Diana Elena Comes

Different Approaches to Semantic Web Service Composition

Semantic web service composition is about finding services from a repository that are able to accomplish a specified task if executed. The task is defined in a form of a composition request which contains a set of available input parameters and a set of wanted output parameters. Instead of the parameter values, concepts from an ontology describing their semantics are passed to the composition engine. The parameters of the services in the repository the composer works on are semantically annotated in the same way as the parameters in the request. The composer then finds a sequence of services, called a composition. If the input parameters given in the request are provided, the services of this sequence can subsequently be executed and will finally produce the wanted output parameters. In this paper, three different approaches to semantic web service composition are formally defined and compared with each other: an uninformed search in form of an IDDFS algorithm, a greedy informed search based on heuristic functions, and a multi- objective genetic algorithm.

Heuristic Approaches for QoS-Based Service Selection

In a Service Oriented Architecture (SOA) business processes are commonly implemented as orchestrations of web services, using the Web Services Business Process Execution Language (WS-BPEL). Business processes not only have to provide the required functionality, they also need to comply with certain Quality-of-Service (QoS) constraints which are part of a service-level agreement between the service provider and the client. Different service providers may offer services with the same functionality but different QoS properties, and clients can select from a large number of service offerings. However, choosing an optimal collection of services for the composition is known to be an NP-hard problem.

Designing socio-technical applications for ubiquitous computing: results from a multidisciplinary case study

A major challenge for ubiquitous system design is creating applications that are legal-compatible and accepted by their intended users. Todays European data protection principles contradict the ideas of ubiquitous computing. Additionally, users have to deal with unconventional interaction concepts leading to a low amount of trust and acceptance in such systems. Current development approaches do not sufficiently cover these concerns, as they do not systematically incorporate expertise from the relevant disciplines. We present a novel development approach for ubiquitous systems that explicitly addresses these concerns. Our primary task was to manage the increased number of stakeholders and dependencies, respectively conflicts between requirements of the particular disciplines. The approach incorporates predefined artifacts and a defined workflow with responsibilities, as well as suggesting how to develop mutual understanding. We apply this multidisciplinary approach to develop the ubiquitous application Meet-U.

A Flexible Approach for Business Processes Monitoring

Business processes and their implementation as Web Service Compositions are not only dependent on Web Services and partners all over the Internet, but also on their failsafe execution. Service providers have to obligate their services to perform according to negotiated Quality of Service (QoS) parameters. For example, response time and throughput are important parameters to achieve fast and efficient services. Overloaded or failing services may compromise the reliability and execution of whole enterprise processes. In this paper we introduce a flexible monitoring approach for the measurement of QoS in BPEL (Business Process Execution Language) processes. We propose a generic algorithm for QoS aggregation in BPEL processes. The novel generic aggregation algorithm applies customized aggregation functions for QoS dimensions. Furthermore, we present a BPEL monitoring system which supports ad-hoc sensor deployment and efficient runtime and offline data aggregation not only for whole process descriptions but also sections inside service processes.

Context-Aware Prediction of QoS and QoE Properties for Web Services

Web Services are commonly used for integrating applications between partners over the Internet. Since services with the same functionality are advertised with different Quality of Service (QoS) levels and are assessed with different Quality of Experience (QoE), choosing the right service may be quite challenging. It is essential for a user to predict QoS and QoE values as accurately as possible in order to find a suitable service. Usually collaborative filtering is applied using similar users and services for predictive purposes. We hypothesize a correlation between context data and QoS and QoE dimensions which can be additionally incorporated to improve predictive accuracy and scalability. In this paper we present the two algorithms PredReg and PredNet in order to predict QoS and QoE values for Web Services. The PredReg algorithm is based on multiple linear regression. The PredNet algorithm uses additionally a neural network for prediction. Both algorithms include context data of users and services generating personalized predictions for the requesting user. In addition, PredNet is able to process categorical variables so that user profiles can also be considered for predictions. We evaluated PredReg and PredNet and compared them with the state-of-the-art approach WSRec [1] which is a memory-based collaborative filtering approach. Our experiments demonstrated that PredReg and PredNet provide a higher predictive accuracy and a significantly improved scalability. Therefore, we recommend the application of PredReg and PredNet for future personalized predictions.

Management of Business Processes with the BPRules Language in Service Oriented Computing

Quality of Service (QoS) concerns are an important topic for the realization of business processes. While BPEL is considered the de facto standard for web service compositions, QoS requirements are not part of its specification. We present the BPRules (Business Process Rules) language for the management of business processes with respect to QoS concerns. BPRules is a rule-based, declarative language which brings novel benefits in the management of business processes, like QoS dependability for sub-orchestrations and corrective actions tailored to the specific needs of the clients. We present the main constructs of the BPRules language and how they support the flexible adaptation of the business process during runtime. Decision making is done according to the behavior of several process executions. An illustrative scenario shows how BPRules is applied to a business process.

Automated Integration of Web Services in BPEL4WS Processes

In order to fully exploit the potential of dynamic service-oriented architectures based on Web Services we provide a novel self-integration service infrastructure that supports automatic service discovery and reconfiguration for BPEL4WS processes. Our service discovery approach takes into account possible runtime transformations such that a service in a BPEL4WS process can be replaced by a semantically similar service even if service interfaces and message structures do not match. We present the main building blocks of our solution, i.e. a standard-conformant WSDL schema extension called Mediation Contract Extension (MECE), a corresponding semantic discovery algorithm, and a runtime mediation system that generates appropriate XSLT transformations on the fly. Our solution dynamically instantiates mediators to bind services to service processes specified with BPEL4WS.

QoS-based Self-Management for Business Processes

Business processes are commonly implemented as compositions of Web Services, using the Business Process Execution Language (BPEL) as an orchestration specification. Business processes do not only require an appropriate setup but also need to be monitored throughout their runtime, especially when Quality-of-service (QoS) constraints have to be met. Monitoring results may be used for the automated reconfiguration and optimization of business processes. We show how we achieve self-management based on QoS constraints within our system. The BPRules Language that we set up can be used to improve the QoS behavior of business processes by triggering appropriate management actions on the process. Also we propose a service selection strategy for the dynamic selection and replacement of services within business processes.