Two papers have been accepted for publication that are related to EA Debts research. The first paper presents a workshop approach to identify EA Debts and their symptoms (i.e., EA Smells) in organization and illustrates it on two case studies. The work will be presented at the SoEA4EE workshop hosted at the EDOC conference. The abstract is as following:
The Enterprise Architecture (EA) discipline evolved during the past two decades and is now established in a large number of companies. Architectures in these companies changed over time and are now the result of a long creation and maintenance process. Such architectures still contain processes and services provided by legacy IT systems (e.g., systems, applications) that were reasonable during the time they were created but might now hamper the introduction of better solutions. In order to support handling those legacies, research on the notion of EA debts has been started. The concept of EA debts widens the scope of technical debts to cover also organizational aspects offering a mean for managing EA in dynamic environments. The research encompasses the development of methods for managing debts together with a repository of typical EA debts. Identifying EA debts for the repository is challenging as required knowledge is usually not documented. Therefore, a structured approach is needed to externalize this knowledge. The paper presents a workshop format that is used to identify EA debts in organizations. Corresponding workshops are performed in two distinct companies to support them in understanding certain issues they face. First results from those workshops are presented in the second part of the paper.
The second paper analyses EA Smells in EA models that are first converted to their graph representation (see the figure below). Then cypher queries are used to determine if there are smells in the respective model. First experiments show that the new solution performs better than the previously proposed prototype, while being more flexible. The work will be presented at the PoEM conference that might be happen as hybrid event at the end of November. The abstract is as follows:
Hitherto, the concept of Enterprise Architecture (EA) Smells has been proposed to assess quality flaws in EAs and their models. Together with this new concept, a catalog of different EA Smells has been published and a first prototype was developed. However, this prototype is limited to ArchiMate and is not able to assess models adhering to other modeling languages. Moreover, the prototype is not integrate-able with other EA standard tools. Therefore, we propose to enhance the extensible Graph-based Enterprise Architecture Analysis (eGEAA) platform that relies on Knowledge Graphs. To align these two approaches, we show in this paper, how ArchiMate models can be transformed into Knowledge Graphs and provide a set of queries on the Knowledge Graph representation that are able to detect EA Smells. This enables enterprise architects to assess EA Smells on all types of EA models as long as there is a Knowledge Graph representation of the model. Finally, we show that eGEAA performs better than the previous prototype implementation by analyzing a set of 347 EA models.
Architecture of eGEAA platform
Related Publication
Jung, Jürgen; Hacks, Simon; Gooijer, Thijmen; Kinnunen, Matti; Rehring, Kevin
@inproceedings{9626297,
title = {Revealing Common Enterprise Architecture Debts: Conceptualization and Critical Reflection on a Workshop Format Industry Experience Report},
author = {Jürgen Jung and Simon Hacks and Thijmen Gooijer and Matti Kinnunen and Kevin Rehring},
doi = {10.1109/EDOCW52865.2021.00058},
year = {2021},
date = {2021-12-01},
urldate = {2021-01-01},
booktitle = {2021 IEEE 25th International Enterprise Distributed Object Computing Workshop (EDOCW)},
pages = {271-278},
abstract = {The Enterprise Architecture (EA) discipline evolved during the past two decades and is now established in a large number of companies. Architectures in these companies changed over time and are now the result of a long creation and maintenance process. Such architectures still contain processes and services provided by legacy IT systems (e.g., systems, applications) that were reasonable during the time they were created but might now hamper the introduction of better solutions. In order to support handling those legacies, research on the notion of EA debts has been started. The concept of EA debts widens the scope of technical debts to cover also organizational aspects offering a mean for managing EA in dynamic environments. The research encompasses the development of methods for managing debts together with a repository of typical EA debts. Identifying EA debts for the repository is challenging as required knowledge is usually not documented. Therefore, a structured approach is needed to externalize this knowledge. The paper presents a workshop format that is used to identify EA debts in organizations. Corresponding workshops are performed in two distinct companies to support them in understanding certain issues they face. First results from those workshops are presented in the second part of the paper.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The Enterprise Architecture (EA) discipline evolved during the past two decades and is now established in a large number of companies. Architectures in these companies changed over time and are now the result of a long creation and maintenance process. Such architectures still contain processes and services provided by legacy IT systems (e.g., systems, applications) that were reasonable during the time they were created but might now hamper the introduction of better solutions. In order to support handling those legacies, research on the notion of EA debts has been started. The concept of EA debts widens the scope of technical debts to cover also organizational aspects offering a mean for managing EA in dynamic environments. The research encompasses the development of methods for managing debts together with a repository of typical EA debts. Identifying EA debts for the repository is challenging as required knowledge is usually not documented. Therefore, a structured approach is needed to externalize this knowledge. The paper presents a workshop format that is used to identify EA debts in organizations. Corresponding workshops are performed in two distinct companies to support them in understanding certain issues they face. First results from those workshops are presented in the second part of the paper.
In: Serral, Estefanía; Stirna, Janis; Ralyté, Jolita; Grabis, Jānis (Ed.): The Practice of Enterprise Modeling, pp. 48–63, Springer International Publishing, Cham, 2021, ISBN: 978-3-030-91279-6.
@inproceedings{10.1007/978-3-030-91279-6_4,
title = {Using Knowledge Graphs to Detect Enterprise Architecture Smells},
author = {Muhamed Smajevic and Simon Hacks and Dominik Bork},
editor = {Estefanía Serral and Janis Stirna and Jolita Ralyté and Jānis Grabis},
doi = {10.1007/978-3-030-91279-6_4},
isbn = {978-3-030-91279-6},
year = {2021},
date = {2021-11-15},
urldate = {2021-11-15},
booktitle = {The Practice of Enterprise Modeling},
pages = {48--63},
publisher = {Springer International Publishing},
address = {Cham},
abstract = {Hitherto, the concept of Enterprise Architecture (EA) Smells has been proposed to assess quality flaws in EAs and their models. Together with this new concept, a catalog of different EA Smells has been published and a first prototype was developed. However, this prototype is limited to ArchiMate and is not able to assess models adhering to other EA modeling languages. Moreover, the prototype is not integrate-able with other EA tools. Therefore, we propose to enhance the extensible Graph-based Enterprise Architecture Analysis (eGEAA) platform that relies on Knowledge Graphs with EA Smell detection capabilities. To align these two approaches, we show in this paper, how ArchiMate models can be transformed into Knowledge Graphs and provide a set of queries on the Knowledge Graph representation that are able to detect EA Smells. This enables enterprise architects to assess EA Smells on all types of EA models as long as there is a Knowledge Graph representation of the model. Finally, we evaluate the Knowledge Graph based EA Smell detection by analyzing a set of 347 EA models.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Hitherto, the concept of Enterprise Architecture (EA) Smells has been proposed to assess quality flaws in EAs and their models. Together with this new concept, a catalog of different EA Smells has been published and a first prototype was developed. However, this prototype is limited to ArchiMate and is not able to assess models adhering to other EA modeling languages. Moreover, the prototype is not integrate-able with other EA tools. Therefore, we propose to enhance the extensible Graph-based Enterprise Architecture Analysis (eGEAA) platform that relies on Knowledge Graphs with EA Smell detection capabilities. To align these two approaches, we show in this paper, how ArchiMate models can be transformed into Knowledge Graphs and provide a set of queries on the Knowledge Graph representation that are able to detect EA Smells. This enables enterprise architects to assess EA Smells on all types of EA models as long as there is a Knowledge Graph representation of the model. Finally, we evaluate the Knowledge Graph based EA Smell detection by analyzing a set of 347 EA models.
