Benchmarking Big Data OLAP NoSQL Databases

With the advent of Big Data, new challenges have emerged regarding the evaluation of decision support systems (DSS). Existing evaluation benchmarks are not configured to handle a massive data volume and wide data diversity. In this paper, we introduce a new DSS benchmark that supports multiple data storage systems, such as relational and Not Only SQL (NoSQL) systems. Our scheme recognizes numerous data models (snowflake, star and flat topologies) and several data formats (CSV, JSON, TBL, XML, etc.). It entails complex data generation characterized within “volume, variety, and velocity” framework (3 V). Next, our scheme enables distributed and parallel data generation. Furthermore, we exhibit some experimental results with KoalaBench.