Data Warehouse Architectures

Back to basics:

Data warehouse as a concept has been around for over twenty years. The concept keeps evolving, changing everyday. Data warehouses now are used for a variety of purposes from day to day operational reporting to predictive analytics.

Newer technologies keep emerging, computing technology keeps getting better. Vendors keep advertising benchmarks for faster, bigger data warehouse technologies. You will also hear several terms like data marts, operational data stored, DW appliances used quite frequently. There is a sea of information out there and navigating through the information may be a daunting task.

Key Data Warehouse Architectures Terms:

Hub and spoke (popularized by Bill Inmon): The data warehouse provides a centralized location for the data to be managed and stored and made available to populate the various data marts. Sources of data will be integrated in the data warehouse allowing data marts to be populated with data from more than one original source.

Bus (Kimball) Architecture: This architecture primarily is based on Ralph Kimball’s definition. Simply put, Data warehouse is the sum of all constituent data marts held together by a conform layer.

Hybrid Architecture:

Each organization is unique in terms of its requirements for a data warehouse. The needs of your organization may require a hybrid architecture that implements key concepts from both architectures.

Key Implementation Approaches:

Top down approach: Start with an enterprise model and architecture and through phased implementation construct an enterprise wide data warehouse.

Bottom up approach: Start with a subject area and build incrementally keeping a standardized conform layer.

Common mistakes to avoid:

• Lack of executive sponsorship and active business involvement
• Assuming the data exists and the data quality is fine
• Presentation layer complexity not aligned with end user skill sets
• Slow response times
• Too many, too few, or the wrong BI software tools
• Going it alone, not getting the right kind of expertise
• Underestimating the effort needed to operate and maintain the DW
• Inflexible (neither scalable nor extensible) design
• Modeling the DW to reflect the source data rather than the business
• Not solving a business problem

So how can Xtivia help?

Xtivia’s proven approach and methodologies combine industry standards with sole focus on your specific needs.

Xtivia will help you find a strong sponsor who understands the cost and level of effort required; market the solution and its value; highlight and advertise the benefits

Our team is well versed with an approach that builds data warehouses incrementally; deliver value early and often by identifying the right business problem/s and building solutions that solve the problem/s.

We will also help you set expectations with the business users; keep them in the loop, build a good BI team and use the right tools appropriately.

Find out more –

http://www.xtivia.com/business-intelligence-a-data-warehousing/business-intelligence-assessments

http://www.xtivia.com/pdfs/Xtivia-BI-DW-Assessment.pdf

Request More Information