On the Conceptualization of Tools for Traceability-based Change Impact Analysis using a Domain-specific Modeling Approach

In the context of system architecture design, design changes are deemed as an important issue in the field of cyber-physical systems. This is due to their costs that can increase up to several folds with each step of the life cycle, from conceptual design to operation. Accordingly, before putting such changes into practice, their impacts need to be analyzed based on traceability of dependencies among elements in system artifacts. However, traceability-based Change Impact Analysis (CIA) is time consuming and error-prone process, if performed manually. Thus, a number of tools have been developed in this regard. Yet, in the absence of approaches guiding the specification and implementation of such tools, research reveals their inadequacy for the needs of the industry. To overcome this, this paper proposes a domain-specific modeling approach for the conceptualization of traceability-based CIA tools. Such approach is composed of three building blocks. First, a language for traceability-based CIA designed around the concept of "CIA parameters" (i.e., all kind of information that could contribute to quantitative or qualitative change impact assessment). Second, a modeling procedure referring to a multi-view modeling approach that explains how to use the traceability language to allow creation and machine processing of models (i.e., analysis of models in terms of change impact). Third, mechanisms and algorithms focusing on the implementation of the modeling procedure on a metamodeling platform. The proposed approach is thus an attempt to bridge the gap between design and implementation of traceability-based CIA tools. The approach is illustrated on an example from the aviation field.