Architecture and Design
Solution Architecture Document
The Solution Architecture Document is the key deliverable of the solution architect and describes the whole solution including the data warehouse and the BI application layer. It provides a definition for all the components that comprise the system (but does not describe how they are implemented – this is done by the technical specifications). Also documented are strategies and guiding principles for the data integration, data repositories and analytics of the solution. Finally, it reviews the software and technology used for the solution and defines requirements of the systems management strategy including configuration management, quality management, backup and restore, security, deployment and operations.
A good Solution Architecture Document will include justification for the decisions made in the document, such as, why one technology was selected over another.
Technical Architecture Document
The Technical Architecture Document is produced by the technical architect and describes in detail the technology used in the solution. It describes the required hardware and includes details of the approach to sizing the hardware. It also describes all software used including operating systems, database systems as well as the BI applications including the version of all the items and the version dependencies between them. It should also provide a roadmap for upgrades of the versions.
The technical architecture document also details data storage and archiving, network deployment architecture, backup and restore and details of high availability if required.
Infrastructure Deployment Guide
The Deployment Guide is either produced by the Technical Architect of the System Administration team. It documents in detail all the steps required to deploy the full system on an environment, from operating system installation and configuration, database installation and details of the installation and configuration of the data integration software and the BI application.
Data Integration Framework Document
The DI Framework document defines the general data integration process method used in the data warehouse. This will include how the flat files containing the source data are received processed and archived and how bad data is rejected and managed. It will also define how operational metrics are obtained and recorded for example volume of data processed, duration of data processing. It will also document the approach to data quality and validation.
Data Integration Standards, Practices and Methods Document
The standards guide ensures that all developers adopt the same approach when implementing the required data integration code. This ensures better quality of code across the project and allows each developer to understand more readily each other’s code. The document includes naming conventions, standard approach to common data integration activities, for example, how to implement type II change data capture, logging methods and standards for error management, system failover and restart.
Both the Data Integration Framework and the Standards, Practices and Methods documents are produced by the Data Integration Architect.
Data Modelling Standards, Practices and Methods Document
A standards guide for a data model defines common areas such as naming conventions, common suffixes and prefixes to use, common field lengths and types. For example, columns that are person names (customer, employee, suppliers) are all 64 chars in length. Identity columns are all suffixed “ID”. The document also provides guidance on change management, what software is used to define the data model and where the files are saved. The document is produced by the Information Architect.
Analytics Standards, Practices and Methods Document
The standards guide covers all BI application content including reports, dashboards and semantic layers (e.g. SAP BusinessObjects universes). For reports and dashboards the document defines their look-and-feel and provides guidance on naming conventions and standards for report prompts and navigation paths.
For the semantic layer the document also provides naming conventions and guidance on design issues, for example, in a SAP BusinessObjects universe the document defines when to use contexts rather than views, when to use derived tables, object naming conventions, guidance for list of values and prompts. It also defines standards on connections to the data warehouse, limits on data volumes, security considerations.
Access Layer Security Model
Security models in a BI system can quickly become complex and therefore they should be fully documented. Furthermore, this document must be updated with any security model changes as they occur. The document should clearly state the different business groups and how they access the different data areas.
The standards guide and the access layer security model documents are both produced by the Analytics Architect.