Design, Implementation, and Evaluation of a Relational Database Engine for Variable Length Records

This paper reports the design, implementation, and evaluation of a relational database engine. The engine has special purpose hardware, the nucleus of which is a pipeline two-way merge sorter, to execute sorting and relational algebra for variable length records. Jn the execution of these operations, key fields are extracted from records. The engine has on-the-fly processing in which it inputs data directly from disk. The evaluation shows the characteristics of processing time, the effect of variable length records, and the effectiveness of on-the-fly processing.