Optimization of join strategies in distributed databases

The paper presents a structured approach to the problem of minimizing the join cost in a relational distributed environment. A tree model is used to represent a query and a set of tree equivalence classes for query representation is identified corresponding to the space of all the feasible strategies to execute the query. The optimal strategy is then chosen by a dynamic programming approach which exploits the properties of the tree model, although the computational complexity remains exponential in the size of the problem