Big-Parallel-ETL: New ETL for Multidimensional NoSQL Graph Oriented Data

Abstract

Abstract The quantitative explosion of digital data derived from social networks, smart devices, IoT sensors, etc is eventuated by the Big Data concept considered as a very important aspect in the performance improvement of traditional decision-making systems since it reveals serious challenges to be addressed. Therefore, the main purpose of this research paper is the integration of NoSQL Graph-oriented Data into Data Warehouse to deal with Big Data challenges especially with the absence of similar approaches to the best of our knowledge. In this paper, we propose a new approach called Big-Parallel-ETL that aims to adapt the classical ETL process (Extract-Transform-Load) with Big Data technologies to accelerate data handling based on the famous MapReduce concept characterized by its efficient parallel processing feature. Our solution proposes a set of detailed Algorithms based on several rules able to conceive rapidly and efficiently the target multidimensional structure (dimensions and facts) from the NoSQL Graph oriented database.