IST 318 Database Administration Lecture 1 What Is DBA?

Post on 18-Jan-2016

223 views 2 download

Tags:

Transcript of IST 318 Database Administration Lecture 1 What Is DBA?

IST 318 Database Administration

Lecture 1

What Is DBA?

Topics

DBMS and DB ApplicationsDatabase, Data, and System AdministrationDBA TasksTypes of DBAsTest and Production Environments

DB Applications

Data is the lifeblood of computerized applications In many ways, business today is data

Using DBMS is the efficient way for data persistence and manipulation

DBA is at the center of the development lifecycle

Database vs. DBMS

DatabaseAn organized store

of data Data can be

accessed by names

DBMSSoftware that

enable users or programmers to share and manage data

Enterprise IT Infrastructure – a big picture

Application Development Lifecycle

Data Administration

Concentrate on the business aspects Business lexicon logical data modelRequirements gathering, analysis, and design

Typical tasks include Identifying and cataloging business dataProducing conceptual and logical data modelsCreating enterprise data modelSetting data policies and standards

Concerns more about metadata

Metadata

Metadata is often described as data of dataDefinitionBusiness nameAbbreviationType and length/accuracyDomain, or range of valid values

Data Models – three levels

Conceptual modelOutlines data requirements at a very high levelDescribes data mostly in business context

Logical modelProvides in-depth details of data types, lengths,

relationships, and cardinality

Physical modelDefines the way data is organized in physical medium

DBA vs. DA

DBAs have to take care the first two levels if no DA roles are implemented in an organization

System Administration

SAs are more concerned about the installation and setup of DBMS

Typical SA tasks includeUnderlying OS systemsDBMS Installation, modification, and supportSystem configurations enabling DBMS to work with

other software systems

DBA Tasks

Ensuring data and databases are useful, usable, available, and correct

Typical DBA tasks includeDB design and implementationPerformance monitoring and tuningAvailabilityDB security and authorizationBackup and recoveryData integrityDBMS release migration

DB Design & Implementation

Understand and adhere to sound relational design principlesRelational theories and ER diagramsDBMS specifics

Understanding conceptual/logical models and being able to transform to physical DB implementationPoor design can result in poor performance

Performance Monitoring & Tuning

Performance = the rate at which the DBMS supplies info to its users

Influenced by five factorsWorkloadThroughputResourcesOptimizationContention

Availability

Multifaceted processKeep the DBMS up and runningMinimize the downtime required for admin tasks

Technologies and up-front planning can help

DB Security & Authorization

Ensure data is available only to authorized users by granting privileges to different (groups of) users

Actions need to be controlledCreating/altering DB objects and/or their structuresReading/modifying data from tablesStarting/stopping DB and/or associated objectsRunning stored procedures or DB utilities

Backup & Recovery

Be prepared to recover DB in the event of Improper shutdown of DB applications, due to

Software errorHuman errorHardware failure

Types of recoveryRecover to currentPoint-in-time recoveryTransaction recovery

Data Integrity

Store the correct data in the correct wayPhysical integrity

domains and data types

Semantic integrityquality data with no redundancy

Internal integrity internal structures and code

Desired Skill Set

SQL + PL/SQLSystem specific operations and practicesData modeling methodologies and toolsNetworking (client/server)O/SProgramming (conventional and web-oriented)Transactional/messaging systems

Types of DBAs

System DBADB architectDB analystData modelerApplication DBATask-oriented DBAData warehouse administrator

Test & Production

At least two separate environments must be created for quality DB implementationTesting (aka development) QA (aka staging)Production

DifferencesThey should share the same configurationThey don’t need to be identical

Testing DB may have only a subset of data

Multiple DB Environments