Why do we emphasize on the architecture? in a nutshell, if the architecture blueprint for the data warehouse is not defined and designed correctly, it starts slowing down the multiple layers of data processing leaving the users frustrated and IT baffled. Incorrectly designed systems often lead to program failures and that often gets blamed on poor requirements definition and so on.

There are several layers of architecture in a data warehouse, you start with data architecture, ETL architecture, database architecture, BI architecture and security architecture. All these layers also are integrated and presented in a solution architecture. Although one can argue that a large part of this is infrastructure, event that team requires and architect to lead the effort.

in order to ensure success of any data warehouse and business intelligence initiative, I strongly recommend that all areas of architecture be considered as a team and be chartered with the responsibility to build out the best suited architecture. As we move towards the cloud computing era, data warehouses will exist in clouds and be accessible across the enterprise. This will demand better architects as just discussed.

Hence in the end of the day it is all about the architectures.

1. A business case - For every change in this environment a business case is needed. Without a strong business case, no changes are feasible in the data warehouse. Simple changes maybe an easy in, but anything new or complex will need investigations and sponsorship.

2. IT driven and business used - The data warehouse largely becomes IT driven. This may not be a good idea, since the business usage will be not embraced by all users. The business may also try to use workarounds to get the data and the information they need

How do you enable a successful chargeback model

1. Organizational Alignment - A key driver, that I have always emphasized on, this is a critical component.

2. Business stakeholders - Involve a business stakeholder in every steering committee and make the business own facets of the problem. This will ensure that even if IT drives the solution, business owns the problem.

3. Roles and Deliverables - A set of clearly defined roles and deliverables will be needed to clarify who owns what components.

4. Defined SLAs - Another key component is the definition of service level agreements on data, availability etc that need to be defined.

For a successful chargeback model to be implemented, a few ideas are outlined here, more analysis and customization of the concepts will be needed to make this a possibility in the real world.

Why do we need to look at this entire ecosystem rather than specific aspects. In my humble opinion, the data warehouse architecture is the backbone of your business intelligence solutions. If this backbone is not designed and built with the appropriate layers and the right sizing, there are other issues downstream which result in complex end solutions.

We as an industry have built data warehouses and datamarts to solve the ever growing need for more information in the enterprise. As we continue to mature and grow in providing solutions, let us look at the holistic picture than the silo that needs a solution. As a part of this channel, I will be covering an entire spectrum of the architecture layers across the data warehouse.