Evaluating push-down on NoSQL data sources

Abstract

Reducing execution time of data integration processes is a challenging and still only partially solved research and technological problem. A few optimization techniques have been developed and applied in practice. One of them is called push-down. It consists in moving the execution of some data processing tasks into a data source. This technique has been developed and evaluated on relational data sources. In this paper, we report our evaluation of push-down on NoSQL data sources, i.e., MongoDB and Cassandra, deployed on a micro-cluster.

NODW Framework for Data Warehousing -A NoSQL Big Data Perspective

“…It does not have any built-in mechanisms for ensuring referential integrity across tables. In a traditional relational database, foreign keys can be used to link records in different tables, but Apache Cassandra does not support this [3]. Relational databases are typically designed with normalization in mind, which is the process of organizing data into separate tables to reduce data redundancy and improve data integrity.…”

“…Target Cassandra Denormalized Schema sandra approaches data modeling differently than relational modeling. Instead of starting with the data model, it starts with the query model [3]. This means that rather than modeling the data and then writing queries, model the queries first and organize the data around them.…”