Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 13

Further Topics in Database

Technology

CMSC 127
Current Trends in DB Technology:
Centralized vs. Distributed DBs

● Centralized Databases - all data at single computer


site.
 Advantage: greater control of database
 Disadvantage: inaccessible to remote users, high
communication costs, if central server down -- all
down
Current Trends in DB Technology:
Centralized vs. Distributed DBs
● Current Trend --> towards distributing data.
 data at multiple sites interconnected via
communications networks
 data may be in heterogeneous platforms/networks
● Advantage: increased reliability/availability of data,
local control, modular growth, reduced communication
costs, improved performance (faster response to
requests for data)
● Disadvantage: costly and complex software, high
processing overhead, difficult to maintain data integrity,
slow response to requests if data not distributed
properly.
DISTRIBUTED DATABASES
To run distributed database, need Distributed DBMS
(coordinates data access to nodes)
1. checks if transaction is local (can be processed at
local site) or global (need access to data at nonlocal
site)
2. if global -- check data directory and route to nonlocal
site as necessary.
DISTRIBUTED DATABASES
HOW IS DATA DISTRIBUTED?
● data replication
● partitioning
 horizontal partitioning
 vertical partitioning
DISTRIBUTED DATABASES
WAYS TO CONNECT
● LAN (local area network) -- if systems are in
close proximity
● Long Haul Networks - distributed over large
distances (use telephone lines/satellites)
● may use different network topologies (i.e fully
connected, partially connected, star, ring, tree).
DISTRIBUTED DATABASES
BASIC RULE OF DIST’D DATABASES -->
transparency (system appears non-distributed to
users)
1. location transparency - user is not aware of site of
data being accessed.
2. replication transparency - user is not aware that
more than 1 copy of data exists.
3. failure transparency - all actions of a transaction
are completed at each site or none are committed.
4. concurrency transparency - allows multiple users
to use dbms at same time and not worry about
updating the same data at same time.
● The concept of Client/Server Architecture evolved out of
distributed database development.
CLIENT/SERVER
ARCHITECTURE
● Data centralized on 1 computer (the server) and each
user’s computer (the client) connected to the server via
a network.
● Concept is to distribute users and centralize data.
● Usually used to network personal computers &
workstations, and replace old mainframe appns
(downsizing)
CLIENT/SERVER
ARCHITECTURE
● Advantage: greater flexibility to develop applications in
convenient environments, more sensible distribution of
duties to form a cooperative computing situation, makes
use of desktop PC’s already existing, GUIs -- easier to
use, replace expensive mainframes with less expensive
hardware (downsizing)
● Components: local area networks, database servers,
application programming interfaces, application
development tools.
Object Oriented Data Modeling

● Centers around objects and classes


● Involves inheritance
● Encapsulates both data and behavior
● Benefits of Object-Oriented Modeling
 Ability to tackle challenging problems
 Improved communication between users, analysts,
designer, and programmers
 Increased consistency in analysis and design
 Explicit representation of commonality among system
components
 System robustness
 Reusability of analysis, design, and programming
results
Figure 14-1 –
Phases of object-oriented systems development cycle
OO vs. EER Data Modeling
Object Oriented EER
Class Entity type
Object Entity instance
Association Relationship
Inheritance of attributes Inheritance of attributes
Inheritance of behavior No representation of
behavior

Object-oriented modeling is frequently accomplished using the


Unified Modeling Language (UML)
Summary
● We have discussed the following topics:
 Distributed databases
 Client-Server Architecture
 Object-Oriented data modeling

You might also like