Far more than a technical project, the Rensselaer Data Warehouse Project requires an understanding of the Rensselaer environment and underlying business processes that describe how information is produced and consumed.
Given the breadth of this project and its potential impact on many types of users, the data warehouse environment is to be developed in an evolutionary fashion, starting with a well-defined set of prioritized requirements.
Implementation of the data warehouse project is organized into 3 distinct phases, summarized in the column to the right, and explained in detail below.
infrastructure phase (July 2001 - December 2001)
This initial stage includes all activities necessary to get the project off the ground, including user group definition, hiring and training of staff, hosting vendor visits, etc.
high-level business requirements
This stage involves high-level business analysis with Rensselaer executives and decision-makers, aiming to:
- understand divisional key strategic initiatives
- identify divisional key performance indicators and success metrics for each strategic initiative
- determine core business processes monitored by each division/unit and want to impact
- determine potential impact on division/unit performance metrics with improved access to information
business requirements prioritization
Faced with the potential of many demands from all the administrative and academic divisions and units at Rensselaer, this stage involves the prioritization of all high-level business requirements.
develop detailed project scope and plan
Based on the prioritized list of requirements, this stage involves the definition of project scope and the project plan.
build technical infrastructure
This stage involves procuring and installing the base system hardware and software necessary to implement the data warehouse architecture.
implementation phase (January 2002 - December 2004)
For each subject area (i.e. each business area), there are 6 basic groups of activities, as described below. Note that the actual implementation time of each subject area will vary based on the complexity of the subject areas (determined during the "gather user requirements" stage). On average, approximately 6 months should be allocated for each subject area.
gather user requirements
Business users and their requirements impact almost every decision made throughout the implementation of the data warehouse. Therefore, this stage consists of those activities necessary to obtain a full understanding of the specific business area.
This stage involves the logical design of the subject area and the detailed analysis of data sources and necessary transformation rules.
This stage translates the logical modeling into a physical database.
data staging design & development
This core stage of building the data warehouse involves the design and development of the processes that load data from the operational systems into the data warehouse.
The general front-end templates are designed during this stage. If necessary, standardized report frames are created. Note that the actual timeframes will vary based on the subject areas defined in the "gather user requirements" phase; for planning purposes, an average of 2 months should be allocated for each subject area.
subject area deployment
This stage involves the actual deployment of the data warehouse environment to the business community.
(I) The Infrastructure Phase involves purchasing the hardware and software necessary to build the data warehouse, hiring and training the appropriate personnel, prioritizing the business requirements, and generally getting the project off the ground.
(II) The Implementation Phase represents the core of the project in which business subject areas will be analyzed, prioritized, and implemented as data marts within the data warehouse.
(III) The Growth Phase is dedicated to addressing the growing and evolving needs of the business users.
for up-to-date status information
growth phase (January 2005 and beyond)
Unlike projects involving traditional operational systems, the maintenance and growth phase of the data warehouse project will require a different approach.
In traditional operational systems development, the completion of the deployment phase generally indicates that the project is done, though maintenance tasks do exist.
Conversely, the data warehouse project is generally never done; there is little decline in the ongoing needs of the users during the data warehouse maintenance and growth phase.
Specifically, it will be necessary to:
- provide ongoing support
- provide ongoing user education
- maintain technical infrastructure
- monitor and respond to end-user query performance
- monitor and respond to data staging and transformation performance
- monitor ongoing success
- identify changes and enhancements
- identify needs for new internal or external data sources
- identify needs for new subject areas
- identify changes in business and propagate such changes to metadata