby facilitating its interpretation:
Figure 5: Co-authorship network for authors ’Y. Zhang’, ’K. K. R. Choo’ and ’T. Taleb’ respectively
4.3. Keyword Analysis
Most Mentioned Keywords: We have identified 854 distinct keywords regarding the column
‘IEEE Terms’ in the source dataset. As it was mentioned before, “Author Keywords” are
not subject to this analysis, for reasons such as inconsistency and redundancy.
Table 4
Occurrences per Keyword
Rank Keyword Occurrences
1 Internet of Things 393
2 Cloud Computing 127
3 Security 114
4 Sensors 90
5 Wireless Sensor Networks 76
6 Monitoring 75
7 Protocols 72
- Real-Time systems 72
9 Servers 70
10 Task analysis 68
Keywords per Year: In this analysis, we were aiming to grasp insight on the shift of interest
in topics related to IoT through categorizing keyword frequency by year. It turns out, logically,
‘Internet of Things’ is the most recurring keyword. It was omitted from this analysis
because we are trying to highlight what other keywords are associated with it. The years ‘2022’
and ‘2023’ were not considered as there were too few samples to build a meaningful viewpoint
(38 and 5 publications, respectively). Table 5 shows the obtained results:
Table 5
Keyword Occurrence per Year
Keyword 2018 2019 2020 2021
Cloud Computing 25 29 36 30
Security 9 27 31 36
Sensors 25 20 19 27
Wireless Sensor Networks 20 21 15 19
Wireless communication 16 15 24 12
Edge computing 17 12 24 13
Monitoring 15 25 19 17
Protocols 15 21 21 15
Real-Time systems 19 19 14 17
Servers 7 14 23 22
Task analysis 8 9 23 26
Co-occurrence: through a process analog to the one employed for establishing co-authorship
networks, we were able to plot the 3 most mentioned keywords on a network graph each. To
avoid complexity and intra-convolutedness, only keywords having occurred for more than 5
times together were assigned a link in-between. and a node was placed for each keyword, as in
Figure 6:
Figure 6: Co-authorship network for keywords ’Internet of Things’, ’Cloud computing’ and ’Security’
respectively
4.4. Affiliation Analysis
Most Contributions by Country: After analyzing the number of contributions by each
country, 90 countries were initially deduced from the results. However, upon manual inspection,
there seemed to have been some confusion in a few entries, for example, some articles were
listed as ‘United Arab Emirates’ and others under the abbreviated label ‘UAE’, other examples
include ‘USA’ and ‘US’, ‘China’ and ‘R.O.C’, ‘Republic of Korea’ and ‘South Korea’. Some other
papers were published under their 2-letter country code (i.e., DK and Denmark, CA and Canada).
These entries were merged and then ranked as in Table 6:
Table 6
Contributions per Country
Rank Country Contributions Citations
1 China 804 38883
2 USA 407 22465
3 Australia 226 12191
7 U.K. 153 9799
4 India 140 10573
5 Canada 134 9224
6 South Korea 125 9116
9 Italy 117 5910
9 Spain 95 4335
10 Pakistan 87 5578
Using the python module ‘matplotlib’, we were able to draw the following pie chart
(Figure 7) to put into perspective a wider range of the results, displaying the top 20 countries by
percentage of contributions:
Figure 7: Distribution of Contributions by Country
Most Citations by Country: By sorting the previous results in terms of overall citation
count, the ranking of the top countries somewhat changed, as the tabular data (Table 7) suggests:
Table 7
Citations per Country
Rank Country Contributions Citations
1 China 804 38883
2 USA 407 22465
3 Australia 226 12191
4 India 140 10573
5 U.K. 153 9799
6 Canada 134 9224
7 South Korea 125 9116
8 Malaysia 67 5980
9 Italy 117 5910
10 Pakistan 87 5578
Figure 8 shows 20 entries instead of just 10, illustrating an expanded portion of the data:
Most Mentioned Institutes: the organizations to which authors are affiliated were extracted
and ordered by number of contributions. A maximum of 1647 different institute (laboratories,
universities, research centers) was obtained and is shown in (Table 8) along with the number of
times assuming position of main (first) contributor:
Most Cited Institutes: similarly, the output holds information on 1647 records, however,
their ranking differs as some institutes have made valuable contributions with less articles
published, and others have published more but without considerable impact. The analysis
in Table 9 is more significant in identifying the value of the contributions by said institutes,
contrary to previous results which simply reflect the activity of institutes in recent years.
Figure 8: Percentage of Citation Count Shared by Each Country
Table 8
Contributions per Institute
Rank Institution’s name Contributions Main
Center for Innovative Integrated Electronic Systems,
1 26 0
Tohoku University, Sendai, Japan
Department of Information Systems and Technology,
2 23 21
Mid Sweden University, Sundsvall, Sweden
Department of Electronic Systems,
3 18 17
Aalborg University, Aalborg, Denmark
Institute for Communication Systems,
4 17 9
University of Surrey, Guildford, U.K.
Department of Computer and Information Sciences,
5 16 16
Towson University, Towson, MD, USA
School of Computer Science and Engineering,
6 15 10
Nanyang Technological University, Singapore
School of Electrical and Electronic Engineering,
7 14 14
Yonsei University, Seoul, South Korea
Department of Information Engineering,
8 14 10
University of Brescia, Brescia, Italy
Department of Electrical and Computer Systems Engineering,
9 13 13
Monash University, Melbourne, VIC, Australia
Department of Systems and Computer Engineering,
10 12 10
Carleton University, Ottawa, ON, Canada
Table 9
Citations per Institute
Rank Institution’s name Citations Contributions
Concordia Institute for Information Systems Engineering,
1 2802 12
Concordia University, Montreal, QC, Canada
Department of CSE and IT,
2 2604 4
Jaypee Institute of Information Technology, Noida, India
Department of Electronic Systems,
3 2189 18
Aalborg University, Aalborg, Denmark
Department of Information and Communication Engineering,
4 2008 8
Yeungnam University, Gyeongsan, South Korea
Wireless Communication Center,
5 1893 3
Universiti Teknologi Malaysia, Johor Bahru, Malaysia
Center for Wireless Communications,
6 1792 4
University of Oulu, Oulu, Finland
Real-Time Power and Intelligent Systems Laboratory,
7 1464 4
Clemson University, Clemson, SC, USA
Department of Information Systems and Technology,
8 1328 23
Mid Sweden University, Sundsvall, Sweden
5G IoT Lab, Sikkim Manipal Institute of Technology,
9 1308 2
Sikkim Manipal University, Majitar, India
Office of the CTO, CISCO Systems, Inc.,
10 1260 2
San Jose, CA, USA
5. Discussion
5.1. Results
The research questions raised in section 2 have found their answers through the findings we’ve
gathered. In this section we’ve mapped each question to its answer:
RQ1: “How well does the literature fare regarding SoS and IoT in recency?”
A1: we can fairly say that there is ample interest if we consider publication frequency for a
metric, the preliminary results show 7032 publication in total before applying any filters and
data preprocessing. The number of citations per article shown in Table 1 is admirable for such
recently published articles.
RQ2: “What are the most resourceful entities known to partake extensively in the enrichment of
the IoT-based SoS state-of-the-art?”
A2: The findings of this study shed light on key players shaping the state of the literature
concerning Iot and SoS. By observing the data in sections 4.2. and 4.4., we can deduce a number
of things. Namely, influential authors and institutions are identified in tables 2, 3, 8, and 9, as
well as distribution by countries, shown in tables 5, 6 and figures 7, 8. Collaboration networks
(Figure 5) also help significantly in providing valuable insight on the leading actors of this
interdisciplinary field.
RQ3: “What fields of research are dominantly associated with IoT/SoS in recent years?”
A3: the study unveils the intersection of IoT and SoS with various other fields. Tables 4 and 5
present the most common keywords associated with IoT and their popularity through the years.
The networks in Figure 6 offer a clear perspective on the confluence taking place between these
research domains.
RQ4: "What other fields is the research on IoT shifting towards?"
A4: There are entries in Table 5 that exhibit a gradual rise in interest across the years. The most
prominent fluctuations can be observed in the following keywords; "Security", "Servers"
and "Task analysis". It is, cordially, assumed that IoT research is progressively leaning in
the direction of said topics.
5.2. Limitations
Throughout the process, several obstacles have effectively hindered the quality of the final
results. We recognize a series of limitations inherent in the chosen scope and used tools. First,
we were constrained by the quantity of preliminary data to work on. 701 sample is not a terrible
number by any means; however, it is not great in any sense either.
Secondly, the choice of data sources was limited on a singular database (ieeexplore). That
posed a decrease in the quality of the data. For example, the data downloaded from ieeexplore
comprises of exactly 1 mention of ‘systems of systems’ in the ‘IEEE Terms’ column despite the
command query clearly stating it with the corresponding wildcards to avoid syntactic confusion.
That could be due to the SoS-related papers either being deemed irrelevant by the sorting
algorithm on the website, or for having very little popularity and therefore not making it in the
top 2000.
Lastly, the plotted graphs, especially the networks, greatly undermine the analysis. They only
show a fraction of the data and in a very simplified manner so as to match human readability
standards.
6. Conclusion
SoS and IoT have been the of subject to ever-expanding surge of interest in the last decade. The
sheer diversity of real-world applications stemming from them are so diverse that a conventional
narrative literature review could hardly encompass all the concepts anchored to the overarching
themes of SoS or IoT.
In recognition of this, our bibliometric analysis is distinctly focused on the recent history of
IoT-related publications. Leveraging the capabilities of the Python programming language and its
accompanying libraries proved handy in extracting, processing, analyzing and plotting raw data
in a multifaceted manner. We were able to transform the retrieved data, along structured tables
and labeled graphs, into a visual narrative that encompasses pie charts, bar charts and network
graphs that unveil meaningful patterns and relationships of co-occurrence and collaborations
between authors.
Given the promising relevance of SoS and IoT, we encourage authors to persist in their
exploration of this field. Moreover, we incite on the combination of adjacent and complementary
modeling approaches such as Digital Twin technology, Deep Learning, and the emerging
landscape of Quantum Computing. The intertwining of these domains could potentially unravel
great things in the future.
Acknowledgment
This work was partially supported by the LABEX-TA project MeFoGL: "Méthode Formelles
pour le Génie Logiciel"
References
[1] F. Alkhabbas, R. Spalazzese, P. Davidsson, Iot-based systems of systems, Proceedings of the
2nd edition of Swedish Workshop on the Engineering of Systems of Systems (SWESoS2016)
(2016) 34–37.
[2] G. Fortino, C. Savaglio, G. Spezzano, M. Zhou, Internet of things as system of systems: A
review of methodologies, frameworks, platforms, and tools, IEEE Transactions on Systems,
Man, and Cybernetics: Systems 51 (2021) 223–236. doi:10.1109/tsmc.2020.3042898.
[3] B. Lebtag, P. Teixeira, V. G. Neto, A systematic mapping on executable models for the
architectural design of systems-of-systems, Anais do IV Workshop em Modelagem e
Simulação de Sistemas Intensivos em Software (2022) 11–20. doi:10.5753/mssis.2022.
225662.
[4] A. Seghiri, F. Belala, N. Hameurlain, A formal language for modelling and verifying
systems-of-systems software architectures, International Journal of Systems and Service-
Oriented Engineering 12 (2022) 1–17. doi:10.4018/IJSSOE.297137.
[5] J. Delsing, G. Kulcsár, O. Haugen, Sysml modeling of service-oriented system-of-systems,
Innovations Syst Softw Eng (2022). doi:10.1007/s11334-022-00455-5.
[6] B. Karaduman, M. Challenger, Smart cyber-physical system-of-systems using intelligent
agents and mas, in: Engineering Multi-Agent Systems, volume 13190, Springer, Cham,
2021. doi:10.1007/978-3-030-97457-2_11.
[7] Y. Nomaguchi, K. Fujita, A framework of system of systems design with scenario, multi-
agent simulation and robustness, Procedia CIRP 60 (2017) 133–138. doi:10.1016/j.
procir.2017.01.008.
[8] S. K. R. Harbo, E. P. Voldby, J. Madsen, M. Albano, A diagram-centric modeling tool for
systems of systems, Proceedings of the 25th International Conference on Model Driven
Engineering Languages and Systems: Companion Proceedings (MODELS ’22) (2022) 51–55.
doi:10.1145/3550356.3559093.
[9] A. Gassara, I. B. Rodriguez, M. Jmaiel, A tool for modeling sos architectures using bigraphs,
Proceedings of the Symposium on Applied Computing (SAC ’17) (2017) 1787–1792. doi:10.
1145/3019612.3019802.
[10] I. G. Vargas, T. Gottardi, R. T. V. Braga, An approach to integrate systems towards a
directed system-of-systems, Proceedings of the 12th European Conference on Software
Architecture: Companion Proceedings (ECSA ’18) (2018) 1–7. doi:10.1145/3241403.
3241431.
[11] M. Chaabane, I. B. Rodriguez, R. Colomo-Palacios, W. Gaaloul, M. Jmaiel, A modeling
approach for systems-of-systems by adapting iso/iec/ieee 42010 standard evaluated by
goal-question-metric, Sci. Comput. Program. 184 (2019). doi:10.1016/j.scico.2019.
102305.
[12] N. Akhtar, S. Hussain, Formal modelling of a mail transport system based on multi-agent
system of systems, Journal of Information Communication Technologies and Robotic
Applications 10 (2019) 68–79.
[13] A. Gassara, I. B. Rodriguez, M. Jmaiel, K. Drira, Executing bigraphical reactive systems,
Discrete Applied Mathematics 253 (2019) 73–92. doi:10.1016/j.dam.2018.07.006.
[14] W. Manzano, V. V. G. Neto, E. Y. Nakagawa, Dynamic-sos: An approach for the simulation
of systems-of-systems dynamic architectures, The Computer Journal 63 (2020) 709–731.
doi:10.1093/comjnl/bxz028.
[15] A. Gassara, I. B. Rodriguez, M. Jmaiel, K. Drira, A bigraphical multi-scale modeling
methodology for system of systems, Computers & Electrical Engineering 58 (2017) 113–
125. doi:10.1016/j.compeleceng.2017.01.016.
[16] F. Lonetti, V. O. Neves, A. Bertolino, Designing and testing systems of systems: From
variability models to test cases passing through desirability assessment, Journal of Software:
Evolution and Process 34 (2022). doi:10.1002/smr.2427.
[17] K. A. Raz, P. C. Wood, L. Mockus, D. A. Laurentis, System of systems uncertainty quantifica-
tion using machine learning techniques with smart grid application, Systems Engineering
23 (2020) 770–782. doi:10.1002/sys.21561.
[18] G. Ahamer, K. A. Kumpfmüller, Education and literature for development in responsibility:
Partnership hedges globalization, in: Handbook of Research on Transnational Higher
Education, IGI Global, 2014, pp. 526–584. doi:10.4018/978-1-4666-4458-8.ch027.
[19] E. Ahmad, Model-based system engineering of the internet of things: A bibliometric
literature analysis, IEEE Access 11 (2023) 50642–50658. doi:10.1109/ACCESS.2023.
3277429.
[20] O. Ellegaard, J. A. Wallin, The bibliometric analysis of scholarly production: How great is
the impact?, Scientometrics 105 (2015) 1809–1831. doi:10.1007/s11192-015-1645-z.
[21] N. Donthu, S. Kumar, D. Mukherjee, N. Pandey, W. M. Lim, How to conduct a bibliometric
analysis: An overview and guidelines, Journal of Business Research 133 (2021) 285–296.
doi:10.1016/j.jbusres.2021.04.070.
[22] K. Mattick, J. Johnston, A. de la Croix, How to. . . write a good research question, Clin
Teach 15 (2018) 104–108. doi:10.1111/tct.12776.
[23] I. Zupic, T. Čater, Bibliometric methods in management and organization, Organizational
Research Methods 18 (2015) 429–472. doi:10.1177/1094428114562629.
[24] Resurchify.com, Ieee access, 2023. URL: https://www.resurchify.com/impact/details/
11300153734#:~:text=The%20overall%20rank%20of%20IEEE,this%20journal%20is%
20ranked%201.448.
[25] Ieee xplore, 2023. URL: https://ieeexplore.ieee.org/Xplore/home.jsp.
[26] L. Chettri, R. Bera, A comprehensive survey on internet of things (iot) toward 5g wireless
systems, IEEE Internet of Things Journal 7 (2020) 16–32. doi:10.1109/JIOT.2019.
2948888.
[27] V. Hassija, V. Chamola, V. Saxena, D. Jain, P. Goyal, B. Sikdar, A survey on iot security:
Application areas, security threats, and solution architectures, IEEE Access 7 (2019)
82721–82743. doi:10.1109/ACCESS.2019.2924045.
[28] O. Elijah, T. A. Rahman, I. Orikumhi, C. Y. Leow, M. N. Hindia, An overview of internet of
things (iot) and data analytics in agriculture: Benefits and challenges, IEEE Internet of
Things Journal 5 (2018) 3758–3773. doi:10.1109/JIOT.2018.2844296.
[29] C. Mouradian, D. Naboulsi, S. Yangui, R. H. Glitho, M. J. Morrow, P. A. Polakos, A
comprehensive survey on fog computing: State-of-the-art and research challenges, IEEE
Communications Surveys & Tutorials 20 (2018) 416–464. doi:10.1109/COMST.2017.
2771153.
[30] K. Shafique, B. A. Khawaja, F. Sabir, S. Qazi, M. Mustaqim, Internet of things (iot) for
next-generation smart systems: A review of current challenges, future trends and prospects
for emerging 5g-iot scenarios, IEEE Access 8 (2020) 23022–23040. doi:10.1109/ACCESS.
2020.2970118.
[31] P. Porambage, J. Okwuibe, M. Liyanage, M. Ylianttila, T. Taleb, Survey on multi-access edge
computing for internet of things realization, IEEE Communications Surveys & Tutorials
20 (2018) 2961–2991. doi:10.1109/COMST.2018.2849509.
[32] S. Kuutti, S. Fallah, K. Katsaros, M. Dianati, F. Mccullough, A. Mouzakitis, A survey of
the state-of-the-art localization techniques and their potentials for autonomous vehicle
applications, IEEE Internet of Things Journal 5 (2018) 829–846. doi:10.1109/JIOT.2018.
2812300.
[33] M. Ayaz, M. Ammad-Uddin, Z. Sharif, A. Mansour, E.-H. M. Aggoune, Internet-of-things
(iot)-based smart agriculture: Toward making the fields talk, IEEE Access 7 (2019) 129551–
129583. doi:10.1109/ACCESS.2019.2932609.
[34] M. Elhoseny, G. Ramírez-González, O. M. Abu-Elnasr, S. A. Shawkat, N. Arunkumar,
A. Farouk, Secure medical data transmission model for iot-based healthcare systems, IEEE
Access 6 (2018) 20596–20608. doi:10.1109/ACCESS.2018.2817615.
[35] Y. A. Qadri, A. Nauman, Y. B. Zikria, A. V. Vasilakos, S. W. Kim, The future of healthcare
internet of things: A survey of emerging technologies, IEEE Communications Surveys &
Tutorials 22 (2020) 1121–1167. doi:10.1109/COMST.2020.2973314.