A New Algorithm for Performance Evaluation of Homogeneous Architectural Styles
Sima Emadi, Aghaee Ghazvini Golnaz · 2012
Software ar chitecture is considered one of the most important indices of software engineer ing today. S oftware A rchitecture is a technical descr iption of a system indicating its component structures a nd their relationships, and is the principles and r ules governing designing. The success of the software depends on whether the system can satisfy the quality attributes. One of the most critical aspects of the quality attributes of a software sy stem is its performance. Performance analysis can be useful for assessing whether a proposed architecture can meet the desired per formance specifications and w hether it can help in m aking key architectural decisions. Architecture sty le is a set of principles which an ar chitect uses in designing software ar chitecture. S ince software ar chitectural sty les have frequently been used by architects, these styles have a specific eff ect on qua lity attributes. If this eff ect is m easurable for each ex isting style, it w ill enable the architect to evaluate and make architectural decisions more easily and precisely. In this paper an eff ort has been made to introduce a new algorithm for investigating performance in hom ogeneous architectural sty les based on M arkov chains. Moreover, How to use this Markov chains model for performance evaluation is shown a nd sim ulation and the implications of this transformation are descr ibed completely. Finally, to represent the usage of our proposed algorithm, a case study is presented as an example. K eyw ord s: Software ar chitecture, M arkov model, Architectural sty les, Performance attribute, homogeneous styles