Consider a set of locally separated parties that do not trust each other and that can only communicate bilaterally. Furthermore, consider a designated sender among them who is to consistently distribute a message to all parties in the set. Is there a protocol among the parties that allows for the reliable distribution of this message even when the sender and possibly some other parties are cheating by deviating from the protocol? Reliability thereby means that all honest parties are guaranteed to receive the same message, and that this is the original message chosen by the sender if the sender is honest himself. This problem is a special case of the so-called Byzantine agreement problem. Byzantine agreement is an important building block for fault-tolerant distributed computing and cryptographic protocols.
This work starts with an overview over existing models and variations of this problem and the discussion of some fundamental well-known solutions to the problem as well as respective impossibility results. Thereby, often an alternative, more intuitive representation is chosen than was done before. Finally, some natural generalizations of the problem are introduced and solved.
From Klappentext im Buch Generalized Communication and Security Models in Byzantine Agreement