Trust negotiation frameworks allow communicating parties
to incrementally establish trust in one another to achieve
security goals. The goals we focus on deal with sharing
sensitive information that is protected by a disclosure policy.
Traditionally, such policies are satisfied through the use
of signed credentials that express role memberships or attributes.
The requirements in these policies are known as
provisions, and represent past and present state. Although
useful, provisions can sometimes prove too rigid, are susceptible
to schema-matching problems, and cannot provide assurances
on how information is used once it has been shared.
In this paper, we propose a means of augmenting trust
negotiation frameworks to support obligations, which are
commitment-based requirements to perform certain actions
in the future. We provide a metamodel for such a framework
along with a method of converting provisions into sets
of obligations. We analyze the complexity of this conversion,
and then provide a study of obligation optimality during the
negotiation.
