Extending Agents by Transmitting Protocols in Open Systems

Agents in an open system communicate using interaction protocols. Suppose that we have a system of agents and that we want to add a new protocol that all (or some) agents should be able to understand. Clearly, modifying the source code for each agent implementation is not practical. A solution to this problem of upgrading an open system is to have a mechanism that allows agents to receive a description of an interaction protocol and use it. In this paper we propose a representation for protocols based on extending Petri nets. However, this is not enough: in an open system the source of a protocol may not be trusted and a protocol that is received may contain steps that are erroneous or that make confidential information public. We therefore also describe an analysis method that infers whether a protocol is safe. Finally, we give an execution model for extended Petri nets.