Provenance awareness adds a new dimension to the engineering of service-based systems, enabling them to increase their accountability through answering questions about the provenance of any data produced. Provenance awareness can be achieved by recording provenance data during system execution. In our previous work we have proposed an overall research agenda towards a design and analysis framework for provenance awareness of composite services. A fundamental element of this framework is the provenance model, Service Provontology, capturing the structure of the provenance data collected which allows designers to query and reason over provenance data instances of composite services. With this paper we contribute an industrial case study exploring real-world provenance data from a service-based system (ORBI). In our study ServiceProv becomes the tool for enabling representation and reasoning over ORB provenance data instances in order to answer specific provenance questions formalized as SPARQL expressions.

With the fast adoption of Services Computing, even more driven by the emergence of the Cloud, the need to ensure accountability for quality of service (QoS) for service-based systems/services has reached a critical level. This need has triggered numerous researches in the fields of trust, reputation and provenance. Most of the researches on trust and reputation have focused on their evaluation or computation. In case of provenance they have tried to track down how the service has processed and produced data during its execution. If some of them have investigated credibility models and mechanisms, only few have looked into the way reputation information is produced. In this paper we propose an innovative design for the evaluation of feedback authenticity and credibility by considering the feedback's provenance. This innovative consideration brings up a new level of security and trust in Services Computing, by fighting against malicious feedback and reducing the impact of irrelevant one.

Cloud Computing provides simplicity to its consumers by saving them the efforts to deal with their own infrastructure, environments or software. This simplicity relies on the shifting of problems from the client to the provider, introducing new paradigms (virtualization, scalability, flexibility, pay-per-use, etc.). This simplicity comes with a price for the consumer that may accurately, or not, reflect the costs of the provider. In this paper we propose to identify the different points, in the Cloud Computing architecture, where the costs are generated, how their reduction/optimization are considered, and finally we point-out which of these key points need to be further investigated, according to their foreseeable efficiency.

Even though Cloud Computing becomes the trend, the reputation in the Cloud is still a critical issue for its adoption by cloud customers. In the recent years researches toward improving and refining the reputation have been motivated by the emergence and rapid growth of Service and Cloud Computing. These researches have produced numerous models and methods, taking into account different aspects of the reputation. Such methods often rely on statistical tools to project and estimate the reputation of services and of some of their characteristic. However, they usually fail to completely describe the inner complexity of reputation's context. In this paper we propose a definition of the Reputation relying on the use of a holistic ontology of the Cloud, ensuring both an optimal coverage of the domain and a deep semantic rooting. This definition is based on the identification of the key aspects requiring the support of the ontology for their evaluation.

Establishing and managing reputation has become a critical issue for Cloud Computing. In this paper we investigate this issue and propose our conception for the management of Reputation in Cloud Computing. This conception is based on the extensive use of SLA through the whole life cycle of the reputation. With this objective in mind we focus on defining and describing the close relations that exist between the establishment of reputation and the management of SLA. In this perspective we identify key aspects and mechanisms to develop a truly efficient Reputation management based on the use of SLAs.

With the evolution of service computing ecosystems and the promises of the Cloud, there is an increasing need for service evaluation and monitoring across a wide range of QoS dimensions. Although in the literature service monitoring is tightly coupled with Service Level Agreement establishment, existing approaches do not provide a holistic model for Cloud service evaluation. In this paper, we present an SLA ontology, encompassing the whole service lifecycle, in order to drive and improve the evaluation of services. We propose a common QoS ontology to express specifications, requirements and feedback. Based on this, we introduce a symmetric representation of service monitoring information and human feedback, allowing a unified management of the service evaluation. Finally, the SLA ontology supports the profiling of users, enabling a refined reputation calculation through the dynamic consideration of customer segments.

In this paper, we describe the design, the development and the use of devices collaboration rules for the PCSCW (Pervasive Computing Supported Collaborative Work) Model. These rules rely on the precise description of roles, tasks, actions, resources required by these actions and constraints associated to these resources to select the proper way to make devices cooperate with the final objective to facilitate the collaboration of humans. We suggest that by defining constraints on resources as triplets composed of a parameter, a value and an associated criticality it allows us to quantify, estimate, compare and then choose between several candidate rules. The finality given by these rules is a simple but efficient way to make devices choose automatically the most appropriate way to cooperate.

In this paper we define an ontological model to accurately represent the context in Pervasive Computer Supported Collaborative Work. A major issue in this domain is the mass of information required to correctly depict a situation. As we need to represent users and devices according to multiple aspects (physical, computational, social ...) the amount of information can quickly become unmanageable. Besides, as a PCSCW context model has to be usable on limited resources devices such as cell phones, GPS, ADSL Modems we needed a more efficient way to represent information. In this perspective the model we propose offers the possibility to represent a situation with more or less precision; that is to say with more or less abstraction. The final goal of this work is then to provide a model able to reason with a precise or fuzzy description of a situation.

In this paper we present an hybrid architecture for a computer supported collaborative work in pervasive computing environment simulator. This architecture supports pure simulation mode where all agents are simulated, real deployment where it seamlessly integrates pervasive computing to the computer supported collaborative work and finally an hybrid mode where real devices can interact with simulated ones. This specific architecture provides an original framework to develop and evaluate complex behaviours in pervasive computing environments by allowing scenario to be played in a semi-simulated way and avoid some of the constraints and difficulties of real conditions testing. In addition to theses obvious advantages it also provides a reliable base to capture and analyse real devices behaviour in pervasive computing environments.

computer supported Collaborative Work (CSCW) has been a hot point in researches for some decades. Recent advances in software and hardware technologies have allowed the use of more and more advanced applications and services. It also brought computing capabilities to mobile devices such as smartphones and laptops. This has led to an extensive use of computers to collaborate in some unexpected manners. To support this collaboration, numerous models have been designed and tried. Among these models, some particularly caught our attention: models based on the description of tasks, role-based models and collaboration’s context models. Simultaneously major advances were made in the young pervasive computing domain. We propose a role-based model, which aims at modelling the cooperation between devices to support pervasive collaborative work. Leading to the emergence of the Pervasive Computing Supported Collaborative Work (PCSCW).

In this paper we present a pervasive service to enhance the collaboration between members of a community. This service allows scattered users using various communication devices (PDA, Smartphone, computer...) to communicate and work efficiently. To provide an ergonomic way for people to collaborate, we propose a system of pervasive brainstorming based on the use of SOMCQ.

With the rapidly growing development of Computer Supported Collaborative Work technologies, the evaluation of these services becomes an essential aspect. This evaluation mixes technical, business, social, perceptive, ergonomic aspects which can't be considered independently. In this paper we propose a new taxonomy of CSCW evaluation methods based on previous works and on our analysis of current evaluation methods. With this new taxonomy and the description of CSCW development process and life-cycle we are able to propose evaluation strategies that can be adapted and tailored for most of systems.

In this paper we present a pervasive service to enhance collaboration between scattered users of a community. We base our system on the use of SOMCQ sent into simple mails and published on a collaborative website. This allows users using non-ergonomic devices to efficiently collaborate.