This paper describes a pipelined processing technique for queries involve data from distributed, autonomous relational databases. We highlight the basic idea, present a preliminary result, and sketches the planned work towards the development of a more complete multi-database query processor based on the proposed idea.
