Minimization of the total volume of data transmissions (and consequently the time required) is an important goal in the processing of queries in distributed database systems. The problem of finding an optimal schedule of operations for a general class of queries in an arbitrary system is shown to be NP-hard [1]. In this paper, a simple subclass of queries in a special kind of network (namely a chain connected network) is considered and optimal solution strategies are developed. The authors show that if the result node coincides with one end of the chain, the optimal solution strategy is either a serial Semi-Join Reduction Program (SJRP), or two serial SJRPs merging at the result node. This extends the results obtained by other researchers which have shown that parallel transmissions are not cost effective in networks of arbitrary topology if the objective is total minimization.
This paper, like other notes appearing in Information Processing Letters, is fairly short and thus does not contain background information (material). The reader, therefore, must have the technical knowledge necessary in this subject, or should first read the earlier work by Gurse [2] cited in the article.