PROJECT IMPLEMENTATION METHODOLOGY
KEYSOFT Solutions follows the Kimball Method, where independent data marts are built and then a logical union amongst these various data marts or department specific data warehouse are formed, leading up to an enterprise data warehouse. This union between the datamarts is achieved through common dimensions, also termed as conformed dimensions. We use the Kimball method, as it is less risky for the customer as compared to the CIF (Corporate Information Factory) an Inmon method, since it ensures quick wins for the customer, as opposed to the longer timeframes required by Inmon approach.
This methodology uses dimensional models, as opposed to the traditional ER models used in traditional databases. The primary reason being while the ER model is mainly optimized for storing data, the dimensional models allows business users themselves to query, analyse data and generate reports without depending on their IT department, or requiring to have significant technical or programming skills.
Apart from our use of Kimball’s methodology to create the logical design of the Data warehouse, we also use RAD (Rapid Application Development) method, which is an Iterative process model, for executing the overall project life cycle. These methodologies together are used as a basic framework and tailored to meet specific project needs.
The RAD methodology is designed for speedy delivery of tasks based on a work breakdown structure. The major difference between RAD and other methodologies is the management techniques that are optimized for speed.
The detailed steps followed by KEYSOFT in the execution of Data warehouse and BI projects, are:
1. Requirements Gathering Phase
During the requirements gathering phase we carry out the following activities :
| a) |
Prepare for business requirements gathering |
| b) |
Gather/Validate business requirements from the Business & IT users |
| c) |
Analyze business requirements and re-prioritize, if required. |
| d) |
Identify & Analyze source systems |
| e) |
Prepare Requirement specification document |
| f) |
Assist the customer to prepare Acceptance Test Plan and Acceptance test cases |
2. Modeling Phase
The modeling phase involves following activities:
| a) |
Prepare the data architecture & source-to-target Mappings |
| b) |
Prepare the technical architecture |
| c) |
Prepare the ETL architecture |
| d) |
Prepare system test plan and test cases |
3. Design Phase
The design phase involves following activities:
| a) |
Prepare detailed design for ETL |
| b) |
Prepare detailed design for end user applications |
| c) |
Prepare Unit test plan and test cases |
4. Development Phase
The development phase involves following activities:
| a) |
Inspect DW infrastructure |
| b) |
Create Physical data model |
| c) |
Develop source codes |
| d) |
Perform unit testing |
5. Testing Phase
The following is the standard approach for testing:
| a) |
Create the Test Approach |
| b) |
Planning for the Test |
| c) |
Preparing the Test |
| d) |
Executing the Test |
| e) |
Managing the Test |
| f) |
User Acceptance Test |
| g) |
Acceptance Methodology and Criteria |
| h) |
Perform System Testing |
| i) |
Perform Acceptance Testing |
6. Implementation
| a) |
Delivery of Data Marts to the Production |
| b) |
Integration of Data Marts using Conformed Dimensions |
| c) |
Schedule Production data loads |
| d) |
Schedule Production Reports/Cubes builds |
7. Maintenance
| a) |
New Functionality / Feature implementation |
| b) |
Enhancements |
| c) |
Post live defect fixing |
These are the broad steps that we follow for any BI & Datawarehousing implementations. Of course, these are tailor made to suit specific business requirements.