Otto Preiss

Stakeholder discovery and classification based on systems science principles

It is the goal of the research work presented to elaborate on improvements to software development methods so that quality attributes can be handled more systematically. By quality attributes, we mean the large group of typically systemic properties of a software system, such as availability, security, etc., but also reusability, maintainability and many more. We define quality attributes as stakeholder-centric conditions on the behavior or structure of a system. The importance of the notion of a stakeholder cannot surprise, but the lack of a general theory on how to define and identify the relevant set of stakeholders does. Drawing from systems theory, we claim that four basic, generic types of stakeholders are sufficient to be able to derive a specialized set of stakeholders for any considered system and domain of inquiry. It is only when we understand the generic concepts and principles behind quality properties of systems, that we can properly derive methods and build tools to cope with them.

On quality attribute based software engineering

Software components are an incarnation of architectural means to better cope with the variety of quality aspects of software systems. Unfortunately, architectural artifacts appear somewhat magically sometimes, and so do components. Components are not a major extension to OO in the programming language or functional modeling sense, but a basis to address many of the quality requirements, be they discernable or non-discernable at system runtime. CBSE, being the discipline of engineering with components, is a promising basis to more explicitly and systematically design with and for quality attributes. After defining the context and classifying quality attributes, we first illustrate the important relationship of quality attributes to use case realizations. Second, we argue for components as the fulcrum point for the realization of functional and extra-functional roles. Third we identify ongoing research directions that we consider conducive towards a software engineering process that supports the design for functional and extra-functional requirements.

Unified Information Models in Support of Location Transparency for Future Utility Applications

For future scenarios of power system operations, like the self healing grid vision, system and communications architecture are key. Applications will require unlimited access to a large variety of information and to a large amount of information. One important architectural feature is location transparency of its applications. The latter should be designed as components that can be run at any location where processing power is sufficient and the relevant data available. Key to location transparency is the agreement on a data model comprehensive enough to endure the foreseeable time to come. In order to safeguard current development investments, applications must already now be designed with such features in mind. This paper discusses how the data models of the IEC 61970 and IEC 61850, when integrated, can be used as a start towards such a uniform information model. However, inconsistencies and gaps justify a concerted effort in further extending such a basis.

MDA in enterprise architecture? The living system theory to the rescue

Integrating business and information technology (IT) is essential for enterprises to be competitive. The integration of business and IT forces project teams to analyze and design a hierarchy of systems such as: groups of companies collaborating in business systems, people and IT systems collaborating in business processes, software components collaborating in IT systems, and programming language objects collaborating in software components. The current tools, and in particular MDA, have not been tailored for the design of hierarchical systems. As a result the project teams have difficulties designing integrated business and IT systems. In this paper we present a technique for supporting the design of hierarchical systems. We were inspired by Millers general theory of living system, a theory that addresses the analysis of hierarchical living systems. We provide concrete guidelines on how to introduce levels in UML-based system modeling, in general, and in MDA in particular. This allows the project team to develop one, common, integrated enterprise (or project) model. This in turn enables the different specialists to reason about and design business and IT systems that are truly integrated. The overall benefit is an increase in the project success rate.

Towards a composition model problem based on IEC61850

In order to derive real-world model problems for the breath of research challenges in the area of predictable assembly of certifiable software components, this paper introduces the substation automation domain, as a representative for the wide spectrum of data acquisition and process control systems. Special emphasis in the paper is put on those aspects of the application area, which make its employment as a model problem provider attractive. First, the compositions are required to meet several well defined quality requirements besides functionality and must therefore allow for the prediction of these qualities before the final system is assembled and is undergoing a potentially costly factory and/or onsite acceptance test. Second, the upcoming IEC61850 standard defines agreed upon domain models and quality attribute requirements on system operations. In order to make this problem domain attractive to research in the area of compositional reasoning of component assemblies, this paper provides the needed step of relating the domain models and the domains quality attribute requirements to software components and assembly properties.

Data communications standards: a case for the UML

This paper argues for the UML as a means to rigorously specify data models and communications services of industrial data communications standards. Such standards contain among others the syntactic and semantic description of the application data to be exchanged among devices and systems of different vendors. As a consequence of the growing number and complexity of types of application data and their resulting specific communications services, it becomes almost impossible to maintain specification consistency and avoid ambiguities on the basis of informal, textual notations. By means of example, this paper discusses the recently accepted IEC 61850 standard “Communications networks and systems in substations” and its shortcomings due to the lack of formal notations. It is shown how many of these shortcomings can be overcome and additional benefits provided through the use of UML models – models that could ultimately find their way into the normative parts of such standards.