Vous êtes sur la page 1sur 4

Karlstad University

Department of Information Systems

Adapted for a textbook by Date C. J.


An Introduction to

Database Systems

Pearson Addison Wesley, 2004

Database System Architecture


Remigijus GUSTAS
Phone: +46-54 700 17 65
E-mail: Remigijus.Gustas@kau.se
http://www.cs.kau.se/~gustas/

ANSI/SPARC architecture
American National Standards Institute (ANSI)
Standards Planning And Requirements Committee (SPARC)
External level
(individual view)

Conceptual (logical) level


(community view)
Internal (implementation) level
(storage view)

Remigijus Gustas

2-2

Conceptual Level
 Global logical representation
 Shared by all users
 Consensus on underlying meaning of the data
 Foundation for database design
 Defined by conceptual schema
 Implemented via conceptual DDL



2-3

Must not involve any implementation considerations


Definitions of information content only

Remigijus Gustas

Internal Level
 Physical representation of the entire database
 Consist of many occurrences of many types of
stored records

 Defined by (yet another) the internal schema DDL


 Also specifies what indexes exist, how stored
fields are represented, what is the physical
sequence of the stored records
 Includes pointers
 Hardware dependent
Remigijus Gustas

2-4

External Level
 Supports single or groups of users
 An external view is defined by external schema
 Different users will require different views
 Most systems allow SQL to be used as a stand
alone query language or embedded in other
languages (Java, C++, PL/I)
 Implemented via SQL views
 References subsets of overall information
architecture
Remigijus Gustas

2-5

Mappings
One Conceptual/internal mapping
Defines correspondences between the
conceptual view and stored database
Several External/conceptual mappings
 Different views can overlap
Most relational systems permit
External/external mappings
 Views are mapped to other views


2-6

Remigijus Gustas

Database Administrator
 Participates in conceptual database design
 Determines how to implement conceptual schema
 Teach users, help them to define external
schemas, consulting on application design

 Implement security and integrity constraints


 Implement unload/reload (dump/restore) utilities


Very large databases for continuous availability require


careful and sophisticated administration

 Monitor and tune database performance




Reorganizing database to ensure acceptable


performance
Remigijus Gustas

2-7

DBMS versus File system


 File managers are not aware of the internal
structure of records and cannot handle requests
that rely on knowledge of record structure
 Little or no support for recovery, concurrency,
security or integrity control
 Files are typically not integrated or shared, but
instead are usually private
 File managers provide much less data
independence than DBMS does
Remigijus Gustas

2-8

Client/Server Architecture
 The purpose of DBMS is to support database
applications



2-9

Clients are various applictions that run on top of DBMS


Server is DBMS itself

Remigijus Gustas

Distributed Processing: One server


and many client machines
 Server machine can be tailored to DBMS function
 Client machine might be a personal workstation
 Several client machines are able to access the
server machine
Communication
Network

Remigijus Gustas

2-10

Distributed Processing: each machine


runs both as a client and a server


One portion of data is stored on one computer and another portion on


a different machine, occasional access to data on the other computer
1) A client might be able to access any number of servers (one at a time)
2) A client might be able to access many servers simultaneously

Case 2 is called a
Distributed DBMS

2-11

Communication
Network

Remigijus Gustas

Vous aimerez peut-être aussi