The main deliverable of the build team is obviously the actual software – the data integration code, the reports and dashboards. As well as this, members of the build team are also responsible for producing the following set of deliverables.
Software Deployment Guide
The Software Deployment Guide provides detailed instructions on how to deploy the software produced by the data integration development team and BI team, essentially how to deploy the data integration code, reports, dashboards and any other delivered components. This guide typically includes instructions on how to replace the previous version and also how to uninstall the code that is being deployed in case there is a need to roll back to a previous version.
During development, issues arise particularly around design and implementation and the issue log is maintained by the build manager to track and manage these.
Once in testing phase the defect tracker is used to log and manage the defects that arise. If a spreadsheet is being used for this then it is owned and managed by the Test Manager. However, for larger projects where there are multiple developers, defect management software is used.
Data Integration Technical Specifications
The data integration technical specifications are produced by the data integration build team. The primary feature of the data integration spec is the mapping from source to target for each layer in the data warehouse. This defines what source column maps to what target column and any required data transformations.
The technical specification should refer to the standards guide when implementing common features. It should also provide guidance on unit testing.
Report, Semantic Layer and Dashboard Technical Specifications
The BI development team produces these technical specifications, one per report, dashboard or universe. For reports and dashboards, the specification defines what data items are used (cross referencing the data dictionary), the design and layout of the report or dashboard and the required user interaction either in prompts or in interacting directly with the report or dashboard.
The technical specification also contains details on implementation of non-functional requirements such as security, audit and on data volumes and performance of the reports and dashboards.
Physical Data Model
The physical data model is essentially the scripts that create the tables, views and any other database object in the database system that is hosting the data warehouse. These are typically created using data modelling software such as Erwin or Oracle SQL Developer Data Modeller and are produced by the data modelling team.
Unit Test Scripts and Results
A unit test script should define positive and negative testing and the data set to use. A separate unit test script is produced for each individual component developed for the solution. The following table lists typical components of a data warehouse
SAP BusinessObjects Universe
Module of the Security Model
|Data Integration||A unit of code that moves data from a source to a target.
A database procedure
an operating system script
Version Control Strategy
This document sets out the method by which all software will be version controlled, what version control numbering system will be used and what version control software if any will be used and how.
Configuration Item Catalogue
The Configuration Item Catalogue documents which version of each component is in each build that is produced and is necessary to make sure compatibility of the versions of the data model, data integration code and the BI content. The version control strategy is defined by the Configuration Manager prior to start of the development.
Each build that is released is accompanied by a release note that informs the testers and end users of the content of the release and includes,
- What new features are in the release,
- What features were dropped,
- What defects were fixed and
- What defects remain and any other known issues
This is also produced by the Configuration Manager.