Trustworthy Service Selection Using Long-Term Monitoring of Trust Contracts

Composing a distributed software system out of independently developed and publically hosted services leads to savings in effort, time, and cost. Such an approach requires a careful selection of a set of services for realizing a particular distributed software system. Hence, identifying and quantifying the trust of such publically available services becomes a prerequisite for composing distributed software systems. The quantification of the trust associated with a service is a challenging task due to the subjective nature of trust, lack of standards, and associated uncertainty. Moreover, the trust of a service is not a static value and should be updated periodically. To achieve a proper quantification of a trust associated with a service, there is a need to periodically collect, monitor, and aggregate various evidences available about that service. This paper augments previously defined quantification of the trust by monitoring and aggregating various available evidences. This enhanced approach is used to improve the service selection process. Publically available mobile app services from the Android marketplace are utilized, as the dataset, to empirically validate this approach.