Abstract

1 Introduction The management of replicated data in a distributed database is an old problem of great practical importance. Distributed data warehouses and data marts contain very large replicated databases distributed among a number of sites. Depending upon the application, the number of sites may range from only a few sites to several hundred sites. Telecommunication applications require rapid distribution of updates to all replicas with strong guarantees of consistency and availability. Early work on replication, (see, e.g. [14] for a survey) focused on eager replica update propagation. Under eager

Citations