Distributed Environment for Mobility-Aware Computing

Within the modern services society, mobility has become one of the main driving factors. Supported by the trend of decentralization in information and communication technology, a great diversity of portable and networked devices has evolved. The use of these devices has increased continuously and leads to a ubiquitous availability of information and services for mobile users. Next to the portability of devices, software systems themselves have become mobile - enabled by an increased awareness of the devices' mobility and thus by the ability to adjust to highly changing execution contexts.

With this background, the research field of Mobile Computing subsumes efforts to adapt modern mobile information systems, middleware platforms, and applications to new evolving needs of mobile users. In doing so, context-based systems play a major role, as these systems supplement mobile applications by adding context knowledge to achieve flexibility and adaptability for such systems. However, in reality most of these currently existing systems are rather ad-hoc static, monolithic structured, and closed. In consequence they are mostly restricted to support rather specific and short time tasks. On the other hand, more general systems being able to support also long running and spontaneous user tasks, are still hardly realizable with existing methods and techniques.

Therefore, this project adresses these shortcomings of current mobile systems and supporting middleware platforms in a detailed study. On the basis of the results of this study, the concept of context-based cooperation is proposed and elaborated in detail. As a consequent evolution of concepts derived from established context-based systems, this approach serves to integrate support of user-centric, long running, and ad-hoc tasks into (new) mobile system environments. Finally, by merging context-based cooperation with well-established concepts and methods of Service-Oriented Software Architectures, the realization concept of Mobile Processes is elaborated. Based on that, a prototype implementation of a system platform for Mobile Processes is realized in order to prove the practical realizability as well as to evaluate the feasibility of the introduced concepts of context-based cooperation and Mobile Processes.

Finally, the practical result of this thesis is a middleware system platform that enables (mobile) environments to execute ad-hoc arising, complex processes without a central coordination mechanism. Thus, the introduced approach represents an important enhancement of context-based systems towards the vision of Ubiquitous Computing.

Based on the concept of Context-based Cooperation, service-based (business) processes can be initialized on mobile devices and can be executed in a decentralized and cooperative way by dynamically using suitable stationary and mobile resources. The DEMAC project realizes the concept of such long-running, user-centric processes for mobile systems. It therefore provides a context-aware middleware to support the design, the description and the distributed execution of these so-called Mobile Processes.

However, the mobility and autonomy of process participants, the required flexibility and dynamism and the decentralized execution strategy lead to new challenges for a reliable and purposive execution of such processes. The recent objective of this project therefore focuses a cooperative and self-contained process management to enable adequate steering and controlling mechanisms for decentralized mobile processes. Therefore, an advanced process lifecycle management system is being developed, integrating and supporting the phases of process design, selection and allocation of resources as well as the processes' monitoring, recovery and analysis.