Quality-oriented software architecture

Abstract

This paper proposes a method to select the software architecture for a family of systems that meets user requirements. The method is based on a set of universally accepted design principles and tactics that aims to establish a systematic correlation between the quality requirements of a system and an architectural style that best implements it. The proposed method can also be a valuable assistant to a novice software engineer in selecting an optimal architectural style because the selection of an appropriate architectural style plays an important role in success or failure of a system. IntroductionSoftware architecture is the very first step in the software lifecycle in which the nonfunctional requirements (NFR) are addressed [7,8]. The nonfunctional requirements (e.g., security) are the ones that are blamed for a system re-engineering, and they are orthogonal to system functionality [7]. Therefore, software architecture must be confined to a particular structure that best meets the quality of interest because the structure of a system plays a critical role in the process (i.e., strategies) and the product (i.e., notations) used to describe and provide the final solution.In this paper, we propose a quality driven approach to software architecture that attempts to bridge the chasm between the problem domain, namely requirement speculations, and the first phase in the solution domain, namely software architecture. The proposed method is a systematic approach based on the fact that the functionality of any software system can be met by all kinds of structures but the structure that also supports and embodies non-functional requirements (i.e., quality) is the one that best meets user needs. To this end, we have developed a method based on non-functional requirements of a system; the method employs computational model, communicational model, and coordination model in order to identify the corresponding set of nonfunctional requirements (NFRs) associated to that model. We then use NFRs to select the software architecture for a family of systems. The identification of the computational (i.e., the nature of computation activities), communicational (i.e., the nature of interactions), and the coordination models (i.e., the nature of binding by separate components are put together to become one). As result, the identifications of computation, communication, and coordination models should minimize the original set of requirements to what is known as a set of architecturally significant requirements or nonfunctional requirements (NFRs) of systems.Examples of computational model based on sequential behavior (or concurrent) include state-based, event-based, object-based, data-based, or feedback-based models. Examples of communication model include message passing, such as query-based, event-based, shared memory, and data stream. Examples of coordination models include blocking (or synchronized) and non-blocking (or asynchronized).To identify the models of a system (i.e., computation, communication, and coordination), we can app...

Evaluation of the Software Architecture Styles from Maintainability Viewpoint

“…As compared to [4], [8], both the proposed method and the method used in [9] give the quantitative results about the maintainability of SASs that is basis of the systematic recommendation and selection of SAS.…”

“…These methods verify specific features such as consistency and satisfaction of some properties by SASs that are different from the quality attributes required in this paper. The above points and table 8 clearly show the position of the proposed method as compared to the methods of [4], [8] and [9].…”

“…The impact of a distributed software system's architectural style on the system's energy consumption has been estimated in [3]. A method for specifying the relation between six SASs and quality attributes such as maintainability has been proposed in [8]. The relationship between the quality attributes, design principles and some SASs has been specified in [8] using a tree-based framework.…”

“…A method for specifying the relation between six SASs and quality attributes such as maintainability has been proposed in [8]. The relationship between the quality attributes, design principles and some SASs has been specified in [8] using a tree-based framework. In [4], the impacts of SASs on quality attributes are determined based on the description of style in [2].…”

“…In [4], the impacts of SASs on quality attributes are determined based on the description of style in [2]. The methods offered in [4] and [8] are not able to determine the amount of style support from quality attributes, do not offer quantitative results about their maintainability, and are not precise. SASs are evaluated in [9] from maintainability viewpoint based on the scenario-based evaluation method that is less precise, less reliable and less analyzable as compared to the measurement-based evaluation method utilized in this paper.…”