Logical foundation of data replication

Abstract

Copy management practiced by transactions bears some principal disadvantages. The most critical one is their bad performance. Another fundamental inconvenience depends on their restricted intensional expressiveness. We demonstrate that sophisticated copy management reduces this defect. Today many new strategies for distributed data management are developed (e.g. the snapshot concept, the escrow method), which enhance the intensional meaning of transactions. In this paper we investigate the logical foundations of data replication by relating it to the theory of logical equality. Therefore we define equality classes, which also determine the level of quality or consistency of a replicated data item. Also data processing mechanisms, by which indirectly data are assigned to certain equality classes, are discussed and assessed with regard to their extensional and intensional meaning.