AN ACL FOR SPECIFYING FAULT-TOLERANT PROTOCOLS

Agent communication languages (ACLs) should allow the developer to adopt human-like communication mechanisms in agent programming, facilitating the development of distributed protocols in multi-agent systems (MASs). However, to implement robust protocols, ACLs should provide a way to deal with the failures of agents, as MASs are prone to the same failures that can occur in any distributed software system. In this paper, we address this issue showing how an asynchronous ACL that provides high-level mechanisms to deal with crash failures of agents can be effectively used to specify fault tolerant protocols.