Abstract:With the explosive growth of the Internet and the distributed applications it supports, there is a pressing need for secure group communications -the ability of a group of agents to communicate securely with each other while allowing members to join or leave the group. Prompted by the success of other researchers in applying finite-state model-checking tools to the verification of small security protocols, we decided to attempt a larger security protocol: a recently published protocol for secure group communication. Not surprisingly, creating an ad hoc abstract model suitable for model-checking required cleverness, and state explosion was always a threat. Nevertheless, with minimal effort, the model checking tool discovered two flaws in the protocol, one of which has not been reported previously. We conclude our paper with a discussion of possible fixes to the protocol, as well as suggested verification tool improvements that would have simplified our task.

No abstract

No abstract