ChaseFUN: a Data Exchange Engine for Functional Dependencies at Scale

Despite their wide use and importance, target functional dependencies (fds) are still a bottleneck for the state-of-the-art Data Exchange (DE) engines. The consequences range from incomplete support to support at the expense of an important overhead in performance. We demonstrate here ChaseFUN, a DE engine that succeeds in effectively mitigating and taming this overhead, thus making target fds affordable even for very large-sized, complex scenarios. ChaseFUN is a custom chase-based system that essentially relies on exploiting chase step ordering and constraint interaction, so as to piecemeal process, parallelize and dramatically speed-up the chase. Interestingly, the structures and concepts at the core of our system moreover allow it to seamlessly uncover a range of usually opaque details of the chase. As a result, ChaseFUN’s two main strengths are: (i) its significant scalability and performance and (ii) its ability to provide detailed, granular insight on the DE process. Across our demonstration scenarios, we will emphasize our system’s practical performance and ability to scale to very large source instances and sets of constraints. Furthermore, we will aim at providing the user with a novel, behind-the-scenes view on the internals of the ongoing chase process, as well as on the intrinsic structure of a DE scenario. CCS Concepts •Information systems Ñ Data exchange;