In a service-oriented system (SoS) service requests define tasks to execute and quality of service (QoS) criteria to optimize. A service request is submitted to an automated service composer in the SoS, which allocates tasks to those service that, together, can "best" satisfy the given QoS criteria. When the composer cannot optimize simultaneously the given QoS criteria, users need to specify priorities over the said criteria. Accounting for users' QoS priorities is therefore necessary during service selection. Once specified by the requester, quality properties will be used by the composer to lead autonomic optimization of the service selection process. We outline and test a selection approach that accommodates priorities and that is based on available Multi Criteria Decision Making techniques.

Knowledge-Base Recommendation (or Recommender) Systems (KBRS) provide the user with advice about a decision to make or an action to take. KBRS rely on knowledge provided by human experts, encoded in the system and applied to input data, in order to generate recommendations. This survey overviews the main ideas characterizing a KBRS. Using a classification framework, the survey overviews KBRS components, user problems for which recommendations are given, knowledge content of the system, and the degree of automation in producing recommendations.

In a service-oriented system, a quality (or Quality of Service) model is used (i) by service requesters to specify the expected quality levels of service delivery; (ii) by service providers to advertise quality levels that their services achieve; and (iii) by service composers when selecting among alternative services those that are to participate in a service composition. Expressive quality models are needed to let requesters specify quality expectations, providers advertise service qualities, and composers finely compare alternative services. Having observed many similarities between various quality models proposed in the literature, we review these and integrate them into a single quality model, called QVDP. We highlight the need for integration of priority and dependency information within any quality model for services and propose precise submodels for doing so. Our intention is for the proposed model to serve as a reference point for further developments in quality models for service-oriented systems. To this aim, we extend the part of the UML metamodel specialized for Quality of Service with QVDP concepts unavailable in UML.

Abstract. In a Service-Oriented System (SOS), service requesters specify tasks that need to be executed and the quality levels to meet, whereas service providers advertise their services' capabilities and the quality levels they can reach. Service selectors then match to the relevant tasks, the candidate services that can perform these tasks to the most desirable quality levels. One of the key problems in QoS-aware service selection lies in managing tradeoffs among QoS expectations at runtime, that is, situations in which service requesters specify quality levels that cannot be simultaneously met. We propose a service selection approach that can deal with tradeoffs. The approach consists of: (i) rich QoS models to be used by service requesters when expressing QoS expectations and service providers when describing services' QoS, and for representing preference and priority relationships between QoS dimensions; and (ii) a multi-criteria decision making technique that uses the models for service selection.

Service Level Agreements (SLAs) are used in Service-Oriented Computing to define the obligations of the parties involved in a transaction. SLAs define the service users' Quality of Service (QoS) requirements that the service provider should satisfy. Requirements defined once may not be satisfiable when the context of the web services changes (e.g., when requirements or resource availability changes). Changes in the context can make SLAs obsolete, making SLA revision necessary. We propose a method to autonomously monitor the services' context, and adapt SLAs to avoid obsolescence thereof.