Now, let's take a look at what we need in the host PC or a high-end mobile phone-like device.

Host OS
This the OS that originally resides on the host PC or the high-end mobile phone. the virtualized OS interacts with this host OS to access the hardware resources of the host PC or the high-end mobile phone.

Processing unit
This unit consists of CPU, RAM, IO and other related hardware needed for the processing the requests from the virtualized mobile storage device.

Display Unit
This is the display screen or monitor which is used by the virtualized mobile storage device to display the responses to the requests from the miniaturized data warehouse which is hosted in the virtualized PC environment.

USB
The mobile storage device and the host PC or high-end mobile phone communicate with each other via a high-speed USB 2.0 port or USB interface.

It should be noted that USB is not strictly needed. Rather, there's a need for some way of connecting the high-capacity storage to the processing environment. For example, this could well take the form of a Compact Flash card or SD card within, say, a Palm Treo smartphone.

Benefits of a Miniature Data warehouse
A miniature data warehouse can be used to bring about personalization in the business intelligence and data warehouse reporting space. The personalized and miniaturized data warehouses can be hardware agnostic and hence can be run from any PC at any location (with minimal constraints), thus freeing the dependencies on the connectivity related infrastructure.

The miniaturized data warehouse concept can herald the rise of personalized business intelligence tools, personalized miniature databases, thus creating a separate miniaturized data warehouse ecosystem. this can eventually lead to creation of a host of analytical hand-held business intelligence gadgets, for use by both consumers and enterprises alike.

Constraints of a miniature data warehouse
While there are no technology constraints with respect to high-capacity mobile storage devices and virtualized OS implementations for these storage devices, there can be some constaints with respect to the compatibility between the virtualized mobile storage devices and the host PC or high-end mobile phones.

There can also be some constaints with respect to using the current DBMSs like MySQL (and others) in this virtualized PC environment. Similarly there can be some constraits with respect to using the current business intelligence applications in virtualized PC environments. However most of the software constraints can be overcome by the respective vendors of the tools and technologies.

Summary
While there has been much emphasis on making data warehousing systems larger and larger, the same technologies can be used in combination with the mature mobile storage devices and OS virtualization technologies to take the data warehousing systems to the other end of the spectrum -- miniaturization of data warehousing systems.

This article gives the concept connectivity between the various components that together form a miniaturized data warehousing system. These miniaturized data warehouses can be used for multiple applications like personalization, analytical hand-held business intelligence gadgets, mobile data access, and PC agnostic business intelligence in a virtualized PC environment. Together, I predict these will bring a paradigm shift in the traditional definitions of data warehousing systems, thus creating a separate miniaturized data warehouse ecosystem that is self-sustained and focused.

Harikrishna S. Aravapalli is a Senior Technical Architect at Infosys SETLabs and has 13 years of experience in databases, data warehouses, and business intelligence technologies. He worked for Wipro and Accenture prior to Infosys. He may be reached at harikrishna_sa@infosys.com.