Denial of service detection in mobile agent systems with dependencies in computations

In connection with mobile agents, security aspects are extensively being discussed. In this context, denial of service attacks can cause tremendous economic damages and are, therefore, of considerable interest to researchers as well. Here, a special kind of denial of service attack is considered, focusing on malicious hosts that either delete received agents or prevent them from continuing their route. In general, the prevention of such attacks is not possible. A detection method for the a posteriori identification of such malicious hosts is presented and discussed. Agent owners can use the output of this procedure when composing future routes of their agents. This method relies on a novel protocol that combines the application of well known cryptographic primitives and a set of behaviour rules. The protocol takes into account given visiting-order constraints for routes, which arise due to dependencies of computation results on an agent's journey. The protocol also work s in the case of colluding hosts.