Database Replication Using Epidemic Update

Due to severe performance penalties associated with synchronous replication,there is an increasing interest in asynchronous replica management protocols inwhich database transactions are executed locally, and the effects of thesetransactions are incorporated asynchronously on remote database copies.However, the asynchronous protocols currently in use either do not guaranteeconsistency and serializability as needed by transactional semantics or theyimpose restrictions on placement of data and on which data objects can beupdated. In this paper we investigate an epidemic update protocol thatguarantees consistency and serializability in spite of a write-anywherecapability. We conducted experiments on a detailed simulation of adistributed, replicated database to evaluate this protocol. Our resultsestablish that this epidemic approach is indeed a viable alternative totraditional eager update protocols for a distributed database environment whereconsistency and full serializability are needed.