Relational Programming

Abstract

Relational databases have become one of the most important software components in the world, ubiquitous in computing systems from smartphones to data centers. Nevertheless, traditional relational databases are struggling to meet the demands of modern data analytics, which involve more and more complex computation over an increasing variety of data. In this dissertation, we reinvigorate relational data processing with a new language foundation for writing relational programs, a new algorithm for the relational join, and techniques to optimize relational queries. We extend the reach of the relational paradigm in both expressiveness and efficiency, ushering a new era of relational programming.