CIDR 2021

The paper entitled “Accelerating Complex Analytics using Speculation” by Panagiotis Sioulas, Viktor Sanca, Ioannis Mytilinis and Anastasia Ailamaki was presented at CIDR 2021 on January 11th.

The paper addresses analytical applications, such as exploratory data analysis and decision support, that process complex workloads including sequences of inter-dependent queries. In particular, it proposes a new query processing paradigm that accelerates inter-dependent queries using speculation. It enables fast and accurate predictions through approximate query processing (AQP), and efficiently validates speculations through a new streaming join operator. In case of mispredictions, instead of discarding progress, it applies corrective actions to incrementally repair the result. The experiments show that task parallelism is improved, queries run faster, and more importantly, the speedup is increased as a function of query complexity.