Towards Interoperability of Systems of Systems Using GraphQL
Eduardo Dantas Luna
1 a
, Vitor Pinheiro de Almeida
1 b
and Eduardo Thadeu Corseuil
2 c
1
Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro, Brazil
2
Department of Informatics, Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro, Brazil
Keywords:
Digital Twins, System of Systems, Knowledge Graph, API Management, GraphQL.
Abstract:
The growing interconnectedness of devices and systems presents a significant opportunity to develop solutions
that leverage data from diverse sources. However, integrating data from these heterogeneous systems, which
may use different protocols and paradigms, poses a considerable challenge. This paper proposes an innovative
solution to address this challenge by introducing an algorithm that generates a GraphQL API management
layer. This layer acts as a bridge between disparate systems, enabling seamless data integration and exchange.
By leveraging GraphQL’s efficient data retrieval capabilities and a knowledge graph to define relationships
between data elements, the algorithm automates the creation of a processing layer that simplifies the integration
process. The proposed solution offers a promising approach to overcome the complexities of data integration,
paving the way for more robust and adaptable data-driven applications.
1 INTRODUCTION
In recent years, the advent of the Internet of Things
(IoT) and Web technologies has led to an exponential
growth of devices connected via the Web. All these
devices create and expose vast amounts of data, which
has tremendous potential for developing solutions that
require data from distinct sources.
For example, by joining data from different trans-
port data services (such as taxi, bus, metro, boat, etc.),
a system could precisely calculate which transport or
combinations of transports to arrive at the desired des-
tination with minimal time and money cost. However,
connecting data from those sources can be a challeng-
ing task.
1.1 The Problems
One of the challenges is how to combine data from
multiple systems since systems can describe the same
data in different ways or even partially describe the
data. An example of such a situation would be two
companies that sell clothes, but in one, the money is
described in dollars and another in euros; without an
explicit description of which currency, it is impossible
a
https://orcid.org/0009-0002-2663-0732
b
https://orcid.org/0000-0002-6544-9541
c
https://orcid.org/0000-0002-7543-4140
to combine this information effectively.
Another concern is combining systems using dif-
ferent data transfer protocols and paradigms. There
are different use cases for each paradigm. Therefore,
it can be challenging to create a communication be-
tween such systems. An example of this involves two
paradigms: a movie streaming service and a movie
catalogue service. A streaming service needs to send
a continuous data flux. In contrast, a catalogue ser-
vice receives a request and sends a response. If poorly
implemented, the streaming service can create more
requests than the catalogue can handle.
Additionally to all these issues is the need to de-
velop those data exchange interfaces. To implement
the connections between systems, developers from
particular systems must learn to deal with different
programming languages and technologies to under-
stand how to connect systems. Additionally, if created
without standardization, different developers could
develop similar interfaces, creating unnecessary re-
work. Also, these interfaces may be composed of
more than just two systems, increasing the difficulty
and the time cost to develop those services.
1.2 Objective
This paper addresses the problem of combining data
from systems that may use different protocols and
paradigms. To achieve this, we will systematically
282
Luna, E., Pinheiro de Almeida, V. and Corseuil, E.
Towards Interoperability of Systems of Systems Using GraphQL.
DOI: 10.5220/0012997100003825
Paper published under CC license (CC BY-NC-ND 4.0)
In Proceedings of the 20th International Conference on Web Information Systems and Technologies (WEBIST 2024), pages 282-287
ISBN: 978-989-758-718-4; ISSN: 2184-3252
Proceedings Copyright © 2024 by SCITEPRESS – Science and Technology Publications, Lda.

search the most relevant academic papers and pro-
pose a solution based on the findings. The pro-
posed solution will focus on generating a communi-
cation layer that can effectively integrate data from
diverse sources, addressing the challenges of hetero-
geneity, protocol differences, and the need for effi-
cient data exchange. By leveraging existing research
and proposing a novel approach, this paper seeks to
contribute to the development of more robust and
adaptable data-driven applications.
1.3 Next Sections
The remainder of this paper is structured as follows:
Section 2 establishes a theoretical foundation, defin-
ing key concepts used throughout the paper. Section
3 presents a systematic analysis of related academic
work, providing insights into existing approaches and
their limitations. Section 4 introduces a practical use
case from the oil and gas industry, highlighting the
challenges of data integration in real-world scenarios.
Section 5 details the proposed solution, including the
conceptual model and the algorithm for generating
the API management layer. Finally, Sections 6 and
7 discuss unresolved issues and potential future work
to address these challenges and enhance the proposed
solution further.
2 THEORETICAL FOUNDATION
An interface is indispensable for retrieving data from
a system. These interfaces are predominantly imple-
mented on the web as Web Applications Program-
ming Interfaces (APIs). An API consists of a collec-
tion of interfaces, often referred to as endpoints. To
manage these APIs, a system must control multiple
aspects of the APIs, such as Authorizations and Rate
Limiting; this control system is called API Manage-
ment. ((Bondel et al., 2021))
API Management has two main parts: the API
Gateway and the API Portal. The Gateway commu-
nicates services and the API, thereby decoupling the
client interface. It also intercepts all incoming re-
quests, routing them to the correct service. It includes
many features like caching, scaling, load balance, etc.
On the other hand, the API Portal serves as a frontend
for both API developers and consumer systems de-
velopers. It provides documentation and a guide plat-
form for developers. One of its core features is the
endpoint catalogue for each service provided through
the multiple APIs. ((Bondel et al., 2021))
API Management is often used in a system archi-
tecture called microservices. A microservice archi-
tecture uses multiple decoupled systems instead of
a monolith system. By using this type of architec-
ture API Management acts as a system that orches-
trates these various systems. Since API Management
is used to connect multiple systems, it can be used
to build systems composed of interaction with other
systems. These systems are referred to as Systems
of systems (SoS) and are often employed in various
scenarios, including Digital Twins. A Digital Twin is
a digital representation of a physical system that mir-
rors real-world behaviours by integrating and present-
ing updated information across the various technolo-
gies that compose the Digital Twin. ((Shi et al., 2016;
Anacker et al., 2022; Olsson and Axelsson, 2023))
An API can use several paradigms, including
REST, SOAP and gRPC. However, this thesis primar-
ily explores the GraphQL paradigm. GraphQL, which
stands for Graph Query Language, is a standard that
defines ways to query and handle data. Advantages of
GraphQL include: efficient data retrieval since it al-
lows for precise data requests, avoiding over-fetching
and under-fetching; It has a single endpoint in con-
trast to other paradigms; Strongly typed data; It can
aggregate data from multiple data sources. (GraphQL
Foundation, 2024)
The architecture of GraphQL is built around two
core components: the GraphQL Server and the query
language. The server is responsible for processing
and executing queries using three main elements:
types, fields and resolver functions. The types and
fields are used to define how the data is structured in
the API. The types and fields also defines all possi-
ble queries a client can make. Resolver functions are
blocks of code responsible for executing the query,
each query defined in the schema corresponds to a
resolver function. The query language closely mir-
rors the type definition language, utilizing defined
queries combined with several operands to fetch data
described by the query. (GraphQL Foundation, 2024)
Similar to how GraphQL is a standard, there ex-
ists a standard for REST APIs known as OpenAPI.
OpenAPI files serve as descriptor files that REST API
libraries extensively utilize to provide detailed infor-
mation about the API. These files encompass data
such as the url of all endpoints, the response codes
implemented by each endpoint, the input and out-
put schemas, and whether specific parameters are re-
quired, among other details. (Initiative, 2024)
A knowledge graph is a sophisticated data struc-
ture frequently employed to model, organize, man-
age, and analyze heterogeneous and intricate datasets.
Owing to its graph-based architecture, it encapsulates
a complex abstraction of knowledge on a particular
domain and delineates the interrelationships among
Towards Interoperability of Systems of Systems Using GraphQL
283

various data entities. (Ramonell et al., 2023)
3 RELATED WORK
This paper undertakes a systematic search for the
most relevant academic papers to be reviewed. This
systematic search is illustrated by Figure 1. The
search process was conducted using the Findpapers
library (Grosman, 2024), which allows users to create
queries based on specific keywords to retrieve aca-
demic papers. The data sources utilized by this li-
brary include ACM, arXiv, bioRxiv, IEEE, medRxiv,
PubMed, and Scopus.
3.1 Metodology
Given the focus of this thesis on Systems of Sys-
tems (SoS) and related aspects of Interoperability, a
query was formulated using two groups of keywords.
The first group included: Integration, Interoperability,
Digital Twins, API Management, Representational
State Transfer (REST), GraphQL and Federated Sys-
tems. To specifically address the SoS focus, a sec-
ond group of keywords related to System of Systems
was created. A conditional AND was applied between
these two groups and a time constraint was set to in-
clude only papers published after 2019.
The initial query returned a substantial number of
papers. To refine the results, an additional filter was
applied using a third group of keywords with AND
NOT conditions. This third group is comprised of
authentication, security, authorization, and cyberse-
curity. This filtering process reduced the number of
papers to 111.
Further refinement was necessary due to the pres-
ence of papers without DOIs and those behind pay-
walls, which were inaccessible. Consequently, the
number of available papers was reduced to 48. A
word count analysis was then performed on the re-
maining papers to verify that their primary focus was
indeed on Systems of Systems and Interoperability.
The article was selected if it had more than 5 occur-
rences of ’Interoperability’ and more than 6 occur-
rences of ’System of Systems’, this two conditions
narrowed down the number of papers to 7. These are
the 7 papers: (Pickering et al., 2020; Mittal et al.,
2020; Mohsin and Janjua, 2018; Weinert and Uslar,
2020; Neureiter et al., 2020; C
ˆ
andea et al., 2023;
Anacker et al., 2022)
Since this systematic selection process was exe-
cuted to find papers focusing on System of Systems
and Interoperability, it was also added two papers re-
lated to the System of Systems, Interoperability and
GraphQL (De F. Borges et al., 2022; Li et al., 2024),
a Survey about Digital Twins and System of Sys-
tems (Olsson and Axelsson, 2023) and a survey about
GraphQL (na Mera et al., 2023).
Figure 1: Systematic Research Diagram.
3.2 Preliminary Analysis
(C
ˆ
andea et al., 2023) focus on implementing Internet
of Things (IoT) and System of Systems (SoS) tech-
nologies in various smart applications, such as Indus-
try 4.0, smart cities, and healthcare. They emphasize
the importance of interconnectivity and interoperabil-
ity for seamless communication and data exchange
between devices. The authors also highlight the need
for reliable quality of service (QoS) to ensure the
performance and effectiveness of these applications.
(Mittal et al., 2020) present the Simulation, Experi-
mentation, Analytics, and Testing (SEAT) framework,
designed to facilitate the development and testing of
autonomous systems within a multi-domain environ-
ment. The framework emphasizes the composability
and interoperability of simulation and analytical ap-
plications, enabling the integration of diverse tools
and capabilities. (Li et al., 2024; De F. Borges et al.,
2022) introduce a framework for using GraphQL.
However, one of them uses a framework that lever-
ages ontologies to generate a GraphQL server, and
the other uses syntactical analysis. Both of them can
query heterogeneous data sources, providing an inte-
grated view of the data.
(na Mera et al., 2023) and (Mohsin and Janjua,
2018) provide comprehensive reviews of GraphQL
WEBIST 2024 - 20th International Conference on Web Information Systems and Technologies
284

and SOA-based software architecture modeling ap-
proaches for SoS, respectively. (na Mera et al., 2023)
highlight GraphQL’s growing adoption and potential
research areas, while (Mohsin and Janjua, 2018) delve
into the limitations of existing SOA-based modelling
techniques for SoS, emphasizing the need for dy-
namic service identification, composition, and provi-
sioning at runtime. (Olsson and Axelsson, 2023) sur-
vey the current knowledge on digital twins in the con-
text of SoS, emphasizing the need for further research
to address challenges such as data sharing and inte-
gration. (Anacker et al., 2022) review the literature
on SoS and patterns, aiming to understand their def-
initions, classifications, and applications in systems
engineering.
(Pickering et al., 2020) propose a time-constrained
SoS discovery process and canvas, demonstrating its
application in an agricultural case study involving an
automated asparagus harvester. The authors empha-
size the importance of understanding the relationships
between constituent systems and their properties to
manage emergent properties effectively. (Weinert
and Uslar, 2020) outline the demand and challenges
for a structured SoS approach in the agriculture do-
main, proposing a reference designation-based sys-
tem architecture documentation approach to address
the heterogeneous infrastructure and lack of interop-
erability. (Neureiter et al., 2020) discuss extending
the concept of Domain-Specific Systems Engineering
(DSSE) to SoS, highlighting the need for interoper-
ability and compatibility between models from differ-
ent domains. They present a case study on integrating
electric vehicle models with a Smart Grid model to
analyze emergent behaviour caused by simultaneous
charging.
3.3 Comparison
After analyzing the related work, a methodical com-
parison was devised. Since this study focuses on the
implementation aspect of Systems of Systems (SoS)
and Interoperability, six questions were formulated
to compare the four implementation-focused papers
(C
ˆ
andea et al., 2023; Mittal et al., 2020; Li et al.,
2024; De F. Borges et al., 2022). The questions are
the following, and the answers to those questions are
on the Table 1.
1. Q1: Does it talk about joining data from multiple
sources?
2. Q2: Does it propose a methodology?
3. Q3: Is the paper related to GraphQL?
4. Q4: Does it have an API paradigm restriction?
5. Q5: Does it provide a unified vocabulary to inte-
grate APIs?
The comparison of implementation papers in 1 re-
veals a diverse landscape of approaches to address
the challenges of data integration in systems of sys-
tems. It is possible to observe that by the table, most
of the problems presented in the Subsection 1.1 are
tackled by (De F. Borges et al., 2022) and (Li et al.,
2024), both relating to GraphQL. This analysis high-
lights the need for further research and development
to create comprehensive solutions that address the di-
verse challenges of data integration in complex sys-
tems.
4 USE CASE
Our use case involves multiple systems within the oil
and gas industry, which could collectively be used
to create a digital twin of an industrial plant. The
key challenge lies in the heterogeneity and complex-
ity of these systems, making data integration a non-
trivial task. These systems are currently intercon-
nected through peer-to-peer connections, which are
increasing exponentially, leading to a complex and
potentially unsustainable network architecture. The
proposed solution aims to address this challenge by
providing a scalable and efficient way to integrate data
from these diverse systems, enabling the creation of a
comprehensive digital twin that can accurately repre-
sent the real-world behaviour of the industrial plant.
5 PROPOSED SOLUTION
After analyzing the work related to this paper, the so-
lution proposed by this paper is an algorithm that can
generate a processing layer between systems that as-
sists in joining data from multiple data sources. The
following subsections illustrate how.
5.1 Conceptual Model
The conceptual model is illustrated by Figure 2. It has
three main layers: the micro-services, the API Man-
agement and the outside layer. The Micro-services
layer encompasses all applications and services man-
aged by an organization. The API management layer
has the API Gateway and API Portal that expose all
endpoints and information of the services. And the
outside layer is applications or systems that consume
data from the organization’s service. This conceptual
model will be employed to explore the creation of in-
terfaces between services within a System of Systems
Towards Interoperability of Systems of Systems Using GraphQL
285

Table 1: Comparison of implementation papers.
(C
ˆ
andea et al., 2023) (Mittal et al., 2020) (Li et al., 2024) (De F. Borges et al., 2022)
Q1 No No Yes Yes
Q2 No Yes Yes Yes
Q3 No No Yes Yes
Q4 Yes No No No
Q5 No No Yes Yes
Figure 2: Conceptual Model.
(SoS) framework to develop a Digital Twin. The so-
lution of this paper is focused on generating the API
Management layer.
5.2 Server Generation Algorithm
The solution of this study will use GraphQL
API Management since other related works like
(De F. Borges et al., 2022; Li et al., 2024) have
shown results using it. The algorithm will gener-
ate a GraphQL API Management based on two main
information sources: A Knowledge Graph and the
OpenAPI files for each service. These information
sources will be employed to define how to merge the
data from multiple applications effectively.
The OpenAPI files will be used to extract informa-
tion regarding the URLs and the input and output for-
mat of each endpoint. On the other hand, the Knowl-
edge Graph will have definitions of how to combine
data from each endpoint by defining which pieces of
data are equivalent in each service.
The algorithm will take all the information
sources as input, and for each service in the Ope-
nAPIs, it will create GraphQL Types, Fields and Re-
solvers for the correspondent service. These resolvers
will be simple endpoint calls defined by the Ope-
nAPI file. By doing it this way, all services can be
queried through the GraphQL API Management and
also these services can be used afterwards to build
new services inside GraphQL.
After generating all the GraphQL Types, Fields
and Resolvers, the next step is to create new GraphQL
Types, Fields and Resolvers based on each equiv-
alence between services in the Knowledge Graph.
After all this process, the algorithm will output a
GraphQL Server code that can be instantiated.
6 DISCUSSION
The proposed solution offers a novel approach to ad-
dress the challenges of data integration in systems
of systems. By leveraging GraphQL and knowledge
graphs, the algorithm automates the generation of a
processing layer that seamlessly connects disparate
data sources. This automation not only reduces the
development time and effort required for building
such integrations but also enhances the flexibility and
adaptability of the system. The use of GraphQL as the
API management layer ensures efficient data retrieval
and a strongly typed schema, contributing to the over-
all robustness and maintainability of the solution.
However, there are potential limitations to con-
sider. The effectiveness of the algorithm relies on the
accuracy and completeness of the knowledge graph,
which defines the relationships between data elements
from different sources. Inaccurate or incomplete
knowledge graphs could lead to incorrect or incom-
plete data integration. Additionally, while GraphQL
offers numerous advantages, it may not be the opti-
mal choice for all use cases. For instance, in scenarios
where real-time data streaming is critical, other pro-
tocols like gRPC might be more suitable.
7 CONCLUSION
This paper proposes an innovative algorithm for gen-
erating a GraphQL API management layer that fa-
cilitates data integration in systems of systems. By
combining the strengths of GraphQL and knowledge
graphs, the algorithm automates the creation of a pro-
cessing layer that seamlessly connects disparate data
sources.
WEBIST 2024 - 20th International Conference on Web Information Systems and Technologies
286

Building upon works like (Li et al., 2024) and
(De F. Borges et al., 2022), which sought to address
the persistent technical challenges in the software en-
gineering process, our research contributes a novel
approach that resonates with organizations grappling
with similar issues. This is particularly salient given
the observations made in study (na Mera et al., 2023),
which underscores the need for innovative solutions,
such as code generative tools, to tackle complex ele-
ments like paging and nested queries. Our findings,
therefore, offer a potentially transformative pathway
for enhancing software engineering practices across a
multitude of domains.
While potential limitations exist, the proposed so-
lution offers a promising avenue for addressing data
integration challenges in complex systems, contribut-
ing to the development of more efficient, adaptable,
and robust data-driven applications. Future work
could implement the algorithm mentioned in this
paper. Another possibility is the creation of such
Knowledge Graphs through various methods such as
syntactical analyses, semantic analyses, or even using
an LLM to generate the knowledge graph.
REFERENCES
Anacker, H., G
¨
unther, M., Wyrwich, F., and Dumitrescu, R.
(2022). Pattern based engineering of system of sys-
tems - a systematic literature review. In 17th Annual
System of Systems Engineering Conference (SOSE),
page 178–183.
Bondel, G., Landgraf, A., and Matthes, F. (2021). Api man-
agement patterns for public, partner, and group web
api initiatives with a focus on collaboration. Pro-
ceedings of the ACM on Programming Languages,
5(OOPSLA):1–28.
C
ˆ
andea, C., C
ˆ
andea, G., and Staicu, M. (2023). Impact of
iot and sos in enabling smart applications: A study on
interconnectivity, interoperability and quality of ser-
vice. Procedia Computer Science, 221:1226–1234.
De F. Borges, M. V., Rocha, L. S., and Maia, P. H. M.
(2022). Micrographql: a unified communication ap-
proach for systems of systems using microservices
and graphql. In 2022 IEEE/ACM 10th International
Workshop on Software Engineering for Systems-of-
Systems and Software Ecosystems (SESoS), pages 33–
40.
GraphQL Foundation (2024). Introduction to graphql.
https://graphql.org/learn/. Accessed: 2024-05-22.
Grosman, J. (2024). Findpapers: A tool for helping re-
searchers who are looking for related works. https:
//github.com/jonatasgrosman/findpapers. Accessed:
2024-05-22.
Initiative, O. (2024). Openapi specification v3.1.0. https:
//spec.openapis.org/oas/latest.html. Accessed: 2024-
05-22.
Li, H., Hartig, O., Armiento, R., and Lambrix, P. (2024).
Ontology-based graphql server generation for data ac-
cess and data integration. Semantic Web.
Mittal, S., Kasdaglis, N., Harrell, L., Wittman, R. L., Gib-
son, J., and Rocca, D. (2020). Autonomous and com-
posable m&s system of systems with the simulation,
experimentation, analytics and testing (seat) frame-
work. In Proceedings of the 2020 Winter Simulation
Conference, pages 2305–2316. IEEE.
Mohsin, A. and Janjua, N. K. (2018). A review and future
directions of soa-based software architecture model-
ing approaches for system of systems. Service Ori-
ented Computing and Applications, 12(3):183–200.
na Mera, A. Q., Fernandez, P., Garc’ia, J. M., and Ruiz-
Cort’es, A. (2023). Graphql: A systematic mapping
study. ACM Comput. Surv., 55(10):202:1–202:35.
Neureiter, C., Binder, C., Brankovic, B., and Lastro, G.
(2020). Extending the concept of domain specific sys-
tems engineering to system-of-systems. In 2020 IEEE
15th International Conference of System of Systems
Engineering (SoSE), pages 391–396. IEEE.
Olsson, T. and Axelsson, J. (2023). Systems-of-systems
and digital twins: A survey and analysis of the current
knowledge. In 2023 18th Annual System of Systems
Engineering Conference (SoSE). IEEE.
Pickering, N., Duke, M., and Lim, S. H. (2020). A time con-
strained system of systems discovery process and can-
vas - a case study in agriculture technology focusing
on an automated asparagus harvester. In 2020 IEEE
15th International Conference of System of Systems
Engineering (SoSE), pages 67–74. IEEE.
Ramonell, C., Chac
´
on, R., and Posada, H. (2023). Knowl-
edge graph-based data integration system for digital
twins of built assets. Automation in Construction,
156:105109.
Shi, W., Cao, J., Zhang, Q., Li, Y., and Xu, L. (2016). A
survey on edge computing for the internet of things.
IEEE Internet of Things Journal, 3(5):637–646.
Weinert, B. and Uslar, M. (2020). Challenges for system
of systems in the agriculture application domain. In
2020 IEEE 15th International Conference of System of
Systems Engineering (SoSE), pages 355–360. IEEE.
Towards Interoperability of Systems of Systems Using GraphQL
287