On distributed real-time systems the mStar environment, net-based learning and context-aware applications

Abstract: This doctoral thesis presents two frameworks for distributed real-time systems, the mStar environment for Computer Supported Collaborative Work (CSCW) and the Alipes architecture for context-aware applications, from the perspective of distributed teamwork and net-based learning. The mStar environment was designed to be symmetric and fully distributed, which allows all users equal access and thus full interactivity, as well as scalable through the use of IP multicast and a server-less design. The environment makes use of reliable multicast, network resource management and packet loss recovery techniques to increase robustness. Heterogeneous networks and terminals are supported through tunneling and transcoding of media. Asynchronous use of the environment is made possible through support for recording and replay of sessions. It is therefore well equipped to meet the requirements for net-based learning, as the inherent time- and location-independence allows students to follow distributed courses, when otherwiselarge geographical distances or time restrictions otherwise would limit where and when education could be offered. The student can be anything from a full-time student attending lectures physically at the university, to a part-time student following courses from his home during evenings and weekends. Students will thus have increased possibilities of taking part in education. The Alipes architecture for context-aware applications allows multiple positioning techniques to be seamlessly interchanged and combined, enabling applications to utilize a single interface, yet benefit from several advantages that single positioning techniques cannot offer alone. Add peer-to-peer interchange of position information using ad-hoc networks, and the platform offers a wide variety of techniques to be interchanged or combined, with obvious advantages such as increased coverage and accuracy. Privacy issues are central to managing a user?s context information, for example his position, as that information could cause serious violations of personal integrity if misused. The Alipes architecture handles privacy through general criteria and contracts between users and location servers. Information exchange is typically carried out on a peer-to-peer basis using ad-hoc networks. Integrated examples could be combining locating a nearby tutor with setting up a conference call to initiate a tutoring session, using context information to enhance the perceived feeling of presence within the mStar environment or to optimize network-usage depending on the user?s context. A final conclusion is that systems aiming at enhancing the social aspects of distributed applications by using context information might be important, if not vital, when creating new real-time services for mobile terminals.