Key points are not available for this paper at this time.
A replicated database architecture is described in which updates processed at a site must be saved to allow reconcilliation of newly arriving updates in a way that preserves mutual consistency. The storage space occupied by the saved updates increases indefinitely, and periodic discarding of old updates is needed to avoid running out of storage. A protocol is described which allows sites in the system to agree that updates older than a given timestamp are no longer needed and can be discarded. This protocol uses a "distributed snapshot" algorithm of Chandy and Lamport and represents a practical application of that algorithm. A protocol for permanent removal of sites is also described, which will allow the discarding of updates to continue when one or more sites crash and are expected not to recover.
Sarin et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: