Fast online analytical processing for Big Data warehousing

In an organizational context where data volume is continuously growing, Online Analytical Processing capabilities are necessary to ensure timely data processing for users that need interactive query processing to support the decision -making process. This paper benchmarks an innovative column -oriented distributed data store, Druid, evaluating its performance in interactive analytical workloads and verifying the impact that different data organizations strategies have in its performance. To achieve this goal, the well-known Star Schema Benchmark is used to verify the impact that the concepts of segments, query granularity and partitions or shards have in the space required to store the data and in the time needed to process it. The obtained results show that scenarios that use partitions usually achieve better processing times, even when that implies an increase in the needed storage space.This work is supported by COMPETE: POCI-01-0145-FEDER-007043 and FCT - Fundacdo para a Ciencia e Tecnologia within the Project Scope: UID/CEC/00319/2013 and the doctoral scholarship PD/BDE/135101/2017; by European Structural and Investment Funds in the FEDER component, through the Operational Competitiveness and Internationalization Programme (COMPETE 2020) [Project no 002814; Funding Reference: POCI-01-0247-FEDER-002814]