Common use of Concluding Remarks Clause in Contracts

Concluding Remarks. Since the pioneering work of [1], the question of what can be computed in a totally anonymous distributed systems has been extensively studied. Some results depended on prop- erties of the communication graph (e.g. [4, 22]). Some work considered shared memory for the “wake up” problem [12], others considered consensus [3]. The power of anonymous broadcast systems, in comparison with anonymous shared- memory systems has also been studied [2]. None of these considered process failures. Anonymous processes with crash failures have been considered more recently [5,6,8,11,16,19]. In [15], Byzantine agreement was studied in a model with a restricted kind of anonymity: processes have no identi- fiers, but each process has a separate channel to every other process and a process can detect through which channel an incoming message is delivered. It was shown that Byzantine agreement can be solved in this model when n > 3t. This paper is the first to study a distributed system model with homonyms, i.e., with a limited number of identifiers. The model unifies both classical (non-anonymous) and anony- mous models and is interesting from both a theoretical and a practical viewpoint. We completely characterized the solv- ability of Byzantine agreement in this model, precisely quan- tifying the impact of the adversary, with some surprising results. We focused however on agreement and many other problems would be interesting to consider. We also focused on computability and complexity is yet to be explored.