In distributed database, each logical data item can be replicated to enhance availability, reliability, and performance. A major restriction to using explication is that replicated copies behave like a single copy, i.e. mutual consistency must be preserved.
In this thesis, we extended the conventional nested transaction model to support replicated abstract data types, and proposed basic concurrency control to preserve mutual consistency of replicated data. Compared with basic concurrency control, a new concurrency control, which minimizes the message traffic, was proposed. Two concurrency control schemes were evaluated in terms of simulation approach.
