DigestJoin: Expediting joins on solid-state drives

Abstract

This demonstration presents a recently proposed join algorithm called DigestJoin. Optimized for solid-state drives (SSDs), DigestJoin aims at reducing intermediate join results and hence expensive write operations while exploiting fast random reads. The demonstration system consists of an implementation of DigestJoin in the open-source PostgreSQL database management system on an Intel SSD. In the demonstration, we will showcase the performance benefits of DigestJoin in comparison to a traditional join algorithm and highlight the workloads in which DigestJoin is particularly favorable. © Springer-Verlag Berlin Heidelberg 2010.