B406 Information Systems

Distributed Systems

See also Whitten et al. (2004), chapter 13

Distributed Systems

This Week

Distributed vs. centralized systems

Client / Server and network architectures
Distributed databases
Distributed Systems

Early Computers

- In the early days,

expensive and monolithic
mainframe computers
- The internet / networks
have been developed
through military drivers
- Distributed control over
machinery in case of
nuclear attacks
Distributed Systems

Control Rooms
Centralized
controls
Complex
interfaces
Very complicate
Distributed Systems

A Today’s office ?!

Then came desktop workplaces

(with Graphical User Interfaces)
- In working environment usually
come with central file storage or
at least central backup
- Central databases
(for employees, customers,
products, finance, …)
- Still dominant today
- Often not used consistently
Æ employers get problems with
data inconsistencies or loose
data
Distributed Systems

Internet

The internet virtually

connects all
computers
- More and more
mobile devices,
Smartphones,
PDA’s, …
Distributed Systems

Today’s Mainframes

• Centralized services
- Data storage
- Data indexing
(google & co.)
- Communication
(email, chat, sykpe)
- Rich Internet
Applications (RIA)
Distributed Systems

Mobile Computing

- Aim is to work
anywhere, anytime
- Home based work
- During travel,
in hotel, …
Distributed Systems

Centralized Systems

• Centralized system comprises a central, multi-use

computer
• Hosts DATA, PROCESS & INTERFACE
components of the information system
• Users interact with the host machine via terminals
(or PCs pretending)
• Virtually all processing performed on the host
– terminals can be weak (i.e. inexpensive)
• Sounds OK, so why distribute the information
system?
Distributed Systems

Why the Trend Toward Distributed Systems?

• Modern businesses are already decentralized (distributed).

• Distributed computing moves information and services
closer to the customers and users who need them.
• Distributed computing consolidates the power of personal
computers across the enterprise.
• Distributed computing solutions are in general more user-
friendly because they use the PC as the user interface
processor.
• Personal computers and network servers are less
expensive than mainframe computers
– Though total cost of ownership is at least as expensive
Distributed Systems

Distribution Cons

• Lot of network traffic involved, which can compromise

performance (without potentially expensive upgrade)
• Security and integrity more easily compromised
– Easier to 'get into' isolated sources of information (unless one goes
for the 'jackpot' of central system)
– Harder to ensure that the system is kept up-to-date
• Nevertheless, the practical advantages of distributed
systems maintain their popularity

Î Many centralized systems are transformed into

distributed systems.
Distributed Systems

The Big Picture

3 architectures
5 layers

[Whitten et al., p.511]

Distributed Systems

Computing Layers

• Presentation layer — the user interface; presentation of

inputs and outputs to the user.
• Presentation logic layer — processing that must be done
to generate the presentation, such as editing input data or
formatting output data.
• Application logic layer — the logic and processing to
support business application (rules, policies, and
procedures)
• Data manipulation layer — all the commands and logic to
store and retrieve data to and from the database
• Data layer—the actual business data (databases)
Distributed Systems

Distributed Architecture Types

• Three main architecture types

– File Server

– Client / Server (2-tier)

– Client / Server (3-tier)

– Client / Server (N-tier)

– Network (N-tier)
Distributed Systems

File Server Architecture

Local area network (LAN) – a set of client computers (PCs)

connected over a relatively short distance to one or more
servers.
File server system – a LAN in which a server hosts the data
of an information system.
– All other layers are implemented on the client computers
– Frequently excessive network traffic to transport data between
servers and clients
– Client must be fairly robust (“fat”) because it does most of the work
– Database integrity can be compromised
(file locked during edit by a single user)
Distributed Systems

File Server Architecture

Distributed Systems

Client/Server Architecture

Client/server system – a distributed computing

solution in which the presentation, presentation
logic, application logic, data manipulation, and
data layers are distributed between client PCs and
one or more servers.

“Currently the dominant model for

distributed computing at current”
Distributed Systems

Client/Server Architecture
Clients
Thin client – a personal computer that does not
have to be very powerful because it only presents
the user interface to the user.
– Largely for interaction with processing layers
– Increasingly PDA’s, Handhelds or Smart phones
Fat client – a typically powerful personal computer
– capable of independent application processes
– Also notebook computer or workstation
Distributed Systems

Client/Server Architecture
Servers
• DATA, PROCESS & INTERFACE layers
distributed across client PCs and one, or more,
servers (typically more powerful / capable than file
server)
Distributed Systems

Client / Server Architecture

Different types of server
• Database server – a server that hosts one or more
databases. (Executing all data manipulation
commands at the server)

• Transaction server – a server that hosts services which

ensure that all database updates for a transaction succeed
or fail as a whole. (Crucial in banking contexts – cash
machines; online shopping …)

• Application server – a server that hosts application logic

and services for an information system. (between interface
& database)
Distributed Systems

Client / Server Architecture

Different types of server
• Messaging or groupware server – a server that hosts
services for e-mail, calendaring, and other work group
functionality.

• Web server – a server that hosts Internet or intranet

websites. (sends data (XML) and documents (HTML) to
clients)
Distributed Systems

Distribution Types

• Various layers = various types of distribution

• Distributed Presentation

• Distributed Data (2-tier)

• Distributed Data & Application (3(+)-tier)

Distributed Systems

Client/Server Architecture
Distributed Presentation
• Distributed Presentation
– Presentation & presentation logic layers reside on client
– Rest of the system could be an existing legacy
database
• Good, because
– Rapid implementation
– Fast, familiar interface to existing database
• Bad because
– System functionality can’t be significantly improved
– Doesn’t maximise client PC potential
Distributed Systems

Client / Server - Distributed Presentation

Distributed Systems

Client/Server Architecture (2-tier)

Distributed Data
Distributed data – a client/server system in which
the data and data manipulation layers are placed
on the server, and other layers are placed on the
client.
– Sometimes called two-tiered client/server computing
– Difference to file server systems is where the data
manipulation commands are executed
– Much less network traffic than file server systems
because only the database requests and the results of
those requests are transported across the network.
– Database integrity is easier to maintain.
Distributed Systems
Distributed Systems

Client/Server Architecture (2-tier)

Distributed Data
• Good, because
- Less traffic (only requests & required records are
transported)
- Integrity is easier to maintain – only records in use need
be locked
• Bad, because
- Requires ‘fat’ clients
- Application logic must be duplicated and maintained
across all clients
Distributed Systems

Problem with 2-tier client/server systems

• With larger number of clients, performance

problems occur (mainly related to application
logic)
• In transaction processing systems (or online
application processing (OLAP)), transactions must
be managed centrally to ensure processing as
single unit
– Example: 2 people want to withdraw money from a
bank account at the same time…
Distributed Systems

Client / Server Architecture (3-tier)

Distributed Data and Application
• Designed to avoid inefficiency of maintaining
application logic on client PCs
• Multi-user transaction processes require careful
monitoring to maintain integrity
• Data and data manipulation layers reside on their
own server(s)
• Application logic resides on its own server
• Presentation logic and presentation reside on the
client PC
Distributed Systems

Client / Server Architecture (3-tier)

• Good, because
– Only application server need be maintained to ensure
integrity (beware stored procedures embedded in
database)
– Simplified, stable client configuration & maintenance
– Greater flexibility
• Bad, because
– Complexity of design / how best to partition the system?
– Available prototypes?
– Again, potential withdrawn from client PCs
Distributed Systems

3-tier system

Data

Business logic /
process /
application

Interface /
presentation
Distributed Systems

Architecture & Layers

• 3-tier system can be decomposed into 'layers'

PRESENTATION
DATA LOGIC LAYER
PRESENTATION USER
FORMAT (GENERATE
LAYER INTERFACE
& EDIT PRESENTATION)

APPLICATION DATA (PROCESSING REQUIRED TO

LOGIC LAYER ANALYSIS SUPPORT APPLICATION)

DATA
STORE & DATA LAYER
MANIPULATION DATABASE
RETRIEVE
LAYER
Distributed Systems

Distributed Data
& Application

• User
• Client
• Application
Server
• Database
Server
Distributed Systems

Internet- and Intranet-based Architectures

(N-tier)
Network computing system – a multi-tiered
solution in which the presentation and
presentation logic layers are implemented in
client-side Web browsers using content
downloaded from a Web server
– presentation logic layer connects to the application logic
layer that runs on the application server, which
connects to the database servers on the backside of the
system
Distributed Systems

Network Architecture
for Information Systems
The greatest potential of this approach is its
applicability to redesign of traditional
information systems to run on an intranet

Intranet – a secure network that uses Internet

technology to integrate desktop, work group, and
enterprise computing into a cohesive framework.
Distributed Systems

Network Architecture (N-tier)

• Presentation and presentation logic reside in web

browsers
• Web server provides browser content and user
authentication (security)
• Application and data layers can be as before (or
whatever configuration best performs the task)
• Intranet = secure, internal network
• Internet = should be familiar!!
Distributed Systems

Network Architecture (N-tier)

• Good, because
– Unified, easy-to-interpret presentation interface
– Enormous flexibility
– Java, etc. allows platform independence
– Unified content structure (HTML)
• Bad, because
– Technologies involved are undergoing rapid
development
– Any others? Security? Pace of working life?
Distributed Systems

Network Computing
System:
Internet/Intranet
• User
• Client
• Web Server
• Application Server
• Database Server
Distributed Systems

One more Architecture

Peer-to-peer networks (P2P)
• network that relies on the computing power and
bandwidth of the participants in the network
rather than concentrating it in a relatively few
servers
• P2P networks are typically used for connecting
nodes via largely ad hoc connections
• Such networks are useful for many purposes
– Sharing content files (see file sharing) containing audio,
video, data or anything in digital format is very common
– and realtime data, such as telephony traffic, is also
passed using P2P technology
http://en.wikipedia.org/wiki/Peer-to-peer
Distributed Systems

Peer-to-peer networks (P2P)

• A pure peer-to-peer network does not have the

notion of clients or servers
• only equal peer nodes that simultaneously
function as both "clients" and "servers" to the
other nodes on the network This model of network
arrangement differs from the client-server model
where communication is usually to and from a
central server
Distributed Systems

Advantages of peer-to-peer networks

• Efficiency - all clients provide resources,

including bandwidth, storage space, and
computing power
• Scalability - as nodes arrive and demand on the
system increases, the total capacity of the system
also increases
• Robustness - in case of failures, data is
replicated over multiple peers
• If true P2P, there is no single point of failure in the
system
Distributed Systems

Relational Databases

• Already met relational databases

• Direct analogy to E-R modelling
• Consist of tables (entities) & relationships
• Table rows = records, columns = fields
• Can be distributed
as with systems
Distributed Systems

A database example
Distributed Systems

Distributed Relational Databases

• A distributed relational database management

system (distributed RDBMS) allows sophisticated
access to and maintenance of an RDB that is
distributed across a network
• Provides backup, recovery and security
• Sometimes called client/server database
management system
• In a distributed RDBMS, the underlying database
engine that processes all database commands
executes on the database server (less data traffic)
Distributed Systems

Distributed Relational Databases

Two types of distributed data
• Data partitioning
– Little duplication between servers
– Vertical partitioning: Different columns
assigned to different database servers
– Horizontal partitioning: Different rows in a
table allocated to different database servers
Distributed Systems

Distributed Relational Databases

Two types of distributed data
• Data replication
– Duplicates some or all tables (rows and columns)
on more than one database servers
– Can duplicate entire tables or subsets on different
servers
• Google indexing (popular search queries vs. rare queries)
• Amazon (popular products vs. rare products)
– Management system with replication technology
controls access to and management of each database
server and also schedules updates between database
servers with duplicated data