COMPARISON OF MULTIDIMENSIONAL DATA MODEL IMPLEMENTATIONS ON COLUMN-ORIENTED NOSQL DATA BASE

• Main Author: EFFENDI (NIM : 23514089), RIDWAN
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/30455
• Institution: Institut Teknologi Bandung

<p align="justify">One alternative to managing large data in a data warehouse is to take advantage of NoSQL database. Previous researches by Khaled Dehdouh et al. (2014) and Max Chevalier et al. (2015) utilize NoSQL database as the technology to develop data warehouses. Both use HBase, a column-oriented NoSQL database that stores data by column and is suitable for analysis needs. Max Chevalier et al transforms the concept of multidimensional data model into a logical model suitable for NoSQL database, whereas Khaled Dehdouh et al. proposes a column-oriented NOSQL aggregation operator called CN-Cube (Columnar NoSQL Cube) since HBase is not provided with aggregation operators. This thesis performs a performance analysis of both studies by implementing and then comparing the two. It is concluded that Max Chevalier et al. benefits from the use of lattices that speed up the query process, but at the same time, it also increases the space usage for data storage. Khaled Dehdouh et al. method requires less space because it only requires a table to store all the dimensions and measures, but in terms of query speed, it decreases along with an increase in the amount of data stored.<p align="justify">