Architecture for virtualization in data warehouse

Abstract

Conventional data warehouse (DW) due to structure of its schema and contents is unable to: a) support any dynamics in its source structure and contents b) unable to support hidden-subjects c) unable to provide data on-thefly i.e. real-time data and populate hidden-subjects on their evolution. To handle these problems the concept of virtualization in DW is floated here. In this study we have proposed architecture of virtualization approach. According to this approach, conventional DW is replaced by: i) a storage component called data-store ii) a Synthetic warehouse (SWH). Datastore is a non-subjective, content consistent, time-variant and integrated storage. On the other hand, SWH is only a structure, with no instances attached. It acts as a schema source for analytical processing and is mapped to its data-store. Subjective conversion is expected to be done on-the-fly. We are hopeful that this architecture will qualify all the evaluation parameters of: i) scalability ii) hidden-subjective support iii) source dynamics. © 2007 Springer.