Part V: Data Administration
Chapter 13: Data and Database Administration
Chapter 14: Data Warehouse
Chapter 13: Data and Database Administration
Data Administrator
Database Administrator
Modeling and Planning Enterprise Data
Data Security
Backup
Concurrent Access
Data Quality
Data Administrator
Custodian of the organization’s dataDevelops procedures to protect and control data
resourceMaintains corporate-wide data definitions and
standardsFunctional responsibilities: planning, analysis, design,
implementation, maintenance, protectionShould be a respected, senior-level middle managerRequires a high level of managerial and technical skills
Database Administrator
Responsible for physical database design and technical issues
Carries through the standards and procedures established by data administrator
Broad technical backgroundManagerial skills are also critical
Corporate Data Management
Functions:data policies, procedures, and standardsplanningdata conflict resolutioninternal marketingmanaging the data repositoryselection of hardware and softwaremanaging data security, privacy, and integritydata backup and recovery
Modeling Enterprise Data
Enterprise architecture: “that set of descriptive representations that are relevant for describing an enterprise such that it can be produced to management’s requirements and maintained over the period of its useful life.”
The conceptual data model must mesh with …business process modelbusiness network logisticswork flow modelsenterprise master scheduleenterprise business plan
Planning for Databases
What data will be included?What processes will the system perform?What business locations will the system
affect?What work units and people will be affected
by the system?How will the system fit with significant
business effects?
Managing Data Security
Threats to data securityaccidental: human, software, hardware errortheft and fraudloss of privacy or confidentialityloss of data integrityloss of availability
Data security planlimit access (4 methods)conceal datafacilitate recovery
Security: Views
Presents only data to which a user needs access
Always shows current dataMore sophisticated security measures are
normally required
Security: Authorization Rules
Restrict access to dataRestrict actions that may be takenAuthorization tables for subjects (users)Authorization tables for objects (data)
Security: User-Defined Procedures
For example, a series of questions
Security: Encryption
Security: Authentication Schemes
Database Recovery
Backup facilitiesstatic data … dynamic data … incremental backup
Journalizing facilitiestransaction log … before-image … after-image
Checkpoint facilityprovides known state from which recovery can be
done
Recovery manager
Recovery and Restart Procedures
Switch (mirrored database)Restore/Rerun: reprocess transactionsTransaction integrity
commit or abortbackward recovery – rollback – undo transactionsforward recovery -- rollforward – apply
transactions
Concurrent Access
Pessimistic: lockingOptimistic: versioningReading
no concurrency issues
Writingincompatible updatesinaccurate data read by another
Locking Mechanisms
Locking level (granularity)database … table … block/page … record … fieldshared locksexclusive locks
Deadlocksprevention by lock all at starttwo-phase: lock all, then releaseresolution: backout transactions
Versioning
Only one user may make changesOther users restricted from changing, must
restart
Managing Data Quality
Security policy and disaster recoveryPersonnel controlsPhysical access controlsMaintenance controlsData protection and privacy
Chapter 14: Data Warehouse
Data Warehouse Architecture
Data Characteristics
Reconciled Data Layer
Data Transformation
Derived Data Layer
User Interface
Two Types of Processing
Operational processingcaptures, stores, and manipulates data to support daily
operations
Informational processinganalysis of summarized data to support decision
making
Data warehouses consolidate and integrate information from many
different sources and arrange it in a meaningful format for making accurate business decisions
Data Warehouse Architectures
Two-level architecturesource files …
read-only data warehouse
Three-level architecturesource files …
enterprise data warehouse (reconciled data) … data mart (derived data)
Data Characteristics
Status vs. event databalance vs. transaction
Transient vs. periodic datacurrent balance vs. balance history
Reconciled Data Layer
Used to build data warehouseCharacteristics of reconciled data
detailedhistoricalnormalizedcomprehensivequality controlled
Capture Scrub Transform Load
Derived Data Layer
Data that have been selected, formatted, and aggregated
Star schemadimension table: “demographic data”fact table: “associative data” – often summarized
User Interface
Traditional query and reporting toolsOn-line analytical processing (OLAP)
cube slicingdrill-down
Data miningexplanatoryconfirmatoryexploratory
Data visualization