=Paper=
{{Paper
|id=Vol-3214/WS4Paper2
|storemode=property
|title=A Reference Architecture for Data Quality in Smart Manufacturing
|pdfUrl=https://ceur-ws.org/Vol-3214/WS4Paper2.pdf
|volume=Vol-3214
|authors=Francisco Fraile,Arcadio Garcia,Cinzia Rubattino,Sabrina Verardi,Manfredi Giuseppe Pistone
|dblpUrl=https://dblp.org/rec/conf/iesa/FraileGRVP22
}}
==A Reference Architecture for Data Quality in Smart Manufacturing==
https://ceur-ws.org/Vol-3214/WS4Paper2.pdf
A Reference Architecture for Data Quality in Smart
Manufacturing
Francisco Fraile1, Arcadio Garcia2, Cinzia Rubattino3, Sabrina Verardi3 and Manfredi
Giuseppe Pistone 3
1
Universitat Politècnica de València, Camino de Vera s/n, 46022, Valencia, Spain
2
EXOS Solutions, S.L., Camino de Vera s/n, 46022, Valencia, Spain
3
Engineering Ingegneria Informatica S.p.A., Piazzale dell’Agricoltura 24, 00144 Roma
Abstract
The i4Q Project aims to provide a complete set of solutions consisting of IoT-based Reliable
Industrial Data Services (RIDS), the so-called 22 i4Q Solutions, able to manage the huge
amount of industrial data coming from cheap cost-effective, smart, and small size
interconnected factory devices for supporting manufacturing online monitoring and control.
This papers presents the reference architecture used to guide the development of IoT-based
Reliable Industrial Data Services for Manufacturing Quality Control. Based on industry
standards and known best practices, the reference architecture adopts a three-tiered
architectural model to represent the main system architectural components, and provides four
different architectural viewpoints to address business, usage, functional, and implementation
concerns.
Keywords 1
Data quality, industrial IoT, system architecture, reference models
1. Introduction
One of the challenges in implementing quality control processes and solutions is the development
of a reference architecture technologies and on relevant sector specific standards. The Reference
Architecture (RA) must not be designed from scratch, but must take into account the most recent
releases of some relevant initiatives for digital industries, industrial IoT and edge computing - namely
those from the Platform Industrie 4.0 initiative (RAMI 4.0) [1], the Industrial Internet Consortium
(IIRA [2] and OpenFog RA [3]), the Big Data Value Association (BDVA RA) [4], Digital Factory
Alliance Initiative (Digital Service RA) [5], the ISO Standard for the Internet of Things (ISO / IEC
30141) [6] – and contextualize them to the specific application domain. Besides, several research
projects provide a wide set of background knowledge on the topic and represent solid backgrounds to
stem upon.
This paper presents the second release of the i4Q reference Architecture, architecture for IoT-
based Reliable Industrial Data Services (RIDS) [7, 8] ready to deal with large amounts of data in
industrial environments, developed in the i4Q project [9].
2. Methodology
The i4Q Project aims to provide a complete set of solutions consisting of IoT-based Reliable
Industrial Data Services (RIDS), the so-called 22 i4Q Solutions, able to manage the huge amount of
Proceedings of the Workshop of I-ESA’22, March 23–24, 2022, Valencia, Spain
EMAIL: ffraile@cigip.upv.es (F. Fraile); arcadio.garcia@exos-solutions.com (A. Garcia); cinzia.rubattino@eng.it (C. Rubattino);
sabrina.verardi@eng.it (S. Verardi); manfredigiuseppe.pistone@eng.it (M.G Pistone)
ORCID: 0000-0003-0852-8953 (F. Fraile); 0000-0001-5370-0527 (C. Rubattino); 0000-0002-3563-8683 (S. Verardi)
© 2022 Copyright for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
Wor
Pr
ks
hop
oceedi
ngs
ht
I
tp:
//
ceur
-
SSN1613-
ws
.or
0073
g
CEUR Workshop Proceedings (CEUR-WS.org)
industrial data coming from cheap cost-effective, smart, and small size interconnected factory devices
for supporting manufacturing online monitoring and control.
The development of the i4Q reference architecture (i4Q RA) followed an architecture-centered,
scenario-driven, iterative development process, based on the ISO/IEC/IEEE 42010 “Systems and
software engineering – Architecture” [10], starting from an initial released that has been reviewed and
refined in consecutive releases. The first release of the i4Q RA, was not designed from scratch, being
strongly based on the most relevant outcomes of other previous research activities and releases of
International Communities, compiled during the initial analysis of related standards and reference
models. According to the analysis and the evaluation performed, it was mainly inspired by the
Industrial Internet Reference Architecture (IIRA) architectural model, based on the three-tiers
approach [11]. Also, according to the IIRA approach, the task of the i4Q RA definition was
performed in parallel with those of its viewpoints analysis, namely the business viewpoint, usage
viewpoint, functional viewpoint, and implementation viewpoint. The viewpoints offer a framework to
think iteratively through the architectural issues that arose throughout the definition of the system
architecture.
Since the first release, a strong connection between the RA and the 22 i4Q Solutions (17 tools and
5 guidelines) was defined. In order to have a coherent inclusive vision provided by the i4Q RA, a
mapping activity of i4Q RIDS against it has been started since the beginning and feedback’s
collection from the solution providers were fundamental to refine the architecture and its building
blocks, and to develop the different viewpoints.
This way, the architecture and the viewpoints are mutually correlated and continuous feedback
from each other need to be exchanged following an iterative approach to guarantee the alignment
between the different elements of the Reference Framework.
So, the analysis across the four key viewpoints served as input for the architecture, which at the
same time influenced them. The combination of the results obtained from the different viewpoints
was derived into detailed reference architecture. The results included: business, regulatory and
stakeholders' key inputs (from the business viewpoint); an identification of tasks, roles or activities to
be performed by the framework (from usage viewpoint); a decomposition into its Control, Operations,
Information, Application and Business domains (from the functional viewpoint); an identification of
associated flows and analysis and selection of the technologies required for its implementation (from
implementation viewpoint).
Besides the architectural viewpoints, the i4Q RA also takes into account all the considerations
coming from the use case scenarios, the requirements’ collection and the specifications of the i4Q
solutions. Special attention is devoted to the pilots' needs involving them in the activities of the
viewpoints’ definition and defining blueprints.
Another key input has been the definition of digital models and ontologies to be used in the i4Q
Framework. The analysis of the reference models performed in the establishment of the first version
of the reference architecture represented the basis for mapping activity to determine the best data
models and ontologies for each architectural tier.
The definition of the use case scenarios and requirements analysis and functional specification
paved the way for an in-depth analysis of the i4Q Solutions, performed in close collaboration with
Solution Providers, to better understand how they fit into the RA and its different viewpoints. The
architectural viewpoints, using the value-driven approach suggested in IIRA, were incrementally
improved, enabling top-bottom (i.e., from the business perspective down to the technical
implementation perspective) and bottom-up feedback loops organizing meetings to make sure that the
different viewpoints are consistent. The operational needs characterizing the realization, deployment
and use of each solution represented a valid input and helped to better identify the RA, refining its
modules, components and functionalities.
The definition of the second version of the RA also took into account all the considerations
formalized in the description of use case scenarios and the requirements’ collection; special attention
was paid to the needs emerging from the pilots, which have been mapped in detailed blueprints to
verify the alignment between the specific components to be used for industrial problems and the RA
building blocks.
It is therefore clear that the design of the i4Q RA has been an iterative and parallel process, a based
on a mix of top-down and bottom-up approaches, in which the results provided by several activities
have been gathered and included in a systematic project vision. All of the information gathered in this
reference architecture lays the groundwork for future project steps.
3. Reference architecture
The i4Q Reference Architecture (i4Q RA) defines a conceptual framework aiming to be the canvas
for the design, implementation and integration of the i4Q Solutions. Being strongly based on the most
relevant outcomes of other previous Research and Innovation activities and releases of International
Communities, the i4Q RA is aligned to the most common standard reference architectures in the
manufacturing domain and is mainly inspired by IIRA architectural model [2].
In particular, two approaches have been adopted:
• The IIRA three-layers architectural pattern as the main design driver.
• The IIRA layered databus pattern in order to make the solution data-driven.
The mapping between the layered databus pattern with the three-tier architecture pattern is as
follows:
• The Machine Databus is mapped with the physical Assets and Smart Products.
• The Unit Databus is mapped with the Edge Tier.
• The Site Databus is mapped with the Platform and Enterprise Tiers.
• The Inter-Site Databus enables ecosystems and inter-factory communications.
The final version of i4Q RA fulfills the considerations coming from the viewpoints’ developments,
the pilots' needs and solution providers mapping activity.
.
. Figure 1: The i4Q Reference Architecture
The i4Q RA Security and Compliance Layer supports the communication between all tiers,
providing a common security layer for accessing and retrieving data and physical assets (sensing and
actuation).
Furthermore, the i4Q RA can support data (and services) description through the most significant
meta- models (i.e., IIRA Characteristics of IIoT Information Models [12]), common ontologies and
digital models. Finally, the Interoperability Assurance Layer guarantees interoperability inside the
i4Q ecosystem considering the broad-spectrum of raw and elaborated data that can be processed by
different solutions.
3.1. Edge tier
The Edge Tier aims to implement the sensor and control domain. It allows connecting physical
assets and systems, collecting and processing data in the “proximity network”. Furthermore, it
supports Digital Twin (enabling advanced use, i.e., data filtering for generating a more precise virtual
object, real-time analytics, etc.) and the Distributed Computing for sustaining distributed workloads
and models. The Edge Tier, typically, process real-time data (Data in Motion).
The Edge Tier includes applications and services for describing and managing data, supports
digital twins reducing the connectivity and latency issue, and the distributed computing devising a
typical edge-cloud orchestration system.
Edge Tier consists of the following building blocks:
• Data Collecting: Solutions for data ingestion, collecting raw data from the facilities and store
them to the data lake or make them available for the processing.
• Data Management: Solutions for data management, transformation (typically pre-processing),
harmonization and loading.
• Data Services: Data Services to enable ingestion use and maintenance.
• Digital Twin: Digital Twin (Operational Services) to achieve a connected 3D production
simulation with a digital twin for manufacturing.
• Distributed Computing: Fog Computing/Edge nodes that allow deploying and running AI
workloads on the edge enable efficient analysis.
3.2. Platform tier
The Platform Tier contains applications and services for supporting the data flow (abstracted data)
acting as a middleware between the Edge Tier and the Enterprise Tier (and between access and
service network). It supports services for quality control (modelling, orchestration, traceability, etc.)
and historical data processing (Data at Rest).
The Platform Tier includes subcomponents for integrating third-party systems and architectures,
Digital Twin services for interacting with the edge-related application, and distribution services (i.e.,
resource management, data, and algorithms orchestration, etc.).
Platform Tier consists of the following building blocks:
• Models Management and Services: Management of Models and related Services enabling
holistic vision of data across i4Q Infrastructure.
• Orchestration Management: Services for orchestration management enabling high resiliency
and responsiveness workloads.
• Distribution Services: Distribution and deployment services enabling plug & play
reconfiguration and installation of workloads.
• Traceability Services: Robust and Rapid Traceability Service (i.e., DLT) to provide an audit
trail for all inserted data, guaranteeing immutability and finality.
• Digital Twin Services: Digital Twin (Management Services) services that enable industrial
companies to virtual validation/visualization and productivity optimization using pre-existing
and/or simulated data and data from different factory levels.
• Data Brokering and Storage: Services for data Brokering and storage management supporting
a high degree of digitization in companies with most manufacturing devices acting as sensors
or actuators and generating vast amounts of data.
• Data Transformation: Services for data transformation (typically post-processing).
• Data Analytics and Services: Services for data analytics on top of the data infrastructure with
several incremental algorithms (i.e. operating on data streams with fast incremental updates)
suitable for analytic processing of high-speed data streams.
• Monitor and Diagnostics: Services for near/real-time monitoring and diagnostics monitoring
the health of workloads, predict problems and take corrective actions, predict failures and
provide alerts.
• Assets Management: Services for assets management for managing and track facilities’ assets.
3.3. Enterprise tier
The Enterprise Tier implements business-specific applications related to quality control, providing
interfaces for end-users. It integrates engineering and management applications for supporting
command generations, operational intelligence operations, big-data analytics and service
orchestration.
The Enterprise Tier includes an “open” (not exhaustive) catalogue of use case/domain-specific
applications. The usage of standard APIs and commons ontologies drives the definition of the
common interfaces and facilitates the integration of (user and system) legacy and newer applications.
Enterprise Tier consists of the following building blocks:
• Biz Analytics for Quality: Applications built on top of underneath services (e.g: Models
Management and Services, Data Analytics and Services) aiming to pursue excellence in
quality focusing on zero-waste context.
• Human Collaboration and Operational Intelligence: CRM, CSM, DSS and ERP/MES and
other services that provide smart alerting and quality diagnosis.
• Engineering and Planning Applications: CAD, CAE, CAM and AR/VR and everything related
to the joint and integrated use of software systems for computer-aided design and computer-
aided manufacturing.
• Orchestration and Distribution Designer: Engineering Environment for defining pipelines and
distributions such as a scalable, easy to use, policy-based distribution mechanism to ease the
task of distributing AI/ML models and other metadata to the edge.
• Simulation and Optimization: Simulation and Optimization applications for optimizing
manufacturing processes.
3.4. Data space
The Data Space Tier extends and enhances interoperability with future manufacturing data spaces
by the means of IDS Connectors, according to main principles defined by International Data Space
Association (IDSA) [13].
The International Data Spaces (IDS) is a virtual data space that leverages existing standards,
technologies and governance models well accepted in the data economy, to facilitate secure and
standardize data exchange and data linkage in a trusted business ecosystem. It thus provides a
foundation for facilitating innovative cross-company business processes and creating smart-service
scenarios and guaranteeing data sovereignty for data owner [13]. i4Q Reference Architecture, thus,
has been enriched with data sovereignty and sharing capabilities taking into account the needs of a
constantly updated knowledge base to determine and present quality factors in manufacturing.
Data Space Tier consists of the following building blocks:
• Industrial Agreement and Coordination framework: A guidance to satisfy requirements in
order to obtain a certification, including business, organizational and operational agreement.
• IDS Metadata Broker: A comprehensive connector that provides the necessary interfaces for
communicating with any other International Data Spaces connector.
• IDS Identity Provider: An intermediary offering services to create, maintain, manage and
validate identity information and for Participants in the International Data Spaces.
• IDS Clearing House: An intermediary providing clearing and settlements services for all
financial and data exchange transactions within the International Data Spaces.
• IDS Connectors: A dedicated communication server for sending and receiving data in
compliance with the general Connector specification.
4. Discussion
The design of the i4Q RA followed an iterative approach and the results provided by several
activities have been gathered and included in a systematic project vision that is currently used to guide
the development of the envisioned i4Q solutions.
The first version of i4Q RA has been taken as the basis, continuing the mapping activities of the
i4Q Solutions against it and collecting feedback from solution providers. Another input for the
architecture consisted in the analysis across the four key viewpoints (business, usage, functional and
implementation), the architecture and the viewpoints are mutually correlated and continuous feedback
from each other. The iterative approach followed for the implementation of the viewpoints was
crucial to guarantee the alignment between the different elements of the Reference Framework.
Another key input is the definition of digital models and ontologies to be used in the i4Q
Framework, of which analysis represents the basis for mapping activity to establish the best data
models and ontologies for each architectural tier.
The definition of the second version of the i4Q RA has also taken into account activities carried
out in the description of use case scenarios, the requirements’ collection and the specifications of the
i4Q RIDS. Special attention was devoted to the pilots' needs involving them in the activities of the
viewpoints’ definition and defining blueprints.
The second and final version of the reference architecture will serve as input for the
implementation activities that will take place in i4Q project as well as sound basic for further
European research projects.
5. References
[1] DIN SPEC91345:2016-04 - Reference Architecture Model Industrie 4.0 (RAMI4.0), 2016. URL:
https://www.din.de/en/wdc-beuth:din21:250940128
[2] Industrial Internet Consortium, The Industrial Internet of Things - Volume G1: Reference
Architecture, 2019. URL: https://www.iiconsortium.org/pdf/IIRA-v1.9.pdf
[3] OpenFog, OpenFog Reference Architecture for Fog Computing, 2017. URL:
http://site.ieee.org/denver-com/files/2017/06/OpenFog_Reference_Architecture_2_09_17-
FINAL-1.pdf
[4] BDVA, European BDVA Strategic Research and Innovation Agenda, 2017. URL:
http://www.bdva.eu/sites/default/files/BDVA_SRIA_v4_Ed1.1.pdf
[5] Digital Factory Allience, Transforming Manufacturing Together. URL:
https://digitalfactoryalliance.eu
[6] ISO, Internet of Things (IoT) - Reference Architecture, 2018. URL:
https://www.iso.org/standard/50508.htm
[7] R. Poler, A. Karakostas, S. Vrochidis, A. Marguglio, S. Gálvez-Settier, P. Figueiras, ... C.
Agostinho, An IoT-based Reliable Industrial Data Services for Manufacturing Quality Control,
in: 2021 IEEE International Conference on Engineering, Technology and Innovation
(ICE/ITMC), IEEE, New York, 2021, pp. 1-8. doi: 10.1109/ICE/ITMC52061.2021.9570203
[8] A. Karakostas, R. Poler, F. Fraile, S. Vrochidis, Industrial data services for quality control in
smart manufacturing–the i4q framework, in: 2021 IEEE International Workshop on Metrology
for Industry 4.0 & IoT (MetroInd4. 0&IoT), IEEE, New York, 2021, pp. 454-457. doi:
10.1109/MetroInd4.0IoT51437.2021.9488490.
[9] CORDIS, Industrial Data Services for Manufacturing Quality Control in Smart Manufacturing,
2021. URL: https://cordis.europa.eu/project/id/958205/en
[10] ISO, ISO/IEC/IEEE 42010:2011 Systems and software engineering - Architecture description,
2011. URL: https://www.iso.org/standard/50508.html
[11] S. W. Lin, B. Miller, J. Durand, R. Joshi, P. Didier, A. Chigani, ... G. Bleakley, Industrial
Internet Reference Architecture, 2015. URL:
https://hub.iiconsortium.org/iira#:~:text=Industrial%20Internet%20Reference%20Architecture%
20(IIRA)&text=The%20IIRA%20is%20a%20standards,guide%20technology%20and%20standa
rd%20development.
[12] IIC, Characteristics of IIoT Information Models, 2020. URL:
https://www.iiconsortium.org/pdf/Characteristics-of-IIoT-Information-Models.pdf
[13] IDSA, Reference Architecture Model, 2019. URL: https://internationaldataspaces.org/