QoS-aware Semantic Matchmaking Algorithms For Web Service Discovery And Selection

In today proliferation of available web services, web services are advertised in inter and intra-organizational registries using common service description languages. They have to be discovered based on users? functional and non-functional requirements called quality of services(QoS). However, the most common registries and service description languages rely on current web service(WS) standard such as web service description language(WSDL), simple object access protocol(SOAP), universal description discovery and integration(UDDI) and so on. They describe the information for service interfaces and operations according to category or syntactic based approaches. This situation creates two problems in web service environment organized with various kinds of structures, languages and descriptions.The first problem is that some service approaches solely focus on functional service information in service management. As a result, the service matchmakers fail to search the best qualified services for the users because there are many functionally equaled services on the web and in turn difficult to discriminate the most suitable services according to users? requirements. The second problem is that syntactic based approaches produce imprecise and inaccurate results because of deficient evaluation factors applied in service methods.The first problem can be addressed by consideration of all possible non-functional QoS properties of WSs as a decisive factor among enormous similar attributes in functional services. The purpose of QoS addition to WS description is not only to satisfy users? expectation about service qualities but also to enhance service matchmaking and discovery process in terms of service quality issues. Presumably, fewer ranked results will be produced that could completely cover all users? expected quality conditions.The second problem can be solved by exploiting semantic technology called ontology. Ontologies are used to understand semantic meanings and concepts of service descriptions. It can lead to semantic based service matchmaking, discovery and selection processes. According to the proof of recent studies, the semantic based approaches could reveal more precise and accurate results than syntactic approaches could.In this thesis book, therefore, the first part is dedicated to QoS ontology design, development and annotating the semantic information to service information. It firstly unveils the features and necessities in describing and understanding of QoS-based service documents. Based on these necessities, afterwards, a new QoS ontology called SQoS-Onto has been carefully designed. Moreover, useful semantic data, which can link between service elements and ontology concepts of SQoS-Onto, are annotated to service description so that semantic service descriptions can suavely understood in an automatic manner.The other part of our thesis deals with how to match QoS metrics and attributes, how to discover user required services with their expected quality levels, and how to allow the users to involve service selection processes with their preferences, experiences and feelings. In these service processes, essential features, which are aggressively demanding in web service system such as self-adaptable, flexible and robustness, are extensively applied in service matchmaking, discovery and selection system. In addition, another enhancement of this part is about ontological semantic technique exploitation in all service approaches proposed by this thesis.The prominent fact of bringing this ontology technology in web service system is that we not only traditionally use semantic concept of service elements, we also consider attribute and hierarchical structure of ontological semantic relationships of QoS service elements. To the best of our knowledge, we are the first of introducing this contribution to this QoS-based service area. We explore and exploit these hierarchical structures and attribute relationships of QoS service information in service matchmaking algorithms.The QoS-aware semantic matchmaking algorithms are proposed for each different service approach such as service ranking, selection and discovery. The impacts of these approaches are the following. First of all, a quite complete design of QoS ontology is introduced to cover every aspect of QoS service information. Second of all, a novel service level agreement(SLA) mapping and provider selection approach is proposed with annotated ontological semantic information to Web Service Agreement(WS-Agreement). It then applies semantic information in SLA mapping to find suitable providers for the users. Thereafter, the advanced service matchmaking algorithms are proposed for web service discovery and selection approaches. In discovering the services, it explores self-adaptable rules to be robust the discovery system in order to avoid erroneous user inputs and no feasible solution occurrences. Another benefit of service selection is that it allows the users to comfortably give service assessments about their experiences and feelings. Last but not least, with QoS ontology and semantic annotation, the advanced service discovery and selection processes become more reflective to users satisfaction and also realistic to implement the real time web service system.