Databases and DBMSs

1

IST 210 Databases and DBMSs

Todd S. BacastowJanuary 2005

2

IST 210 A Process of Mapping

Real WorldConceptualData Model

–High level model–Comprises

–“Things”–“Descriptions”–“How things are connected”

–Relational–Hierarchical–Network–Object Oriented

LogicalData Model

PhysicalData Model

3

IST 210 Data Models A data model describes

the structure of a database data types, relationships, constraintsa set of basic operations

insert, delete, modify, retrieve user-defined operations

4

IST 210 Types of Data Models

Conceptual concepts: entity, attribute, relationship Entity-Relationship model (DBMS-independent)

Logical data represented by record structure E.g. relational, network, hierarchical

Physical describes how data is stored in the disk

5

IST 210

External Level

Conceptual Level

Internal Level

DBMS Architecture

Conceptual Schema

Internal Schema

ExternalView

ExternalView

6

IST 210 External Level

Describes a part of the database for a particular user group and hides the rest

Supports multiple views of a database Same data model as the conceptual schema

External Level

ExternalView

ExternalView

Conceptual Schema

Internal Schema

ExternalView

ExternalView

7

IST 210 Conceptual Level

Data Abstraction hides unnecessary details

Conceptual Level hides physical layer

Data types, Constraints, User Operations Uses both conceptual/logical data models

Conceptual Level Conceptual Schema

Conceptual Schema

Internal Schema

ExternalView

ExternalView

8

IST 210 Internal Level Defines physical storage on the disk Defines data location

path, blocks, pages, … Device specific

STORED_EMP BYTES=20PREFIX BYTES=20, OFFSET=0EMP# BYTES=20, OFFSET=6, INDEX=EMPXDPET# BYTES=20, OFFSET=12PAY BYTES=20, OFFSET=16, ALIGN=FULLWORD

Internal Level Internal Schema

Conceptual Schema

Internal Schema

ExternalView

ExternalView

9

IST 210 DB Schema vs. DB State

Database Schema description of the database is specified during database design

Database State (extension of the schema) current state of the database: a snapshot actual data instances in a DB changes over time by update initially, a database is empty state with no data

10

IST 210 DB Schema vs. DB State

Valid State DBMS checks every state of the database does it satisfy the structure and constraints

specified in the schema? Schema Diagram

Displays database schema

11

IST 210 Example Schema

12

IST 210 Database Schema

Metadata descriptions of the schema constructs and

constraints stored in the database catalog

Schema Evolution Schema change prompted by the change of

application requirements

Designer Goal : develop a schema

that changes infrequently

13

IST 210 DBMS Mapping

Mappings for multi-level DBMS to transform a request specified at one level

into the request at another level access: external conceptual internal DB retrieve: DB internal conceptual external

Three-Schema Architecture advantage: true data independence disadvantage: overhead cost of mappings

Conceptual Schema

Internal Schema

ExternalView

ExternalView

14

IST 210 Data Independence What happens when the schema changes at

some level? Data Independence

the capacity to change the schema at one level without having to change the schema at the next higher level

Two Types of Data Independence logical and physical data independence

15

IST 210 Data Independence (con’t)

1. Logical Data Independence capacity to change the conceptual schema without

having to change the external schema when: logical reorganization of the database

2. Physical Data Independence change the internal schema without having to

change the conceptual schema when: physical reorganization of the files

16

IST 210 DBMS Languages

Data Definition Language (DDL)• to define DB conceptual schema

Data Manipulation Language (DML)• to specify database requests: update, retrieval• high-level DML: describes which data to retrieve• low-level DML: describes how to retrieve it

17

IST 210 DBMS Languages (con’t)

High-level DML: set-oriented, declarative Low-level DML: record-oriented, procedural Types of DML

• data sublangauge: DML embedded in a general purpose language (for DBAs)

• query language: high-level, interactive, stand-alone DML (casual end users)

• user-friendly interface for DML (naïve users)

18

IST 210 Database System Environment

DBMS Component Modules Managers, i.e., disk control Compiler, i.e., query Processors

19

IST 210

20

IST 210 DBMS Interfaces

Menu-based interfaces Forms-based interfaces Natural language interfaces

interpret requests to high-level queries Command line

21

IST 210 System Utilities & Tools

Loading loads existing data files into the database DBMS conversion, reformatting the data

Backup provides a backup copy of the database incremental backup: updates changes only

File Reorganization to improve performance

22

IST 210 System Utilities & Tools

Performance Monitoring monitors database usage provides statistics

Data Dictionary also called information repository stores additional information: (catalog) + design

decisions, usage standards, user information, application program descriptions

23

IST 210

Terminal

Mainframe

Storage

Logic

Presentation

Network

Mainframe/terminal

Mainframe/ terminal

Storage, Logic and Presentation all in same place

No platform specific user interface

Doesn’t take advantage of client machine

24

IST 210

Client

Server

Storage

Logic

Presentation

Network

DBMS

Client Server without stored procedures

Database server handles storage only Logic and presentation in client Takes advantage of client cpu Logic changes require client

redistribution Integrity not maintained if other DB

tool used Each user needs to be a specific

database user

25

IST 210

Client

Server

Storage

Logic

Presentation

Network

DBMS

Client Server with stored procedures

Database handles storage and business logic

Logic changed in one place, no redistribution of client

DBMS dependent code Each user needs to be

specific database user

26

IST 210

Client

Server

Storage

Logic

Presentation

Network

DBMS

Client Server with 3 tiers

Storage in database Logic in Transaction

Monitor Client does presentation

only Authentication and

Access control can be done in TP monitor

Each user does NOT have to be a database user

27

IST 210

Client

Database Servers

Storage

Logic

Presentation

Network

DBMS

Transaction Monitor

Network

DBMS

Storage

Client Server with 3 tiers

A component which sits between the client and the database server to insure reliable updates of information

Used in airline reservation and banking systems

28

IST 210 Why 3 Tiers?

Scalability multiple transaction monitors load balancing

Flexibility Complexity

update multiple data stores Two phase commit with multiple databases

29

IST 210 Classifications of DBMSs

Data Model (OO, Relational, hierarchical) Number of Users ( single vs. multi-user) Number of Database Sites ( centralized vs.

distributed vs. federated) Special-purpose vs. general-purpose