The scope of the thesis is twofold: (i) study trust models synchronized with levels of security and (ii) propose an extension to the current WS-Trust mechanisms. The extension will consider the evaluation of the past interactions between WSs and metrics associated with trust violations of these interactions. The results can be applied to monitor the SLA in WS-based systems and for early threats detection.

We considered a new mechanism to build and self-adapt the mutual trust of the interacting WSs. On the client side, we enforce a start with the less critical applications supported by the client-WS (use-context), while on the provider side, we consider various levels of trust. The levels of trust are induced based on the statistical interactions.

The proposed approach is introducing the trust profile and defining mechanisms to evaluate the evolution of trust based on current interactions. The theory is supported by experiments.