Key-value data warehouse : Models and OLAP analysis
Abdelhak Khalil, Mustapha Belaissaoui
Abstract
OLAP is a technology that allows analysts to perform multidimensional data analysis, and view business data from different angles. It enables users to analyze informations from multiple and heterogeneous databases. Over the last few years, Not-only-SQL(NoSQL) databases were created and gained a lot of interest and popularity in OLTP systems. In despite of that, a few works have investigated the possibility to use NoSQL technologies to implement big OLAP cubes. Our main goal is to propose a new model for managing cubes into NoSQL key value stores, for this purpose, we present in this paper two approaches which allow cubes to be implemented under key value model, and we show how these structures can then be used to support decision making by responding to OLAP basic analytical operations. The first approach discusses how to match the star schema with key value model based on a tabular structure layered on top of the distributed key-value model. The second approach performs OLAP cube computation named CV-CUBE operator by converting dimension members into binary keys.