Some guidelines for choosing models for distributed database systems

A presentation is made of results obtained from a study of the effects of model complexity on design accuracy in the context of distributed database systems. This design problem is studied with respect to one metric: average number of sites accessed by a transaction. As a result the authors conclude that the degree of complexity of a model does not necessarily reflect the accuracy of the resulting design. This suggests that designers can drastically reduce the cost and time involved in computer system design through diligent choice of system models.<
>