Diploma and Master Theses (authored and supervised):
"Trusted Information Sharing";
Supervisor: S. Dustdar, F. Skopik;
Institut für Informationssysteme, AB Verteilte Systeme,
One of the challenges we face in collaborative environments nowadays is efficient sharing of sensible information. Most common solutions to this problem are static role-based access control mechanisms, which require administrators to manage roles and privileges manually. In a system where a number of users is well defined and does not change more often, this kind of security model works sufficiently.
However, a typical scenario in a collaboration system looks different and is more demanding, because of the inherent dynamics in such environments. In flexible collaborations, where teams and roles of members dynamically change, we need more sophisticated mechanisms to enable efficient information sharing. In a typical scenario we have a growing up collaborative network, where more and more information are available every day. Consequently the number of collaborative partners is growing, which in turns leads to increased number of interactions between partners as also software services. Furthermore interactions can be categorized in terms of success or failure. By monitoring interactions and analyzing someone's behavior, we are able to determine, if a particular person is a trustworthy collaboration partner. Assuming having virtual teams, in which team members do not have any personal experience with each other, trust relations can be based on available knowledge about past and current collaborations.
Periodically updated trust relations can be used to improve information sharing by adapting the amount of shared data. Hence, access to, e.g., personal files, is granted to currently close and reliable collaboration partners only. This solution provides a mechanism to cumulate data regarding trust without third-party involvement as also improve future collaborations by reacting to each collaboration behavior. The major advantage of using automatic trust inference instead of manual ratings or static roles is the support for autonomic adaptations of collaboration scenarios. In this thesis, we focus on enabling information dissemination in collaborative environment. In order to explain and clearly illustrate our concept, we present a science collaboration scenario. We also investigate trust related concepts with regard to managing social trust in flexible collaboration environments. In the practical part of this thesis, we describe implementation of our web-based solution on efficient sharing of sensible information.
Created from the Publication Database of the Vienna University of Technology.