API federation in a polystore

Abstract

This thesis describes the design, implementation, test, and evaluation of an extension to a well-known polystore called BigDAWG. As a polystore, BigDAWG binds together multiple diverse data sources into a single system that allows primarily analytical queries to be easily executed against the most appropriate engine for the job. The extension described allows the polystore to source data from APIs such as those publicly available on the internet. Further, an administrative interface was constructed to facilitate the adding, editing, and deleting of access to these APIs. These additions place power into the hands of the BigDAWG polystore user such that not only do they have the ability to source data from many new sources, but can process that data quickly and easily by using the built-in polystore casting facilities, cutting down on the time necessary to create custom ETLs to achieve the same result.