Concepts

Universal Service Terminal

The on-going trend to miniaturized powerful mobile devices has brought the vision of ubiquitous computing and communication much closer to reality. Context-aware information and services will increasingly become accessible from anywhere, at any time. The Universal Service Terminal Enhancement (UST+) project researches selected key aspects of this vision. One such aspect is seamless collaboration of distributed services in different kinds of environments – in the user’s digital home or on-the-road; in the personal domain or in Web 2.0 communities, and in networks with or without supportive infrastructure. The result is a collection of concepts and software tools designed to simplify and accelerate the development of innovative distributed applications. To be able to transfer the developed concepts to “real life” devices a middleware solution was developed that features particular aspects and can be installed on modern mobile phones.



Here, the Universal Service Terminal is a combination of functionality provided by multiple services on one or more devices to fulfill a particular user request. Those services are managed and composed by means of a common middleware layer providing general functions such as service discovery or recovery.

However, the UST+ project does just not aim at the definition of a new all-embracing middleware approach, but is specially geared to convergence of different well-established standards and innovative concepts, such as support for context-awareness. Boundaries between different service environments and networks are reduced, paving the way for a pervasive and user-centric provision of services.

UST+ devices are devices that have the UST+ middleware installed and can therefore take part in Universal Service Terminals. Particular features and concepts that are especially supported in UST+ are:

• A blackboard approach for the seamless exchange of semantically annotated data between UST+ devices,
• Loosely coupled and distributed state management approach to enable the collaboration of multiple device services,
• Consideration of context information influencing the semantically enhanced service selection and service composition processes,
• Framework elements for the support of different service environments and networks: mobile ad hoc and infrastructure based, on the way and at home.

Service Collaboration

Social community applications on the Web compellingly illustrate the appeal and strengths of collaborative behavior. Inspired by the guiding principles of what is often referred to as Web 2.0, a solution suite for mobile devices was realized in UST+ to support device services in terms of four distinct principles:

• Services are enabled to collaborate on a common user task
• Complementary service functionality can be discovered and combined automatically to offer sophisticated context-aware services
• Active and open information exchange among UST devices enhances functionality such as service discovery or recovery
• A loosely coupled service interaction model enables ad hoc connections of devices and seamless exchange of data

The service collaboration approach pursued in UST+ does just not aim at the definition of a new all-embracing middleware approach, but is specially geared to convergence of different well-established standards and innovative concepts, such as support for context-awareness. Boundaries between different service environments and networks are reduced, paving the way for a pervasive and user-centric provision of services.

The resulting solution suite supports semantic description languages which are utilized to represent service functionality, user tasks or profiles, and context information. In this way, UST devices are able to automatically execute and compose available services according to a predefined user goal. Context changes can affect service behavior and may trigger appropriate adaptation or recovery mechanisms.

Service composition

Service Composition became widely known with the emergence of Service Oriented Architecture (SOA), an architectural style based on the composition of loosely coupled services. With SOA a new request does not necessarily need to be met with a newly developed application, but may be answered by a composition of services already on hand.

Within the scope of the UST+ project, the principle of service composition is regarded as key enabler to support the collaboration of multiple mobile of infrastructure based devices. Here, all devices hosting at least one service of the service composition that processes the given request are considered as a UST.

Thus, the UST+ project expands the view from single interacting applications to the process of composing services from different devices in a context-aware manner for optimal exploitation of the functionality of available services and to offer a completely new functionality. In this context semantics play a vital role in enabling a flexible and general approach in terms of describing crucial information such as service interfaces, user requests, and user preferences. The specifics of mobile environments strongly influence the outcome of the service composition process and call for dynamic adaptation, e.g., in terms of service recovery going beyond traditional systems with central controlling instances.

Trust Representation and Evaluation

Automated service discovery mechanisms commonly selecting services based on their functionality only. Since it is not under any control who participate within distributed service environments, it may occur that services do not perform as expected, e.g., by providing inaccurate, insufficient, false, or no service results. Indeed, providing inaccurate or insufficient results may refer to the subjective perception of a user according to its personal preferences. Here, trust may be an appropriate way to handle the mentioned issues.

Within the UST+ project, concepts of a framework that deals with the representation and evaluation of trust in mobile ad hoc networks were developed. The approach is based on a dynamic service behavior observation model that is applied to find and propose the most suitable and reliable service or to restrict the access to local services. Moreover, misbehaving peers can be identified and excluded from the network’s knowledge about other peers.

Two main issues are especially interesting. First, the service rating model to get knowledge about the behavior of nodes is needed. Second, the dissemination of information about other nodes is required to achieve a broader decision basis, e.g., to get a priori knowledge about unknown nodes.

The rating of entities and its provided services depends on the promised effects, which are announced by the service provider, e.g., through common service description frameworks, and the occurred or monitored effects, which were detected by the service consumer. Thus, results can be validated by specialized software components that understand a service’s capabilities and its results.

To extend the information basis about available services, the achieved knowledge is shared with trusted network peers. Here, querying trusted nodes for its recommendations and disseminating own observations to trusted peers are applied. The dissemination of observations allows a service evaluation that is based on the receiver’s user preferences instead of the sender’s preferences and ratings.