The Technical Infrastructure Domain includes all of the hardware, software, policy and standards needed as a foundation to support the Information Warehouse.
Platform refers to the hardware and operating system (OS) software that is required to support the processing requirements of the components within the Information Warehouse Domain as well as within the Business Analysis Domain.
The key platform features required by the Information Warehouse, Data Marts, Data Propagation, Data Access, Data Acquisition, and the Business Analysis Domain are discussed in the remainder of this section.
Data warehouse solutions require platforms that are:
- Readily scaleable
- Highly reliable
- Capable of efficiently processing complex queries and mass updates against large quantities of data
- Capable of keeping pace with rapidly changing technology developments and exploiting best-of-breed industry products in the areas of:
- Database management software
- Multidimensional processing
- Data analysis tools
Platform features common to the Data Warehouse, Data Mart, Data Propagation, Data Access, and Data Acquisition components of the Information Warehouse Domain include:
- Multitasking, multi-threaded, and multi-user operation
- RAID disk technology to support high levels of data protection
- Redundant hardware configurations that range from fault resilient to fault tolerant to provide high availability with limited down time
- Being on the first release tier of the Database Management System(s) (DBMS) designated as the standard for the Data Warehouse and DataMarts
- Hierarchical storage management to facilitate automatic and efficient data archive processing
- High speed tape management system for back-up, recovery, and load processing
- Integrating with systems management tools automatic alarms and alerts
- A level of security that meets standards, audit specifications, and regulatory requirements
- Automatic restart of processing from the point of fault, without the loss of data
- Support of industry standards for high speed data communication (ATM, frame relay, FDDI, fast ethernet)
- Support of high speed channels and I/O bus architecture for rapid and efficient movement of large quantities of data from disk to memory within a platform configuration, as well as between platforms and other specialized hardware.
In addition to the common features described above, the features of the platform that support the Data Warehouse also include:
- The ability to efficiently process mass updates and complex queries against large, complex databases (database size can vary from several hundred gigabytes of data to several terabytes)
- Support by a massively parallel processing (MPP) architecture to efficiently support complex database structures exceeding several hundred gigabytes (Symmetric multiprocessing (SMP) may be acceptable to support the initial Data Warehouse configuration while data volumes are relatively small.)
- A processing family that is readily scaleable with attractive price points.
In addition to the common platform features described above, the features of the platforms that support Data Mart, Data Propagation, Data Access, and Data Acquisition also include:
- Alexibility and scalability to support a wide range of configurations to satisfy the diverse business needs that exist across the enterprise (There should be a natural progression through the hardware/software family to preserve investments. Configurations for the Data Mart can range from Uni-processor to SMP and possibly MPP, depending on business needs.)
- An architecture capable of supporting automated or manual remote operation (lights-out) of platforms which include :
- Automatic power on/off
- Automatic tape stackers or robotics for automatic back-up and restoration of file structures and databases.
In addition to the common platform features described above, the features of the platforms that support the Business Analysis Domain also include:
- support the standard data warehouse client tools
- standard analysis platform configurations which satisfy the processing requirements of various classes of business analysts.
6.2. Systems Management
Systems Management is the collection of disciplines and processes for administering and managing enterprise technology and is specifically required to maintain the content and technology infrastructure of the Information Warehouse Domain. The scope of Systems Management is not, however, limited to the Data Warehousing Architecture. It is, in fact, addressed primarily by the broader Information Technology Architecture, which may be viewed as a superstructure into which Data Warehousing fits.
The sub-components of Systems Management are:
- Database Administration
- Problem Management
- Operations Management
- Version Management
- Recovery Management
- Performance Management
- Configuration Management
- Security Management
- Change Control
- Software Distribution Management
- Data Administration
6.2.1. Change Control
Change Control is the mechanism for managing the flow of changes through the application life cycle, the goal of which is to reduce defect introduction while enhancing application stability and quality.
Change Control features include:
- Processing authorization and sign-off functions needed to ensure application viability and quality
- Providing a communication vehicle, with a flexible reporting facility
- Linking or grouping components associated with a modification.
6.2.2. Configuration Management
Configuration Management refers to the process of identifying and synchronizing the hardware associated with the components of the Business Intelligence Architecture.
Configuration Management features include:
- Asset management
- Facilities to group and track hardware relationships.
6.2.3. Operations Management
Operations Management provides the policies, procedures, and tools to coordinate and manage system and network resources across the enterprise.
Operations Management features include:
- Tracking system and network resources
- Determining and coordinating operational status (e.g., power on / power off)
- An integrated, real-time, logging, reporting, tracking, and interactive management tool.
6.2.4. Performance Management
The Performance Management sub-component is used to monitor and adjust the availability and behavior of the Business Intelligence Architecture components.
Performance Management features include:
- Capturing information used in capacity planning
- Identification of saturation points (i.e., bottle-necks) and proactive tuning
- Facilities for verification of system design and service levels
- Integration of network monitoring tools with Performance Management tools for monitoring database performance in mainframe and distributed environments.
6.2.5. Problem Management
Problem Management comprises the policies, procedures, and tools for collecting, managing, tracking, diagnosing, and resolving problems in the Business Intelligence Architecture.
Problem Management features include:
- An information capture facility, serving as a single source for problem tracking and supported by flexible search capabilities
- A flexible reporting facility
- A knowledge base for recognizing patterns and resolving problems
- procedures for problem escalation and resolution
- integration of Problem Management tools with Operations Management, the Database Management System (DBMS), and Performance Management.
6.2.6. Recovery Management
Recovery Management provides the policies, procedures, and technology to restore system operation.
Recovery Management features include:
- Proceduralized recovery scenarios
- Automated detection and recovery capabilities
- Central backups used in off-site recovery and local backups used in on-site recovery.
6.2.7. Security Management
Security Management provides the policies, procedures, and technology for protecting the Business Intelligence Architecture components from unauthorized access and use.
Security Management features include:
- Database, network, application, and platform level security needed to ensure application viability and quality
- Logging of user access activity and providing user authentication
- Integration with Change Control.
6.2.8. Software Distribution Management
Software Distribution Management provides the vehicle to electronically disseminate Business Intelligence Architecture application executables to platforms throughout the enterprise.
Software Distribution Management features include:
- Automatically detecting file changes and distributing new versions via selective, stage-able push/pull capabilities
- Providing audit trails, control counts, error, and exception handling
- Providing an active inventory and software metering
- Integration with Version Management and Change Control.
6.2.9. Version Management
Version Management provides the capability to maintain multiple copies of test and production Business Intelligence Architecture applications’ and data specifications. These items include, but are not limited to, system specifications, models, and code.
Version Management features include:
- The ability to baseline specifications (e.g., code) and provide back-out and recovery capabilities needed to ensure application viability and quality
- Support for concurrent release development, providing for the check-in / check-out management of source code and other technical specifications.
6.3. Database Management System (DBMS)
The Database Management System (DBMS) is the systems software that manages data that is accessed by one or more Data Warehouse Applications or Business Analysis Domain users. Both the Corporate Data Warehouse and all of the Data Marts will use a DBMS to manage their data. Data Warehouse data is stored in a relational DBMS while Data Mart data is stored either in a relational DBMS or a multi-dimensional DBMS (MDDBMS).
DBMS features include:
- Data integrity through:
- Referential and domain integrity enforcement
- A flexible locking strategy (row / page)
- Multi-level security down to the column level
- DBMS-imbedded column edits.
- A comprehensive user interface and data manipulation capability which including:
- Logical view facilities that insulate the user from physical data constructs
- Interfaces to the Metadata Repository, standard languages, and data modeling tools
- Date and time functions, complex data types, and null values (multi-valued logic)
- Millennium awareness (year 2000 and date calculations)
- An access path optimizer (static and dynamic bind)
- Compliance with the ANSI SQL 92 standards, including join, outer join, and aggregation capabilities
- Program and device independence
- Fully independent explicit data partitioning
- Increased throughput and performance through:
- Indexing strategies (binary tree, bit map, hashing, and join)
- Parallel execution (MPP and SMP)
- I/O independence (asynchronous read and write and caching management)
- Scalability for a large number of active users (greater than 500), large data volumes (terabyte or greater), and aggregate I/O
- Management tools for:
- Performance monitoring and tuning
- Data management (load, reorganization, etc.)
- Capacity planning
- Log management
- Job/usage accounting sufficient to support charge back
- Managing resources through:
- Dynamic internal configuration
- Data compression
- Efficient memory management
- A query governor.
- Distributed database features including:
- Triggers and stored procedures
- Data replication
- Location transparency with directory services
- Multidimensional database (MDDBMS) features including:
- A high maximum number (10 – 12) of dimensions without performance degradation and with large amounts of data
- Implementation of star schema models
- Drill down / drill up (and reach through) capability.
6.3.1. DBMS Software
6.3.2. DBMS Utilities and Tools
6.3.3. Database Administration
Database Administration provides the policies, procedures, and technology to manage the physical databases of the enterprise.
Database Administration features include:
- Definition and management of physical databases that are derived from logical models
- Direct generation of physical specifications from logical models in the Metadata Repository
- Mandatory synchronization between logical and physical models
- Monitoring, tuning, and maintenance of the physical databases.
The Network facilitates connectivity between platforms and enables the transmission of data, voice, image, and video between locations to support business needs. The Data Warehouse Initiative requires network support in the form of LAN, WAN, and remote/mobile connectivity.
Network features capable of supporting the Information Warehouse Domain and the Business Analysis Domain include:
- 7 day by 24 hour (7×24) operational availability
- Capacity in the form of:
- Megabyte to terabyte file transport between sites
- Moderate transaction traffic that can have potentially large result sets
- Sustaining fast transfer rates for sustained periods of time without any degradation of quality or throughput
- Dynamic bandwidth allocation to support burst transmission of large volumes of data between:
- Client tools and the Data Warehouse and/or Data Marts
- Data Warehouse file transfers and one or more Data Marts
- External / Operational Environment data and the Information Warehouse Domain
- Data Warehouse and Archived Data Storage
- Ubiquitous, performance oriented access from either a direct or remote connection
- Secure transportation of transaction data and files
- Delivery of a consistent level of performance
- Error free transmission of data or files.