Abstract:

 During the last two decades, architecture of traditional Data Warehouses (DW) have played a key role in decision support system. A typical architecture of traditional DW involves source systems, components implementing data collection process, stores (a central repository and data marts), as well as consumers: reporting and analytic applications. However, the emergence of several new Internet services, Mobile applications, Web applications, Social media (Facebook, Twitter, and Instagram, so on), Devices, and Sensors, requires new Data warehousing systems and suitable architectures: large volume and format of collected datasets, data source variety, streaming, integration of unstructured data and powerful analytical processing. In this paper, we propose a two level architecture for Data warehousing and OLAP over Big Data. The first level Platform Independent Architecture (PIA) allows identifying and specifying the main components of the architecture to collect, store, transform and process the different kind of data. This level is technology independent and focuses only on the requirements of the data features and the processes needed to design a Data warehousing and OLAP over Big Data. The second level Platform Specific Architecture (PSA) allows specifying the platforms and technologies that would be used to achieve the different steps from the collection of data until the reporting and analytic applications.