=Paper=
{{Paper
|id=Vol-2641/paper_04
|storemode=property
|title=An Agricultural Data Platform iStar
Model
|pdfUrl=https://ceur-ws.org/Vol-2641/paper_04.pdf
|volume=Vol-2641
|authors=Stefan Braun,István Koren,Marc Van Dyck,Matthias Jarke
|dblpUrl=https://dblp.org/rec/conf/istar/0007KDJ20
}}
==An Agricultural Data Platform iStar
Model ==
https://ceur-ws.org/Vol-2641/paper_04.pdf
An Agricultural Data Platform
iStar Model
Stefan Braun1 , István Koren1 , Marc Van Dyck2 , and Matthias Jarke1,3
1
Information Systems & Databases, RWTH Aachen University, Aachen, Germany
{braun,koren,jarke}@dbis.rwth-aachen.de
2
Technology and Innovation Management, RWTH Aachen University, Aachen,
Germany
vandyck@time.rwth-aachen.de
3
Fraunhofer FIT, Birlinghoven Castle, Sankt Augustin, Germany
Abstract. Organizations increasingly rely on external data and ser-
vice exchange within business networks in order to fuel their analytics
and artificial intelligence needs. In Industry 4.0 practices, new ecosys-
tems have evolved, where data and service provisioning often happens
within dedicated platforms. Hereby, the challenge lies in ensuring the
data sovereignty of enterprises in terms of self-determination with regard
to the use of their data. While conceptual modeling of these platforms
inhabits a large number of opportunities, for instance, including auto-
mated generation of access policies, research in this area is scarce. To
this end, we propose a bottom-up approach using the iStar 2.0 modeling
language. In this paper, we first introduce a model describing the mar-
ket participants of a data and service exchange platform in the realm of
smart farming. We then generalize and provide a formalization of rele-
vant aspects in a broader context. The resulting models serve both as a
basis for discussion on the requirements analysis level and as fundamental
groundwork for further value generation in the area of data sovereignty
in complex cross-organizational settings.
Keywords: Conceptual Modeling · Data Sovereignty · Industry 4.0
1 Introduction
Connected sensors, assets, products, and actors in Industry 4.0 continuously gen-
erate an enormous amount of data. The availability, access, and usage of data
and applications by multiple parties enable an increase in productivity due to
faster and more practical insights. They also enable new value propositions, such
as predictive maintenance and dynamic pricing. These effects are reinforced by
economies of scale; in the vision of the Internet of Production [4] the vast inter-
connection even allows new business models. Policies and agreements between
stakeholders are required to allow for the regulated collaboration of different
parties and their subsequent data sharing. An important aspect is thus the ques-
tion of fair value distribution, i.e., a balance of value creation and value capture
Copyright © 2020 for this paper by its authors. Use permitted under 19
Creative Commons License Attribution 4.0 International (CC BY 4.0).
between the stakeholders, and a more precise definition of the benefits for data
sharing [1]. The first step for the identification of the former—also, but not only,
in terms of data use—is to formulate goals, relationships, and interdependencies
of the several parties. One exemplary formalized modeling of a data ecosystem
platform was conducted by Chakrabarti et al. [2]. The authors describe the data-
related interdependencies of users of the alliance-like International Data Spaces
platform [8]. From a business management point of view, however, the steps
before the established data collaboration are also of great importance. To this
end, platform ecosystems evolved that connect various stakeholders from estab-
lished business partners to emerging market entrants like complementors [12],
i.e. businesses that complement the product or service of another company. In
this paper, we model interdependencies of platform stakeholders. First of all, this
view gives stakeholders an insight on collaboration opportunities. As a particular
example, a use case from agriculture is presented. It is based on an on-going large
case study of an existing and evolving smart farming ecosystem [13]. Together
with domain experts who performed the original study, we identified the busi-
ness models of manufacturers, dealers, contractors, and farmers, as well as farm
management platforms and complementors, and formalized it as an iStar 2.0
model. We then generalize and provide relevant aspects in a broader context.
While the model currently serves to derive business model requirements, we see
a multitude of possible further use cases, e.g., in the automated generation of
data usage policies. The remainder of this paper is structured as follows. First,
we introduce a model from the agriculture domain in Section 2 and have a look
at a Strategic Dependency (SD) model. In Section 3, we set our focus on the re-
lationship of the data platform with the previously existing ecosystem. Section 4
concludes the paper, discusses the use case of automated policy generation, and
gives an outlook on future work.
2 Agriculture Scenario
Since the agricultural sector is one of the driving forces behind digitalization [6,11],
it lends itself to closer reflection. In this section, we illustrate identified inter-
sections and interfaces of participants. The Strategic Dependency (SD) model
of the use case is shown in Figure 1. It was collaboratively designed in our Di-
rewolf iStar Modeler tool [5]. In the following, we give a brief description of the
identified actors and their goals:
Manufacturer The manufacturer’s primary goal is to deliver products and
services to those linked to the farm. There are two sub-goals: First, sell-
ing machines via the dealer, and second, developing and offering innovative
services and machines to customers.
Dealer The dealer has the goal of providing land machines and corresponding
services to farmers and contractors with the sub-goal to sell or lease ma-
chines. The dealer buys machines from the manufacturer and either sells or
leases them for a profit to contractors or directly to farmers.
20
D D
D D
D
D
D
D
D D
D
D
D
D
D
D
D
D
D
D D
D
D D
Fig. 1. Stakeholder Relationships as iStar Strategic Dependency View
Contractor The contractor has the goal of providing technical services to farm-
ers by partially or entirely taking over specific farming processes. This goal
is supported by the two sub-goals to efficiently and flexibly allocate ma-
chines on the one hand and flexibly allocate specialized staff on the other.
He obtains the machines from the dealer.
Farmer The farmer’s goal is to efficiently raise living organisms for food or raw
materials. For this, he has the sub-goals to efficiently use labor, the profitable
sale of goods, the efficient use of the machines, and the efficient use of the
inputs for seed and crop protection.
Farm Management Platform The farm management platform is a new ac-
tor in the agricultural value chain. It can be considered a “new software
system for farmers” and integrates all data from the farm and other data
streams (e.g., weather) to provide data-based services to those linked to the
farm with the sub-goals of being enabling innovative services based on data,
facilitation of sales of machines, and the connection of users with suppliers
and complementors. Platforms can be managed by a central player [1] or
governed by an alliance of different stakeholder organizations [8].
Complementor The complementor aims to offer value-adding digital services
to those linked to the farm with the sub-goals to develop and offer innovative
services to customers, leveraging data inputs, and providing new data.
The farm management platform relies on the resources farming data and ma-
chine data, which it obtains from the contractor, and provides the required data
for the development of services to the complementor. The complementor devel-
ops new services, which are made available via the farm management platform to
the users: The contractor, the dealer, and the manufacturer. Notable in Figure 1
is the dependency direction of the new data services going from bottom to top.
The production dependency cycle (i.e., especially the machine dependency cycle)
starts on the left, traversing the bottom to the right, before finally concluding in
the top. Those two dependencies, the first representing the newly added services
21
and the second representing the previously existing ones, are clearly distinguish-
able. However, there is a link between the two, since the participants depend on
the data service early in the production cycle. In contrast, the new data service
depends on the last participants of the production dependency. In total, thus,
they build one big cycle.
3 Existing vs. Emerging Market Participants
Of particular interest is the relationship between actors from the former tra-
ditional agricultural value chain, i.e. the existing market participants, with the
new extended actors centered around the farming platform, i.e. the emerging
market participants. Therefore, we merge the emerging market participants -
complementor and farm management platform - into one actor and the existing
participants manufacturer, dealer, contractor, and farmer into another actor.
The resulting SR model is depicted in Figure 2. We get two actors, the emerging
Emerging Existing
Market Market
Participants input Participants
data
provide additional handle existing
data input task data historically
ingest D available
D data machine data
provide D handle
machine data
support platform source data D
participants with match market data
business optimization participants execute
staff existing
supplement
allocation tasks
data
human-
leverage existing interface for readable
connect existing D machines
data input market participants new services
machine improve
enable development allocation existing
of data services tasks
offer
innovative
novel usage
D services
develop of existing
innovative resources
services
analyze
data
analyzed
data
Fig. 2. Abstracted Strategic Rationale Model of Agriculture Use Case
market participant, depicted on the left, and the existing market participant, lo-
cated on the right. There is an apparent interdependency of those two abstracted
actors: The emerging market participant serves innovative services used by the
existing market participant and provides the matching of market participants,
such that those can find services and providers. Both of these tasks rely on an
interface for the new services and system, e.g., a platform. On the other hand,
the new participants are dependent on the old participants providing data; all
data in the figure revolves around farm and machine data. Therefore we have
a cyclic dependency, where each participant relies on the other. For analysis of
the abstraction, the interior of the abstracted actors is depicted, emphasizing
the most important newly introduced changes within the relationship of the ex-
isting market participant and the emerging market participant. For a detailed
insight, we refer to the International Data Spaces model [8], as the data spaces’
22
data owner and the data provider are part of our existing market participant.
The interface represents the app store provider, data app store, and broker for
new services/systems, which we understand as a resource in our model rather
than a separate participant and, in our case, is handled by the emerging market
participant. The data consumer and data user correspond to the existing market
participant using the new innovative service.
4 Conclusion & Future Work
In this paper, we modeled an agricultural data platform in the Industry 4.0.
As platform ecosystems in industrial settings are characterized by high com-
plexity in terms of technology layers [10] and relationships [9], and, in addition,
ecosystem interdependencies change as they evolve, we found that the iStar no-
tation is superior to previous textual descriptions or non-formalized graphical
abstractions, which favor static and simple environments, to reflect the ecosys-
tem interdependencies and dynamics. Adding to the predominantly conceptual
and qualitative state of studies on platform ecosystems, future research could use
iStar to model different scenarios and compare these to the expected outcomes
based on the existing conceptual and qualitative literature. Using this formal-
ization, we can compare scenarios better concerning, e.g., centralization, the
structure of the data ecosystem (e.g., alliance-like, one existing participant takes
the additional role of a new participant, etc.), the background of new partici-
pants, and especially the relationship of emerging and existing participants. The
abstracted model emphasizes the significant aspects when considering market
entrants, both for comparing different possible scenarios and for using the ob-
tained model for policy generation – as the big picture is to throw a bridge from
goals to policies to code using conceptual modeling. The introduced formalized
models can be used as a groundwork for code generation tasks for all actors, such
as data access policies, user interfaces, and corresponding data models. An ex-
emplary code generation task is permission management to data streams based
on the modeled parameters. For instance, access rights may be automatically
granted if a relationship edge is drawn in the model; or it may just as well be
withdrawn again in the opposite case. To implement this, we may rely on iStar
extensions like Secure Tropos [7] or STS-ml [3]. While these models are a neces-
sary first step towards policy generation, it is an important one nonetheless, as
we, combined with already existing results for data governance, obtain proper
groundwork for future work and already provide some insights.
The synthesis of these ideas by combining the comparison of platform vari-
ants with code generation may lead to a faster and more holistic analysis of data
ecosystem variants. Ultimately, a repository of available graphical representa-
tions and code structures may facilitate automated, easier and faster decision
support for stakeholders in new data-driven ecosystems.
Acknowledgement. Funded by the Deutsche Forschungsgemeinschaft (DFG,
German Research Foundation) under Germany’s Excellence Strategy - EXC-
2023 Internet of Production - 390621612.
23
References
1. Adner, R.: Ecosystem as Structure. Journal of Management 43(1), 39–58 (2017).
https://doi.org/10.1177/0149206316678451
2. Chakrabarti, A., Quix, C., Geisler, S., Pullmann, J., Khromov, A., Jarke, M.:
Goal-Oriented Modelling of Relations and Dependencies in Data Marketplaces. In:
Goulão, M., et al. (eds.) Proceedings of the 11th International i* Workshop co-
located with the 30th International Conference on Advanced Information Systems
Engineering (CAiSE 2018), Tallinn, Estonia, June 12, 2018. CEUR-WS Proceed-
ings, vol. 2118 (2018), http://ceur-ws.org/Vol-2118/iStar2018_paper_4.pdf
3. Dalpiaz, F., Paja, E., Giorgini, P.: Security Requirements Engineering
via Commitments. In: 2011 1st Workshop on Socio-Technical Aspects in
Security and Trust (STAST). pp. 1–8. IEEE (08092011 - 08092011).
https://doi.org/10.1109/STAST.2011.6059249
4. Jarke, M.: Data Sovereignty and the Internet of Production. In: Advanced Infor-
mation Systems Engineering. Lecture Notes in Computer Science, Springer Inter-
national Publishing, Cham, Switzerland (2020)
5. Koren, I., Klamma, R., Jarke, M.: Direwolf Model Academy: An Extensible Col-
laborative Modeling Framework on the Web. In: Michael, J., et al. (eds.) Com-
panion Proceedings of Modellierung 2020 Short, Workshop and Tools & Demo
Papers, Vienna, Austria, February 19-21, 2020. CEUR-WS Proceedings, vol. 2542,
pp. 213–216 (2020), http://ceur-ws.org/Vol-2542/MOD20-TuD5.pdf
6. Kritikos, M.: Precision agriculture in Europe: Legal, social and eth-
ical considerations. European Parliament, Brussels, Belgium (2017).
https://doi.org/10.2861/278
7. Mouratidis, H., Giorgini, P.: Secure Tropos: A Security-Oriented Ex-
tension of the Tropos Methodology. International Journal of Soft-
ware Engineering and Knowledge Engineering 17(02), 285–309 (2007).
https://doi.org/10.1142/S0218194007003240
8. Otto, B., Jarke, M.: Designing a multi-sided data platform: findings from
the International Data Spaces case. Electronic Markets 29(4), 561–580 (2019).
https://doi.org/10.1007/s12525-019-00362-x
9. Schermuly, L., Schreieck, M., Wiesche, M., Krcmar, H.: Developing an Indus-
trial IoT Platform - Trade-off between Horizontal and Vertical Approaches. In:
Ludwig, T., Pipek, V. (eds.) Human Practice. Digital Ecologies. Our Future. 14.
Internationale Tagung Wirtschaftsinformatik (WI 2019), February 24-27, 2019,
Siegen, Germany. pp. 32–46. University of Siegen, Germany / AISeL (2019),
https://aisel.aisnet.org/wi2019/track01/papers/3
10. Sisinni, E., Saifullah, A., Han, S., Jennehag, U., Gidlund, M.: In-
dustrial Internet of Things: Challenges, Opportunities, and Directions.
IEEE Transactions on Industrial Informatics 14(11), 4724–4734 (2018).
https://doi.org/10.1109/TII.2018.2852491
11. Trendov, N.M., Varas, S., Zenf, M.: Digital Technologies in Agriculture and Rural
Areas: Briefing Paper. Food and Agriculture Organization of the United Nations
(2019), http://www.fao.org/3/ca4887en/ca4887en.pdf
12. van Alstyne, M.W., Parker, G.G., Choudary, S.P.: Pipelines, Platforms, and the
New Rules of Strategy. Harvard Business Review 94(4), 54–62 (2016)
13. Van Dyck, M., Lüttgens, D.: Design Faktoren und Strategien für digitale Plattform-
geschäftsmodelle im B2B-Kontext am Beispiel der Agrarindustrie. In: Gausemeier,
J., Bauer, W., Dumitrescu, R. (eds.) Vorausschau und Technologieplanung. pp.
215–232 (2019)
24