A significant characteristic of global computing is the need for secure interactions between highly mobile entities and the services in their environment. Moreover, these decentralised systems are also characterised by partial views over the state of the global environment, implying that we cannot guarantee verification of the properties of the mobile entity entering an unfamiliar domain. Secure in this context encompasses both the need for cryptographic security and the need for trust, on the part of both parties, that the interaction will function as expected. In this paper, we explore an architecture for interaction/collaboration in global computing systems. This architecture reflects the aspects of the trust lifecycle in three stages: trust formation, trust evolution and trust exploitation, forming a basis for risk assessment and interaction decisions.