Distributed database management model and validation

The authors describe a simple, yet effective, distributed database model that simulates database usage and buffer management in the distributed data processing environment. The model is table driven such that database access requirements, file location, and other information defining the database environment are set up internally in several tables, and linked lists represent the directory and data blocks. Each database transaction is defined and represented by a transaction flow diagram (TFD), and a sequence of TFDs representing an operational scenario is input to the model. The model `executes´ input TFDs by looking up tables, and performs buffer management for directory and file data while logging history and gathering various statistics on database usage and buffer management. The performance data are used for database access overhead measurement, database workload characterization, and buffer allocation. Disk access frequency and response time are used to validate the simulation results.