With the advent of Big Data technologies, there is a need for new benchmarks to evaluate information decision systems. In particular, existing benchmarks for multidimensional data warehouses need to be adapted to increasing volume and diversity of data. In this context, we propose a new benchmark dedicated to data warehouses that can support different information systems (relational and NoSQL) and data models (snowflake, star, flat) structured or non struc- tured. In order to scale in volume, it supports parallel generation of data on multiple computers (cluster). It can generate diverse data structures, by supporting generation of multiple different schemas. In this paper, we present a new benchmark for Big data, called KoalaBench and the first experimental results of its use.