JovianDATA: a multidimensional database for the cloud

Abstract

The JovianDATA MDX engine is a data processing engine, designed specifically for managing multidimensional datasets spanning several terabytes. Implementing a terascale, native multidimensional database engine has required us to invent new ways of loading the data, partitioning the data in multi-dimensional space and an MDX (MultiDimensional eXpressions) query compiler capable of transforming MDX queries onto this native, multi-dimensional data model. The ever growing demand for analytics on huge amount of data needs to embrace distributed technologies such as cloud computing to efficiently fulfill the requirements.This paper provides an overview of the architecture of massively parallel, shared nothing implementation of a multi-dimensional database on the cloud environment. We highlight our innovations in 3 specific areas - dynamic cloud provisioning to build data cube over a massive dataset, techniques such as replication to help improve the overall performance and key isolation on dynamically provisioned nodes to improve performance. The query engine using these innovations exploits the ability of the cloud computing to provide on demand computing resources.