Detecting Enterprise Architecture Smells based on Software Architecture Smells

Abstract: Software architecture (SA) smells are design problems in the internal structure and behavior of an SA. These can be seen as a specific category under the umbrella concept of technical debt (TD). TD is a central concept in software development projects and having the means to detect and measure the smells is essential to understand impairments they may cause. However, TD is only limited to the technical aspects and does not describe smells found on an enterprise level. Enterprise architecture debt (EAD) expands the concepts of TD beyond the technical aspects such that it covers the debts that can be found in all layers of an Enterprise Architecture (EA). EA smells give a measurement for EAD by providing means for identifying and detecting the smell, hence enabling a method to quantify the level of debt. The goal of this thesis is to find EA smells derived from existing SA smells. In total, three new EA smells were presented based on existing SA smells. Each new smell was described by a short description that informally summarizes the smell. This was followed by an indication of the smell’s origin and reasoning about the effects on the quality. Then, an illustrative example of the smell was provided. Finally, a detection algorithm was also provided and implemented in a prototype detection program. This thesis serves as a basis for measurements of the quality of an EA and motivation for future research in this area. It is argued that the finding of EA smells can facilitate quality assessment in an EA.