Exploring Automotive Stakeholder Requirements for Architecture Optimization Support

The software product line (PL) architecture (PLA) is one of the most important PL core assets as it is the abstraction of the products that can be generated, and it represents similarities and variabilities of a PL. Its quality attributes analysis and evaluation can serve as a basis for analyzing the managerial and economical values of a PL. We proposed metrics for PLA complexity and extensibility quality attributes. This paper is concerned with the empirical validation of such metrics. As a result of the experimental work we can conclude that the metrics are relevant indicators of complexity and extensibility of PLA by presenting their correlation analysis.

As society is evolving in a digital direction, we should be more oriented in developing new solutions to reduce the investment costs for digitalization solutions. Digitalization foresees the integration of stakeholders in ecosystems, e.g., a factory, an airport, and a bridge. Such integration requires data sharing among different local automation systems owned by different legal bodies, which are possibly located in different countries under their legal systems. The sharing of data enables the optimization of productivity, raw material yield, energy and environmental footprint, etc., which is the basic motivation for automation. Scalability is another key characteristic of a future product, services, and production line architecture. It is a significant investment and should fit for volume and premium products.

The Product Line Architecture (PLA) can be improved by taking into account key factors such as feature modularization, and by continuously evaluating its design according to metrics. Search-Based Software Engineering (SBSE) principles can be used to support an informed-design of PLAs. However, existing search-based design works address only traditional software design not considering intrinsic Software Product Line aspects. This paper presents MOA4PLA, a search-based approach to support the PLA design. It gives a multi-objective treatment to the design problem based on specific PLA metrics. A metamodel to represent the PLA and a novel search operator to improve feature modularization are proposed. Results point out that the application of MOA4PLA leads to PLA designs with well modularized features, contributing to improve features reusability and extensibility. It raises a set of solutions with different design trade-offs that can be used to improve the PLA design.

Microservice has been successfully employed in software industry [1, 11], as they provide modularization and easy management of small and autonomous services, high availability, scalability and short time-to-market. A recent study on microservices shows that most studies generate specific solutions, which emphasize the need for fundamental research, proposals of reusable practices and works that focus on providing information to ease communication between architects and stakeholders [11] -- software architecture can be a powerful tool for this regard. The definition and documentation of the software product line architecture (PLA) is an important activity, especially for inception and extraction of microservice-oriented PLA, because they involve decisions about how to design customizable microservices, how to arrange the communication between microsservices and APIs, etc. In this work, it is proposed a metamodel for the specification of microsserve-oriented PLA design in order to assist the developer in carrying out such an activity. The proposed metamodel was validated in an exploratory study, in which a new PLA was designed through the instantiation of the proposed metamodel and a product was configured from the designed PLA. Both the metamodel and its instantiation were evaluated in a survey involving software developers. The results indicate that the metamodel addresses the structural needs of microservice-oriented architectures. The main contributions of this work are (i) to assist the software product line developers in the specification and documentation of microservice-oriented PLA design and (ii) the lessons about the industrial practice learned from the surveys that are useful to enhance the proposed metamodel.

This paper focuses on an industrial experience with software product lines of analytics-enabled solutions, specifically the evolution of the software product line architecture for a Subject Matter Expert Workbench toolset which supports analytic plugins for multiple software product lines. As context, the toolset product line was intended for integration of expert knowledge into a family of industrial asset health applications at runtime. The toolset architecture is now being evolved to build and manage plugins for multiple Industrial Analytics solutions (software systems and services) beyond asset health. This evolution is driving changes in the desired architecture qualities of the toolset; widening the stakeholder pool and influencing priorities; affecting the architecture tradeoffs and decisions; and triggering updates to the product line architecture, the guidance for applying it, and the current prototype of the toolset. We describe our experiences in handling this evolution, assess lessons learned, and discuss potential relevance to other product line scenarios.

The Product Line Architecture (PLA) design is a complex task, influenced by many factors such as feature modularization and PLA extensibility, which are usually evaluated according to different metrics. Hence, the PLA design is an optimization problem and problems like that have been successfully solved in the Search-Based Software Engineering (SBSE) area, by using metaheuristics such as Genetic Algorithm. Considering this fact, this paper introduces a tool named OPLA-Tool, conceived to provide computer support to a search-based approach for PLA design. OPLA-Tool implements all the steps necessary to use multi-objective optimization algorithms, including PLA transformations and visualization through a graphical interface. OPLA-Tool receives as input a PLA at the class diagram level, and produces a set of good alternative diagrams in terms of cohesion, feature modularization and reduction of crosscutting concerns.

Privacy-enhancing personalized (PEP) systems address individual users' privacy preferences as well as privacy laws and regulations. Building such systems entails modeling two different domains: (a) privacy constraints as mandated by law, voluntary self-regulation, or users' individual privacy preferences, and modeled by legal professionals, and (b) software architectures as dictated by available software components and modeled by software architects. Both can evolve independently, e.g., as new laws go into effect or new components become available. In prior work, we proposed modeling PEP systems using a product line architecture (PLA). However, with an extensional PLA, these domain models became strongly entangled making it difficult to modify one without inadvertently affecting the other. This paper evaluates an approach towards modeling both domains within an intensional PLA. We find evidence that this results in a clearer separation between the two domain models, making each easier to evolve and maintain.

Software product line architecture is one of the most important artifacts defined at the early stage of a product line development process. Since the rest of the products are developed based on the initial product line architecture, it is of high importance to ensure the architecture stability by enabling the software’s evolution possibilities. Industrial evidence shows that companies spend more resources on maintaining and evolving their architecture and products than on the initial development of them. Hence, there is a need for flexible software architecture that stays stable as the requirements evolve. In this paper we propose a structural model, some architecture quality metrics, case-based reasoning methodology to predict the architectural stability and a feature model for business applications. The goal of the proposed architecture model is to develop a framework for business applications development and evaluating the stability of product line architectures in the face of changes in requirements.

The application of design patterns through mutation operators in search-based design may improve the quality of the architectures produced in the evolution process. However, we did not find, in the literature, works applying such patterns in the optimization of Product Line Architecture (PLA). Existing works offer manual approaches, which are not search-based, and only apply specific patterns in particular domains. Considering this fact, this paper introduces a meta-model and a mutation operator to allow the design patterns application in the search-based PLA design. The model represents suitable scopes, that is, set of architectural elements that are suitable to receive a pattern. The mutation operator is used with a multi-objective and evolutionary approach to obtain PLA alternatives. Quantitative and qualitative analysis of empirical results show an improvement in the quality of the obtained solutions.

Search-based algorithms have been successfully employed in Product Line Architecture (PLA) design in the seminal approach named Multi-Objective Approach for Product-Line Architecture Design (MOA4PLA). This approach generates a set of alternative PLA designs, which optimize different architectural properties. In addition to these properties, the alternative PLA designs should have as few architectural smells as possible. Architectural smells can negatively impact PLA variability, PLA extensibility, SPL maintainability, and other non-functional attributes. However, one of the main findings of a previous study is that the tool that automates the application of MOA4PLA adversely introduces architectural smells in the automatically generated solutions. In this work, we present OPLA-Tool-ASP, which is a tool that implements guidelines to detect and prevent the architectural smells Unused Interface, Unused Brick, Concern Overload, and Link Overload in the context of MOA4PLA. An empirical study was carried out to assess the effectiveness of OPLA-Tool-ASP in preventing the aforementioned smells in the resulting PLA designs. The obtained results pointed out that the proposed tool is effective in both preventing the smells and improving the architectural properties selected for optimization.

Intelligent tutoring systems (ITSs) are effective to provide instruction for students in several situations. Many works have been using gamification by adding game elements to learning contexts aiming to engage students and to drive desired learning behaviours. However, the design of gamified ITS should deal with a huge variability. Software product lines (SPLs) promise to offer rapid product development and more affordable development costs to build software from the same family. A key factor to successfully implement a product-line approach is to structure commonalities and variabilities into a product line architecture (PLA). In this paper, we propose a PLA for developing gamified ITSs that uses an ontology-driven feature modelling strategy. We illustrate how our architecture could be applied to instantiate a product on the basic math domain. We also discuss a set of implications of using it as well as how it could support the evolution/changing of gamified ITSs.

Software Product Lines (SPLs) is a reuse approach in which a family of products is generalized in a common architecture that can be adapted to different clients. The Product Line Architecture (PLA) is one of the most important artifacts of a SPL. PLA design requires great human effort as it involves several factors that are usually in conflict. To ease this task, PLA design can be formulated as an optimization problem with many factors, i.e, as a multi-objective optimization problem. In this context, the MOA4PLA approach was proposed to optimize PLA design using search algorithms and metrics specific to the context. This approach supported by OPLA-Tool has already been used in several works demonstrating its applicability. However, MOA4PLA does not take into account aspects that are subjective, such as the preferences of a particular Decision Maker (DM). To do so, this paper presents a proposal to incorporate the user preferences in the optimization process performed by MOA4PLA, through an interactive process in which the DM subjectively evaluates the solutions in processing time. Thus, the solutions generated can be better suited to the DM's needs or preferences. In order to allow the user interaction, modifications were made in MOA4PLA and implemented in the OPLA-Tool. Aiming at an initial validation of the proposal, an exploratory study was carried out, composed of two experiments: a qualitative and a quantitative. These experiments were realized with the participation of a software architect. Empirical results pointed out that the proposed interactive process enables the generation of PLAs that are in accordance with the architect's preferences. Another significant contribution are the lessons learned on how to improve the interactive process.

Product lines are an important system development paradigm in the automotive industry to amortize costs beyond a single product. The paradigm is well established in the mechanical and electrical engineering practise in automotive companies like Bosch. As software is covering more and more functionality in cars, software product lines are getting more attention. The architecture of a software-intensive system is a key asser in developing a software product line.The Architecture Trade-off Analysis Method (ATAM) developed by the SEI assesses the quality of software architecture early in the development process. ATAM is therefore a useful review technique to guarantee important quality attributes of every single product created with the product line architecture later on. This article reports about the experience Bosch made in using ATAM in two cases.Benefits in using ATAM are not only the review results itself but a better documented and better understood architecture. We experienced the most important benefit of ATAM is the rising stakeholder's awareness of architectural decisions, tradeoffs, and risks. It illuminates the software architecture better than any written documentation.Bosch employees are trained in the evaluation roles in order to trasition ATAM to Bosch.The reports conclude with some suggestions fo improving the ATAM itself and the training of ATAM roles.

Product line engineering is a recent approach to software development that has shown to enable organizations to achieve significant reductions in development and maintenance cost as well as time-to-market of increasingly complex software systems. Yet, the testing process has not kept up with these reductions and the relative cost for testing product lines is actually becoming higher than in traditional single system development. Also, testing often cannot keep pace with accelerated development in product line engineering due to technical and organizational issues. This paper advocates that testing of product lines can be made more efficient and effective by considering testability already during architectural design. It explores the relationship between testability and product line architecture and discusses the importance of high testability for reducing product line testing effort and achieving required coverage criteria. The paper also outlines a systematic approach that will support product line organizations in improving and evaluating testability of product lines at the architectural level.

Software product line development is a compromise between customer requirements, existing product line architectural constraints and commercial needs. Managing variability is the key to a successful product line development. Product line models of requirements and features can be constructed that contain variation points. New products can be driven by making requirement selections from a product line model of requirements but as the product line evolves selections are constrained by the design of the existing product line architecture and the cost of making these changes. We present a set of rules that map the selection constraint values of requirements to the selection constraint values of features which in turn map on to the selection constraint values of architectural assets. We illustrate the application of the rules using a worked example.