Non-functional properties (NFPs) represent an important facet of service descriptions, especially when a Service Oriented Architecture (SOA) approach is used. An effective SOA service development approach requires the identification, specification, implemen-tation, aggregation, management and monitoring of service-related NFPs. However, at this point in time, NFPs are either not handled at all or handled partially in proprietary ways. The goal of this thesis is to encourage their availability for use. In this thesis, the focus is on the NFPs relevant from the perspective of service consumers, in opposition to the perspective of service providers (or developers) and to multi-perspectives. In other words, the scope covers only the NFPs that need to be pub-lished to help service consumers determine whether a given service is an appropriate one for their needs (e.g., description of NFPs to be attached to the service along with the functionality description). This thesis provides the following contributions to the SOA knowledge base: definition of a domain-independent catalogue comprising 17 NFPs relevant to the descriptions of atomic services from the perspective of service consumers. These NFPs have been derived from a literature review and have been vali-dated via a two-step survey; formalization of NFP representation by defining data structures to enable quantifying and codifying them, together with a corresponding XML schema; definition, implementation and validation of algorithms to aggregate the NFPs of the composite service based on the NFPs of its underlying services, with a discussion of the NFP aggregation limitations; definition of a modeling approach for the NFP-aware selection of services, which involves aspect-oriented modeling with the User Requirements Nota-tion, in the context of SOA; integration of NFP descriptions into the Web Services Description Language (WSDL); and definition and use of the discriminator operator in service composition, to en-able the creation of fault-tolerant composite services. Overall, this work contributes to research by providing better insight on the nature, rele-vance, and composability of NFPs in a service engineering context. As for industrial im-pact, this work contributes a validated collection of NFPs with a concrete syntax and composition algorithms ready to be used for defining, selecting, and composing NFP-driven services and for evolving current SOA-related standards.

Non-functional properties (NFPs) represent an important facet of service descriptions, especially when a Service Oriented Architecture (SOA) approach is used. An effective SOA service development approach requires the identification, specification, implemen-tation, aggregation, management and monitoring of service-related NFPs. However, at this point in time, NFPs are either not handled at all or handled partially in proprietary ways. The goal of this thesis is to encourage their availability for use. In this thesis, the focus is on the NFPs relevant from the perspective of service consumers, in opposition to the perspective of service providers (or developers) and to multi-perspectives. In other words, the scope covers only the NFPs that need to be pub-lished to help service consumers determine whether a given service is an appropriate one for their needs (e.g., description of NFPs to be attached to the service along with the functionality description). This thesis provides the following contributions to the SOA knowledge base: definition of a domain-independent catalogue comprising 17 NFPs relevant to the descriptions of atomic services from the perspective of service consumers. These NFPs have been derived from a literature review and have been vali-dated via a two-step survey; formalization of NFP representation by defining data structures to enable quantifying and codifying them, together with a corresponding XML schema; definition, implementation and validation of algorithms to aggregate the NFPs of the composite service based on the NFPs of its underlying services, with a discussion of the NFP aggregation limitations; definition of a modeling approach for the NFP-aware selection of services, which involves aspect-oriented modeling with the User Requirements Nota-tion, in the context of SOA; integration of NFP descriptions into the Web Services Description Language (WSDL); and definition and use of the discriminator operator in service composition, to en-able the creation of fault-tolerant composite services. Overall, this work contributes to research by providing better insight on the nature, rele-vance, and composability of NFPs in a service engineering context. As for industrial im-pact, this work contributes a validated collection of NFPs with a concrete syntax and composition algorithms ready to be used for defining, selecting, and composing NFP-driven services and for evolving current SOA-related standards.

Non-functional properties (NFPs) represent an important facet of service descriptions, especially when a Service Oriented Architecture (SOA) approach is used. An effective SOA service development approach requires the identification, specification, implemen-tation, aggregation, management and monitoring of service-related NFPs. However, at this point in time, NFPs are either not handled at all or handled partially in proprietary ways. The goal of this thesis is to encourage their availability for use. In this thesis, the focus is on the NFPs relevant from the perspective of service consumers, in opposition to the perspective of service providers (or developers) and to multi-perspectives. In other words, the scope covers only the NFPs that need to be pub-lished to help service consumers determine whether a given service is an appropriate one for their needs (e.g., description of NFPs to be attached to the service along with the functionality description). This thesis provides the following contributions to the SOA knowledge base: definition of a domain-independent catalogue comprising 17 NFPs relevant to the descriptions of atomic services from the perspective of service consumers. These NFPs have been derived from a literature review and have been vali-dated via a two-step survey; formalization of NFP representation by defining data structures to enable quantifying and codifying them, together with a corresponding XML schema; definition, implementation and validation of algorithms to aggregate the NFPs of the composite service based on the NFPs of its underlying services, with a discussion of the NFP aggregation limitations; definition of a modeling approach for the NFP-aware selection of services, which involves aspect-oriented modeling with the User Requirements Nota-tion, in the context of SOA; integration of NFP descriptions into the Web Services Description Language (WSDL); and definition and use of the discriminator operator in service composition, to en-able the creation of fault-tolerant composite services. Overall, this work contributes to research by providing better insight on the nature, rele-vance, and composability of NFPs in a service engineering context. As for industrial im-pact, this work contributes a validated collection of NFPs with a concrete syntax and composition algorithms ready to be used for defining, selecting, and composing NFP-driven services and for evolving current SOA-related standards.

Context: Organizations strive to improve their business processes, and adaptive business processes have recently attracted much attention in that context. However, much research in that area has a narrow focus and does not consider a comprehensive view of the organization and its goals. In addition, Business Intelligence-based monitoring methods are useful for business process improvement but they often present information in a format that is not entirely suited for decision making. Objectives: The main objectives of this thesis are to provide: • A framework to model goals, processes, performance, situations, and improvement patterns using one modeling notation, in an iterative and incremental manner; • A method for the modeling and analysis of cause-effect relationships between indicators used to measure goal satisfaction; and • A technique allowing the detection of undesirable, sub-optimal conditions and the application of improvement patterns to the context Method: We develop an iterative framework based on the User Requirements Notation (URN) for modeling, monitoring and improving business organizations and their business processes. In addition, we introduce a formula-based evaluation algorithm allowing better analysis of the relationships between the business performance model elements (namely indicators). Furthermore, we use a profiled version of the Aspect-oriented URN (AoURN) with extensions (Business Process Pattern profile), for detecting undesirable conditions and for business process adaptation. We validate the novelty and feasibility of our approach by performing a systematic literature review, by assessing it against Zellner’ mandatory elements of a method, by developing tool support, by performing a pilot experiment and by using real-life examples from different sectors (healthcare and retail). Results: The two examples show that through the framework’s iterative approach, organizations at different levels of maturity in their business improvement journey can benefit from the framework. Furthermore, our systematic literature review shows that although there are existing works that enable our vision, most of them have a narrow focus and do not cover the three organization views that are of interest in this research. AoURN allows analysts to find repeated patterns in a context and bundle goal, performance and process models as a self-contained unit. AoURN hence enables the modeling of complex circumstances together with analysis techniques for what-if analysis and process adaptation, all using a unified and integrated modeling language. Finally, the pilot experiment suggests that, with some level of documentation and training, users who are already familiar with URN can use the profiled AoURN provided in this thesis as well as the discussed improvement patterns.

Context: Organizations strive to improve their business processes, and adaptive business processes have recently attracted much attention in that context. However, much research in that area has a narrow focus and does not consider a comprehensive view of the organization and its goals. In addition, Business Intelligence-based monitoring methods are useful for business process improvement but they often present information in a format that is not entirely suited for decision making. Objectives: The main objectives of this thesis are to provide: • A framework to model goals, processes, performance, situations, and improvement patterns using one modeling notation, in an iterative and incremental manner; • A method for the modeling and analysis of cause-effect relationships between indicators used to measure goal satisfaction; and • A technique allowing the detection of undesirable, sub-optimal conditions and the application of improvement patterns to the context Method: We develop an iterative framework based on the User Requirements Notation (URN) for modeling, monitoring and improving business organizations and their business processes. In addition, we introduce a formula-based evaluation algorithm allowing better analysis of the relationships between the business performance model elements (namely indicators). Furthermore, we use a profiled version of the Aspect-oriented URN (AoURN) with extensions (Business Process Pattern profile), for detecting undesirable conditions and for business process adaptation. We validate the novelty and feasibility of our approach by performing a systematic literature review, by assessing it against Zellner’ mandatory elements of a method, by developing tool support, by performing a pilot experiment and by using real-life examples from different sectors (healthcare and retail). Results: The two examples show that through the framework’s iterative approach, organizations at different levels of maturity in their business improvement journey can benefit from the framework. Furthermore, our systematic literature review shows that although there are existing works that enable our vision, most of them have a narrow focus and do not cover the three organization views that are of interest in this research. AoURN allows analysts to find repeated patterns in a context and bundle goal, performance and process models as a self-contained unit. AoURN hence enables the modeling of complex circumstances together with analysis techniques for what-if analysis and process adaptation, all using a unified and integrated modeling language. Finally, the pilot experiment suggests that, with some level of documentation and training, users who are already familiar with URN can use the profiled AoURN provided in this thesis as well as the discussed improvement patterns.