Net-based Learning and the mStar Environment

Sammanfattning: This licentiate thesis focuses on how to design a robust and flexible environment for net-based learning. The environment is required to be scalable for large groups, robust over lossy networks, flexible to meet different learning scenarios, and support interactivity. The mStar environment presented in this thesis has been developed to meet these requirements. The environment is scalable through the use of IP multicast and a server-less design. Robustness is achieved by separating traffic by loss tolerance, where traffic that accept no loss uses a reliable multicast protocol and traffic that can accept some loss uses repair techniques. To enhance robustness even further, network resource management is suggested. Everything from small group meetings to large lectures is supported, which together with the possibility to use the tools asynchronously gives the flexibility needed. The tools in the environment are fully symmetric, which allows everyone equal access and thus supports full interactivity. Students using the mStar environment is no longer bound by physical distance, since they can easily take part of lectures, seminars and group discussions electronically. They can also be connected via not multicast enabled low bandwidth networks, since they can tunnel concentrated traffic to thir local hosts. The students are also less bound by time, as they can use the mMOD system to access recorded sessions or the WWW to access course information.