logo for information-management-architect.com
leftimage for information-management-architect.com

Architecture Specification  

Improve architecture and design process with David Bowman’s information management guidelines for architecture specification

This site is designed for information technology professionals and information management consultants who want to learn more about data warehouse and business intelligence architecture.

It provides information management guidelines for data warehouse architecture and an Architecture Specification template that can help jump-start a project.

What is an Architecture Specification?


The objective of an architecture specification is to define the architecture for the entire data warehouse and business intelligence solution including:
  • Database architecture and infrastructure;
  • Database sizing and performance expectations;
  • Identification of the extract, transform and load process;
  • Recommendations for the sizing and configuration of hardware; and
  • Access control, backup and recovery guidelines.
Architecture Guidelines

Data Architecture
  • Should complete a physical data model for data warehouse storage;
  • Should create the physical data model by translating the logical data model into a physical data model;
  • Should define partitioning strategy;
  • Should include process metadata columns as required;
  • Should include audit columns as required;
  • Should identify and document reference table codes and descriptions;
  • Should provide table definitions;
  • Should provide column definitions;
  • Should include volumetrics;
  • Should document the processes and procedures required to meet data retention requirement including processes to archive all data including landing area, staging area, data warehouse and data marts, access archived data and restore archived data;
  • Should document the processes and procedures required to meet backup and recovery requirements;
  • Should document strategy to recover data in event of a system-wide failure including the approach that will be used including planning, processes and procedures;
  • Should define the approach that will be used to replicate data to meet redundancy requirements;
  • Should define the approach that will be used to provide an alternate method of accessing data accessing data in case of a system failure; and
  • Should specify any partitioning strategies that will be used in the database design to help meet performance requirements and archiving requirements.
Data Integration Architecture
  • Should provide diagram showing the key architecture components and environments;
  • Should describe what happens in each stage of the data integration solution;
  • Should define the interface capacity;
  • Should provide a link to each interface specification for each interface specified on the data integration diagram;
  • Should specify the approach that will be used to address data migration requirements;
  • Should define the approach for ensuring change data capture including the use of change data capture software if appropriate;
  • Should define the strategy for file/table control, record control and record accounting; and
  • Should provide high level data integration module specifications for each data integration module including any master data management integration routines.
Information Delivery Architecture
  • Should complete a dimensional data model for reporting and analytics;
  • Should create facts and dimensions based on the requirements specification;
  • Should add snow-flake dimensions if appropriate;
  • Should add physical attributes such as table-space names and index names;
  • Should define partitioning strategy;
  • Should include process metadata columns as required;
  • Should include audit columns as required;
  • Should identify and document reference table codes and descriptions;
  • Should be the subject of an internal quality assurance review;
  • Should define the strategy for handling ad hoc queries;
  • Should define the specifications for each report that needs to be converted;
  • Should define specifications for all audit reporting; and
  • Should define specifications for all reconciliation reporting.
Functional Specifications

Should define specifications for any functional requirements that are not already covered.
Data Security Specifications
  • Should define the processes that will be followed to ensure data movement security;
  • Should define the users and processes that are allowed to access the data;
  • Should group users and processes into roles where possible; and
  • Should define the processes that will be followed to ensure test data security.
Data Quality Specifications
  • Should define the processes required to meet data quality requirements;
  • Should include any data checkpoints, audit trails, and metadata management;
  • Should document the type of data integrity methods that will be implemented and specify the processes and procedures performed at the database, application, and network level; and
  • Should address exception and error management and the policies and procedures that should be performed in the event of a data exception or errors.
System Specifications
  • Should define the strategy to ensure legal and policy requirements are satisfied;
  • Should define the systems that will be used in the solution;
  • Should define hardware specifications e.g. machine name, IP address, manufacturer, model, machine details, operating system storage devices, scheduled maintenance;
  • Should define software specifications e.g. software manufacturer,  software name, version, revision, and patch information and software compatibility to operating systems;
  • Should define network specifications e.g. network name available bandwidth, maximum capacity and network availability;
  • Should define the servers that will be used to host all data base environment components;
  • Should list all database environments including development, quality assurance, production and any environments required to store repositories such as data models, data integration code, business intelligence code;
  • Should specify all known facility requirements that are needed to support the project including hardware, software and other equipment;
  • Should identify the strategy to satisfy reliability requirements;
  • Should identify the strategy to satisfy availability requirements;
  • Should identify the strategy to satisfy data currency requirements;
  • Should identify the strategy to satisfy flexibility requirements;
  • Should identify the strategy to satisfy accuracy requirements;
  • Should identify the strategy to satisfy consistency requirements;
  • Should identify the strategy to satisfy maintainability requirements;
  • Should identify the strategy to satisfy scalability requirements;
  • Should identify the strategy to satisfy performance-monitoring requirements;
  • Should identify the strategy to satisfy operational control requirements;
  • Should identify the strategy to satisfy failure management requirements;
  • Should identify the strategy to satisfy logging requirements
  • Should identify the strategy to satisfy error-handling requirements;
  • Should specify the processes, hardware and software to satisfy metadata requirements; and
  • Should specify requirements for hardware/software to satisfy configuration management requirements
Tactics And Traceability
  • Should list any design assumptions;
  • Should define how each architecture component satisfies a requirement and include the requirement id and requirement statement for traceability;
  • Should define how the solution meets corporate architecture guidelines if applicable;
  • Should describe all reusable components; and
  • Should document any known issues with the design.
Architecture Specification Template

Under Construction  

Summary...

The objective of an architecture specification is to define the architecture for the entire data warehouse and business intelligence solution including:
  • Database architecture and infrastructure
  • Database sizing and performance expectations
  • Identification of the extract, transform and load process
  • Recommendations for the sizing and configuration of hardware
  • Access control, backup and recovery guidelines
This site provided information management guidelines for data warehouse architecture and an Architecture Specification template that can help jump-start a project.