Recently, several strategies for transaction processing in partitioned distributed database systems with replicated data have been proposed. We survey these strategies in light of the goal of maintaining reliability. Extensions and combinations are then discussed, and guidelines for the selection of a strategy for a particular application are presented.