Transcript Slides - Ken Cosh

261446
Information Systems
Dr. Kenneth Cosh
Lecture 4
Review

Hardware

Input, Output devices, Processors, Memory
Client/Server Networking


The Micro computer is called the client, while
midrange computers are often servers.
Some processing is performed on the server, and
some on the client;

Thin-client model


In a thin-client model, all of the application processing and
data management is carried out on the server. The client is
simply responsible for running the presentation software.
Fat-client model
 In this model, the server is only responsible for data
management. The software on the client implements the
application logic and the interactions with the system user.
Thin and Fat Clients
Presentation
Thin-client
model
Data management
Application
processing
Client
Presentation
Application processing
Fat-client
model
Client
Server
Server
Data
management
Peer 2 Peer (P2P)


In a fat client model where all the processing and
data is stored on the client, a P2P network can
emerge, where servers are removed and clients
communicate directly with each other.
Grid Computing, still being researched and
developed, but an approach where the processing
power of any machine on the network can be used
and shared by others.
Hardware Trends

Convergence of Hardware & Telecommunications



Nanotechnology



It’s all getting smaller
& Mobile
Edge Computing


Telephones with cameras, television, browser etc.
Internet telephony, Skype
Load balancing across web servers
Autonomic Computing

Systems which can configure and optimise themselves
Languages




Computers only understand ‘0’s
and ‘1’s.
Programming with only ‘0’s and
‘1’s would be very boring and
very error prone.
Low level programming
languages allow us to translate
some basic instructions into a
more readable english code;
 add x y z
High level programming
languages allow us to use a
larger subset of language with a
tight syntax and semantics
Software Trends

Less concern with machine efficiency



Cost per instruction is falling, but personnel costs continue
to rise.
Hence more concerned with human efficiency than
machine efficiency
 Tools to support computer professional efficiency (query
languages, OOP, CASE)
 Tools to support executives (voice recognition, Natural
language interfaces)
More OOP

Quicker development & more attractive applications.
Software Trends

More Purchased Applications?



More User Development?



Quicker implementation
Less organisational re-engineering
4GLs allow anyone to code
Easy one time customisations
More Web based applications

Available everywhere
This Weeks Topics

The ‘Data’ Resource


Organising Data
Databases
Organisational Obstacles

Implementing new data models requires reexamining the role of data within an organisation,




Who has access to what data, and when?
Changing the allocation (or sharing) of data can impact on
current power relationships, and so is often met by political
resistance.
Traditionally data was stored in file format, with each
department having a selection of files.
More recently databases and DBMS allow data to
be shared across multiple departments
So What’s the Problem?

Systems within systems (subsystems),
interfacing systems and adaptive systems



Each system tends to grow and adapt
independently.
Functional units develop systems isolated from
other units.
Each functional unit develops many databases;
personnel has personnel, payroll, medical
insurance, pensions, mailing file….
Problems

Data Redundancy and Confusion




Duplicate Data in multiple data files.
The same data can have different names,
different meanings, different related data in
different places.
The same name might be used for different data
in different places.
Database confusion makes implementing a SCM,
CRM or Enterprise wide system difficult.
Problems 2

Program-Data Dependence




There is a tight relationship between the data in files and
the programs using them.
Any changes to the data, results in necessary changes to
the programs that use the data.
Maintaining data becomes costly.
Lack of Flexibility


Scheduled reports can easily be generated from the data.
Ad Hoc reports however are costly to generate. While the
information is somewhere in the system getting it out is
tricky.
Problems 3

Poor Security



Or poor control.
There is now a lot of data in a lot of databases throughout
the organisation. It is difficult to control or manage the data
– who is accessing what data?
Lack of Data Sharing & Availability


With poor control over data, its difficult to share data
between functions.
Accounts might benefit from some data that manufacturing
has, etc.
DBMS




The DBMS sits between the actual data and the
applications which use the data.
This saves the user from needing to understand the
actual physical way the data is stored, instead
presenting a logical view of it.
The user doesn’t need to know the data definition
language, but instead could use a data manipulation
language such as SQL.
In reality often the manipulation language is hidden
within an application.
DBMS
Data Definition
Data Manipulation
Creating & Changing the
logical structure of a
database
Querying & making
changes to the information
Database
Application
Generation
Data Administration
Menus, data entry
screens, reports and
application software
Who can see what
information; methods
for backup and
recovery
Hierarchical Database
ROOT
FIRST
CHILD
Performance
Ratings
SECOND
CHILD
Employee
Compensation
Job Assignment
Benefits
Salary History
Pension
History
Life Insurance
Health
Hierarchical Data


Suppose from the previous data structure, we
wanted to access the salary history for all
people with the job title “Assistant”, accessing
that data would not be easy.
While certain scheduled reports can be
generated, ad hoc reports are not as flexible.
Relational Databases


Data is organised into tables, which could be
visualised as a spreadsheet. In each table
data is organised into rows / records (or
tuples).
Any piece of data from any table can be
linked to any piece of data in another table,
so long as they have a common data element
(field).
Object-Oriented DB



Hierarchical and Relational databases assume that
data is in character or numerical form.
Some databases store data which can’t easily be
represented in files and tables (such as graphics,
sounds, java applets or any other multimedia).
O-O databases are designed to deal with these
diverse data types, however they tend to be a lot
slower than relational databases.
Data Warehouse



Logical collection of information gathered
from many different operational databases.
Used to create business intelligence, assist
with analysis and decision making.
Multi-dimensional ‘hypercube’ of information.
Data Mining




Query and Reporting Tools
Intelligent Agents
Multidimensional Analysis Tools
Statistical Tools