A conceptual framework for large-scale ecosystem interoperability and industrial product lifecycles
Introduction
Lack of interoperability between computer systems remains one of the largest challenges of computer science and costs industry tens of billions of dollars each year [1], [2]. Standards for data exchange have, in general, not solved the problem: standards are not universal nor universally applied (even within a given industry) leading to heterogeneous ecosystems. These ecosystems comprise large groups of software systems built around different standards that must interact to support the entire system lifecycle. We are currently engaged in the “Oil and Gas Interoperability Pilot” (or simply OGI Pilot), an instance of the Open Industry Interoperability Ecosystem (OIIE) initiative that aims for the automated, model-driven transformation of data during the asset lifecycle between two of the major data standards in the Oil & Gas industry ecosystem. The main standards considered by the project are the ISO15926 suite of standards [3] and the MIMOSA OSA-EAI specification [4]. These standards and their corporate use1 are representative of the interoperability problems faced in many industries today.To enable sensor-to-boardroom reporting, the effort to establish and maintain interoperability solutions must be drastically reduced. This is achieved by developing model transformations based on high level conceptual models.
In our previous work [5] we presented three core contributions: (1) we compared the suitability of different multi-level modelling approaches for the integration of ecosystems in the Oil & Gas industry, (2) introduced the core SLICER (Specification with Levels based on Instantiation, Categorisation, Extension and Refinement) relationship framework to overcome limitations of existing approaches with respect to the definition of object/concept hierarchies, and (3) evaluated the framework on an extended version of the comparison criteria from [6].
The current work extends these contributions by: (1) expanding on the explicit handling of descriptions in the SLICER framework, (2) extending the core SLICER relationships with a complete treatment of attributes, relationships, and their integrity constraints, (3) presenting the formalisation of SLICER core and the treatment of attributes, and (4) illustrating mappings between a SLICER model and alternatives making use of SLICER's finer semantic distinctions to identify patterns of meaning in the original models.
Section snippets
Ecosystem interoperability
The suite of standard use cases defined by the Open O&M Foundation covers the progress of an engineering part (or plant) through the Oil & Gas information ecosystem from initial specification through design, production, sales, deployment, and maintenance including round-trip information exchange. The data transformations needed for interoperability require complex mappings between models covering different lifecycle phases, at different levels of granularity, and incorporating data and
Multi-level modelling techniques
A number of Multi-level modelling (MLM) techniques have been developed to address the shortcomings of the UML-based model. While they improve on the UML-based model in various respects, no current approach fulfils all of the criteria necessary for ecosystem interoperability (see Section 7).
Most current MLM techniques—particularly Deep Instantiation (DI) approaches such as Melanee [9], MetaDepth [10], and Dual Deep Modelling [11]—focus on the reduction of accidental complexity in models [12]. In
A relationship framework for ecosystem modelling
A highly expressive and flexible approach is required to overcome the challenges involved in modelling large ecosystems and supporting transformations across their lifecycles. A key requirement is the identification of patterns of meaning from basic primitive relations that can be identified across modelling frameworks and assist the development of mappings between them. A core observation when building transformations for the real world complexity of the OGI pilot is that a higher level in the
Comparison with dual deep modelling
To illustrate the benefits of SLICER we perform a comparative evaluation against DDM [11] using a common example model. While we compare against the DDM approach, as it is the most flexible and complete of the potency-based MLM frameworks (see Section 7), we limit the comparison to aspects that are applicable across all potency-based methods. The comparison expands on SLICER core (refer Fig. 4) by incorporating a more detailed view of attribute usage in the framework. Fig. 8, Fig. 9 illustrate
Mapping using SLICER distinctions
A key factor in this work is in identifying and explicitly representing the distinguishing factors of relationships: extension, refinement, instantiation, specification, and categorisation. These relationships can be considered as meta-properties that annotate or enrich the families of relationships existing in conventional models. This can be seen as an extension of the classical meta-properties used in OntoClean [45]: Unity (U), whether a concept has wholes as instances and criteria that
Evaluation and comparison
The new relationships and relationship variations in SLICER were developed by examining special cases that are known to cause problems in practice and which obscure modelling concerns with generic domain associations, naming conventions, or both, resulting in highly complex and often inconsistent models [47]. The goal of this approach is to enable development of coherent models in domain areas where persistent and large scale effort has failed to produce workable models since the modelling
Conclusion
Effective exchange of information about processes and industrial plants, their design, construction, operation, and maintenance requires sophisticated information modelling and exchange mechanisms that enable the transfer of semantically meaningful information between a vast pool of heterogeneous information systems. This need increases with the growing tendency for direct interaction of information systems from the sensor level to corporate boardroom level. One way to address this challenge is
Acknowledgments
This research was funded in part by the South Australian Premier’s Research and Industry Fund grant no. IRGP 37.
Matt Selway is a Research Fellow in the School of IT and Mathematical Sciences at the University of South Australia. He was awarded his Doctorate by the University of South Australia in 2016. His research interests include conceptual modelling, knowledge representation, automated software engineering, and natural language understanding. He is working on diverse projects for his group including conceptual modelling for semantic interoperability, and the generation of behaviour/process models
References (50)
- et al.
Ontological and linguistic metamodelling revisited: a language use approach
Inf. Software.Technol.
(2013) - et al.
Supporting ontological analysis of taxonomic relationships
Data Knowl. Eng.
(2001) - N. Young, S. Jones, SmartMarket Report: Interoperability in Construction Industry, Tech. rep., McGraw Hill,...
- Fiatech, Advancing Interoperability for the Capital Projects Industry: A Vision Paper, Tech. rep., Fiatech, Feb....
- ISO, ISO 15926 – Part 2: Data Model,...
- MIMOSA, Open Systems Architecture for Enterprise Application Integration,...
- M. Selway, M. Stumptner, W. Mayer, A. Jordan, G. Grossmann, M. Schrefl, A conceptual framework for large-scale...
- B. Neumayr, M. Schrefl, B. Thalheim, Modeling techniques for multi-level abstraction, in: The Evolution of Conceptual...
- S. Bergamaschi, D. Beneventano, F. Guerra, M. Orsini, Data integration, in: Handbook of Conceptual Modeling, Springer,...
- C. Atkinson, T. Kühne, The Essence of Multilevel Metamodeling, in: Proceedings of the UML 2001, Springer, 2001, pp....
Dual deep modeling: multi-level modeling with dual potencies and its formalization in F-Logic
Softw. Syst. Model.
Reducing accidental complexity in domain models
Soft. Syst. Model.
No silver bullet essence and accidents of software engineering
Computer
Model-driven development: a metamodeling foundation
IEEE Softw.
On the ontological expressiveness of information systems analysis and design grammars
Inf. Syst. J.
A powertype-based metamodelling framework
Soft. Syst. Model.
Multilevel modeling
Bus. Inf. Syst. Eng.
Power types
JOOP
Cited by (24)
Multi-level conceptual modeling: Theory, language and application
2021, Data and Knowledge EngineeringCitation Excerpt :In a previous version of OMLM called MiF [45], the same authors claim that their language does not support cross-level domain relations (F8), even though can potentially be extended in that sense. Selway et al. [39] propose the SLICER conceptual framework, which also accounts for multi-level models. SLICER provides to the user a set of level-aware relations that enable multi-level modeling, such as specializations, instantiations and powertype (“subset by specification”) relations.
Understanding the resilience of coal industry ecosystem to economic shocks: Influencing factors, dynamic evolution and policy suggestions
2020, Resources PolicyCitation Excerpt :The link technology rigidity refers to the technology that enables system members to connect with each other, realizes resource circulation and energy flow, and promotes the formation of a closed-loop by the system, and reflects the system dynamism to a certain extent (Geng and Zhan, 2016). The function positioning refers to the dependence of the system on the original product market under the changes in the external environment, reflecting the skopos of the system (Neves et al., 2019; Selway et al., 2017). From the medium aspect, the influencing factors of CIER mainly come from the industry-level characteristics of structure, behavior, and performance.
TOTEM: Reconciling multi-level modelling with standard two-level modelling
2020, Computer Standards and InterfacesCitation Excerpt :These limitations of the two-level modelling approach have sparked a big interest in multi-level modelling technologies within the MDE community. This interest is witnesses by the MULTI series of workshops dedicated to multi-level modelling [60], a Dagstuhl seminar [1] and the plethora of multi-level modelling tools recently proposed, like DeepTelos [41], DeepJava (Thomas [49]), Deep-Ruby [63], the DPF Workbench [50], Dual Deep Modelling [64], FMMLx [20,31], Kite [35], Melanee [7], MetaDepth [22], ML2 ([17]; Claudenir M [29]), MultEcore [53], NMF [37], OMLM [38], and SLICER [71], among others. Many of these tools offer sophisticated capabilities for modelling in multiple meta-levels, like attribute mutability and durability [33], deep references [25], dual potencies [64], or refactoring support [24].
Towards the Integration of Multi-Level and Multi-View Modelling for Interoperability
2023, Proceedings - 2023 ACM/IEEE International Conference on Model Driven Engineering Languages and Systems Companion, MODELS-C 2023Clabject Typing in MLM - the Double Life of a Clabject: A Position Paper
2023, Proceedings - 2023 ACM/IEEE International Conference on Model Driven Engineering Languages and Systems Companion, MODELS-C 2023
Matt Selway is a Research Fellow in the School of IT and Mathematical Sciences at the University of South Australia. He was awarded his Doctorate by the University of South Australia in 2016. His research interests include conceptual modelling, knowledge representation, automated software engineering, and natural language understanding. He is working on diverse projects for his group including conceptual modelling for semantic interoperability, and the generation of behaviour/process models from textual descriptions.
Markus Stumptner is full Professor of Computing at the University of South Australia, Adelaide, where he directs the Advanced Computing Research Centre. His research interests include conceptual modeling, automated software engineering, knowledge representation, model-based reasoning and data analytics. He is Program Lead - Data Management of the Data to Decisions Collaborative Research Centre. He is on the editorial boards of AI Communications and AI EDAM, on the Board of Directors of MIMOSA (Maintenance Information Management Open Systems Alliance), member of the Australian Computer Society Steering Committee for Artificial Intelligence and fellow of the International Society of Engineering Asset Management.
Wolfgang Mayer is a Senior Lecturer at the University of South Australia. His research interests include knowledge representation and reasoning methods and their applications in industrial practice. He has made contributions to information management and knowledge extraction methods, technologies for software systems interoperability, business process modelling and underlying software implementation, expert systems for process and product customization, and fault diagnosis in software systems. He has published more than 75 refereed publications in journal articles, conference proceedings, and book chapters.
Andreas Jordan completed his Master's Degree in 2009 and is currently a PhD student in the Knowledge and Software Engineering Laboratory based at the University of South Australia. His research interests include conceptual modelling, ontology engineering, and knowledge representation and reasoning methods and their applications in industrial practice. He is currently working on ontology-based semantic interoperability between ISO 15926 and other standards in the Oil and Gas industry.
Georg Grossmann is a Senior Lecturer in the School of IT and Mathematical Sciences at the University of South Australia. He is working on the integration of business processes and complex data structures for systems interoperability and has applied theory successfully in industry projects. Current research interests include integration of service processes, ontology-driven integration and distributed event based systems. He is currently Co-Chief Investigator in the Data to Decisions CRC (D2D CRC), steering committee chair of the IEEE EDOC Conference and secretary of the IFIP WG 5.3 on Enterprise Interoperability.
Michael Schrefl received his Dipl.-Ing. degree and his Doctorate from Vienna University of Technology, Vienna, Austria, in 1983 and 1988, respectively. During 1983–1984, he studied at Vanderbilt University, USA, as a Fulbright scholar. From 1985 to 1992, he was with Vienna University of Technology. During 1987–1988, he was on leave at GMD IPSI, Darmstadt, where he worked on the integration of heterogeneous databases. He was appointed Professor of Information Systems at Johannes Kepler University of Linz, Austria, in 1992, and Professor in Computer and Information Science at University of South Australia in 1998. He currently leads the Department of Business Informatics—Data and Knowledge Engineering at Johannes Kepler University of Linz, with projects in data warehousing, workflow management, and web engineering.