Nicolas Ferry

WComp, a Middleware for Ubiquitous Computing

Ubiquitous computing relies on computers present everywhere, at any times and in any things. Indeed with recent years advance in mobile communication technologies and the miniaturization of computer hardware, processing units are becoming invisible and a part of the environment. Middlewares for ubiquitous computing have to manage three main features specific to their environment: devices’ mobility, devices’ heterogeneity and environment’s dynamicity. The devices’ mobility, due to motion of users and their associated devices, forbids to assume that entities are known and will always be available. The second concept, entity’s heterogeneity, outlines the diversity between devices’ capabilities and functionalities provided by new smart objects. Finally, the environment high dynamicity illustrates the ubiquitous world entropy with the appearance and disappearance of devices. Devices used to create applications are thus unknown before discovering them. Then, ubiquitous computing must deal with such a dynamic software environment (called software infrastructure afterwards). As a result, future ubiquitous computing architectures must take into account those three constraints to solve ubiquitous computing challenges. Our model of middlewareWComp is based on three parts: a software infrastructure, a service composition architecture, and a compositional adaptation mechanism. To manage the dynamicity and heterogeneity of entities in the software infrastructure, we highlight the use of Web Service Oriented Architecture for Device (WSOAD). This will be discussed in section 2. Ubiquitous applications are then based on a set of Web services for devices that must interact with each other. Consumers can not edit these services. Therefore, in order to add new functionalities to the system, an application has to be a composition of services for devices. Such an application, and thus such a composition, must be modifiable at runtime. The second part of the WComp middleware enables us to make such applications by dynamically composing services from the software infrastructure. To allow reusability of newly created functionalities, and for scalability purposes, such composition can be encapsulated as a composite service. This part of the system will be presented in Section 3. Moreover, the infrastructure of ubiquitous computing applications evolves dynamically led by appearances and disappearances of objects or devices. The variation of this infrastructure is dynamic due to arbitrary node mobility, failures or energy constraints. The service composition must be as relevant as possible according to the underlying software infrastructure. Managing these

Toward a Behavioral Decomposition for Context-Awareness and Continuity of Services

Many adaptative context-aware middleware exist and most of them rely on so-called vertical architectures that offer a functional decomposition for contextawareness. This architecture has a weak point: it does not allow the system handling both dynamics of the changing environment and applications. To avoid this, we propose an approach for context-awareness based on a behavioral decomposition, and because each behavior must complete all functionalities necessary for contextawareness, we introduce an hybrid decomposition. It consists in a functional decomposition into a behavioral decomposition. This approach derives benefits from both decomposition, first allowing to handle environment and application’s dynamics, second introducing reusability and modularity into behaviors.

Context adaptative systems based on horizontal architecture for ubiquitous computing

Many adaptative context-aware middleware exist and mostly rely on so-called vertical architectures that offer a functional decomposition for context-awareness. This architecture has a weak point: it leads to data centralization. Our mechanism for adaptation: the Aspects of Assemblies is based on a horizontal architecture. This type of architecture separate the system into behavior and is based on a decentralized approach. However, after having shown some limitations of AAs in the field of context-awareness we will introduce a way to improve them using a multi-cycle weaving approach. Then, using this approach we will be able to build context-adaptative systems that interact directly with their environment. Finally we will evaluate our approach in term of reactivity.

Cascaded Aspects of Assembly for ubiquitous computing

Ubiquitous systems are characterized by using devices and objects of everyday life. The software infrastructure of such systems appears dynamically populated by functionalities of those devices. Because environments nature is highly variable, even the corresponding software infrastructure, ubiquitous systems have to handle those variations. It implies that a designer cannot predict a priori the environment in which its application will be executed. So, the approaches used should not only address specific cases of context. AOP provides a mean to encapsulate adaptations into aspects. Such an encapsulation improves the reusability of adaptation mechanisms. In this context our previous works called Aspect of Assembly (AA) was based on AOP to propose mechanisms for self-adaptation of ubiquitous applications as an assembly of components. But they present few limitations in terms of reusability and to manage the dynamic variability of the environment. The paper presents Cascaded Aspects of Assembly as an extended model of AA to address these limitations. They consist in chaining AAs weaving as cascades.

Low response time context awareness through extensible parameter adaptation with ORCA

Ubiquitous computing applications or widespread robots interactions execute in unforeseen environments and need to adapt to changeful available services, user needs, and variations of the environment. Context-awareness ability addresses such a need, enabling, through adaptation rules, applications to react to the perceived dynamic variations. Responses to adaptation have to be quick enough to maximize the time during which the application is coherent with its environment. Adaptation rules, associating variations of the environment to application reactions, are usually established at design time. However, in unforeseen and partially anticipated environments, we claim that adaptation rules have to be dynamically extensible to match previously unexpected variations. Our approach enables rule composition and ensures a deterministic result. We propose to use parameter adaptation to quickly respond to environmental variations and dynamic compositional adaptation to provide extensibility to the parameter adaptation. To foster even lower response times, we internalize context-awareness processing and decision into the application.

Multi-dynamics adaptations using cascaded aspect of assembly

Middleware for ubiquitous computing are having a mediator role in ubiquitous systems. They have to be able to handle both dynamics: the dynamic of the environment and of the application. But, at least, they must consider the dynamic of the underlying software infrastructure. To handle such a dynamic, we present in this paper a mechanism for adaptation called cascaded aspects of assembly (AA). The functional decomposition proposed by cascaded AAs allows to dynamically compose adaptations to manage the dynamic variability of the application and of its infrastructure. Because cascaded AA can be selected/unselected or edited at runtime, some others processes can manage them in various autonomous and decentralized processes to handle the other dynamics.