DataHabitat Architecture Overview

The DataHabitat product line is comprised of two main components: the engine and the data models (encyclopedias). The data models are source specific, whereas the engine is unique. Data Models are currently available for 13 popular mid-market financial systems including Microsoft Great Plains, ACCPAC and Best Software.

The architecture is designed to allow for the rapid development of Packaged Business Intelligence (PBI) products, that are open, extensible and user manageable. Data Modeling, ETL, Data Store, Data Visualization, Query, Analysis and Reporting tools are all embedded in a single package. This approach greatly simplifies and demystifies a process that has been historically burdened by complexity. In addition, the end-user has access to the ETL engine, being able to load data from different sources and create new metrics in the data warehouse. This feature, we understand, is what makes it possible to deploy extensible data warehouse solutions to the mid-market, where companies have little or no IT staff.

Any ODBC data source, relational or not, can be loaded into the data warehouse. By default the data warehouse is deployed in a Microsoft SQL or Visual FoxPro database (which comes embedded in the product), but can also de deployed in any ODBC-compatible data platform including Oracle, DB2 and SQL Server. It is a simple matter of changing a table property to deploy the data in a different data store.

During the extraction phase, basic metadata is collected from the source databases. The metadata is then extended to include a semantic layer so as to provide business names to tables and fields. The metadata describes objects such as entities, attributes, relationships, summaries, metrics (derived attributes), sorts, filters, graphs, reports and queries.

The data warehouse is relational by default, but can be de-normalized to provide flexibility. The data warehouse can be exported for use by common OLAP tools.

The end-user online analytical tools and report writer are metadata driven. Navigating by relationships, the user has drill-down and drill-up capabilities and a series of tools to filter, slice and dice, summarize, group and graph data dynamically.

The Report Writer takes advantage of the pre-defined relationships and filters to provide users access to the data model, without requiring knowledge of the data structure.

The ETL is Windows-based while the Query and Reporting tools have both Windows and Web-based user interfaces.

Download DataHabitat Architecture Overview