Fault-tolerant broadcast in anonymous systems

The broadcast service spreads a message m among all processes of a distributed system, such that each process eventually delivers m. A basic broadcast service does not impose any delivery guarantee in a system with failures. Fault-tolerant broadcast is a fundamental problem in distributed systems that adds certainty in the delivery of messages when crashes can happen in the system. Traditionally, the fault-tolerant broadcast service has been studied in classical distributed systems when each process has a unique identity. However, very recently have appeared new distributed systems, such as sensor networks, where unique identity is not always possible to be included in each sensor node (due to small storage capacity, reduced computational power, a huge number of elements to be identified, etc.). In this paper, we study the definition and implementability of the fault-tolerant broadcast service in anonymous asynchronous systems, that is, in asynchronous systems where all processes have the same identity, and, hence, they are indistinguishable (they may have the same code).