Chapter 13 Computer Network Design

Adopted from
Dr. Mohammed H. Sqalli COE, KFUPM
Spring 2012

Introduction
What is a Network?
What is Network Design?
Top-Down Network Design
Network Development Life Cycle (NDLC)
Network Analysis and Design Methodology
Types of Network Design
And Then What?

CSE-550-T112

Lecture Notes - 1

What is a Network?
Management view
Technical view

CSE-550-T112

Lecture Notes - 1

The Management View (1/3)

A network is a utility
Computers and their users are customers of the network
utility
The network must accommodate the needs of

customers

As computer usage increases so does the requirements of

the network utility

Resources will be used to manage the network

The Network Utility is NOT free!
Someone must pay the cost of installing and maintaining
the network
Manpower is required to support the network utility
CSE-550-T112

Lecture Notes - 1

The Management View (2/3)

Utilities dont bring money into the organization
Expense item to the Corporation
Cannot justify Network based on Productivity
Improvements
As a network designer, you need to explain to

management how the network design, even with the

high expense, can save money or improve the
companys business

If users cannot log on to your commerce site, they will try

your competitor, and you have lost sales
If you cannot get the information your customers are
asking about due to a network that is down, they may go
to your competitor

CSE-550-T112

Lecture Notes - 1

The Management View (3/3)

You need to understand how the network assists the

company in making money and play on that strength

when you are developing the network design proposal
Try to show a direct correlation between the network

design project and the companys business

Because you want a faster network is not good enough,

the question that management sends back is WHY DO I
NEED A FASTER ONE?

CSE-550-T112

Lecture Notes - 1

The Technical View (1/2)

A Network really can be thought of as three parts and they

all need to be considered when working on a network design

project:

Connections
Communications/Protocols
Services

Connections
Provided by Hardware that ties things together
Wire/Fiber/Wireless Transport Mechanisms
Routers
Switches/Hubs
Computers
CSE-550-T112

Lecture Notes - 1

The Technical View (2/2)

Communications/Protocols
Provided by Software
A common language for 2 systems to communicate with each other

TCP/IP (Internet/Windows NT)

IPX / SPX (Novell Netware 4)
AppleTalk
Other Network OS

Services
The Heart of Networking
Cooperation between 2 or more systems to perform some function Applications

CSE-550-T112

telnet
FTP
HTTP
SMTP
Lecture Notes - 1

Traditional Network Design

Based on a set of general rules
80/20
Bridge when you can, route when you must
Cant deal with scalability & complexity
Focused on capacity planning

Throw more bandwidth at the problem

No consideration to delay optimization
No guarantee of service quality
Less importance given to network RMA (Reliability,
Maintainability, and Availability) compared to throughput

CSE-550-T112

Lecture Notes - 1

Maintainability
Mean time between failures (MTBF)

average time a device or system will operate

before it fails
Mean time to repair (MTTR) average time
necessary to repair a failure within the
computer system

Data Communications and

Computer Networks: A
Business User's Approach,

10

Reliability
Reliability is defined by the equation:
R(t) = e -bt

In which: b = 1/MTBF
t = the time interval of the operation
e = Eulers number = 2.71828
MTBF = Mean Time Between Failures

Data Communications and

Computer Networks: A
Business User's Approach,

11

Reliability
What is the reliability of a modem if the MTBF

is 3000 hours and a transaction takes 20

minutes, or 1/3 of an hour (0.333 hours):
R(t) = e -bt
b = 1/MTBF
R(0.333 hours) = e -(1/3000)*(0.333)
= e -0.000111
= 0.99989
Data Communications and
Computer Networks: A
Business User's Approach,

12

Availability
Availability probability that a particular

component or system will be available during

a fixed time period
Availability % =
(Total available time Downtime) / Total available
time

Data Communications and

Computer Networks: A
Business User's Approach,

13

Availability
Suppose we want to calculate the availability

of a modem for one month (24 hours per day

for 30 days, or 720 hours), knowing the
modem will be down for two hours during that
period:

Availability % = (720 2) / 720

= 0.997

Data Communications and

Computer Networks: A
Business User's Approach,

14

Application Characteristics
Applications

Message
Length

Message
arrival rate

Delay need

Reliability
need

Interactive
terminals

Short

Low

Moderate

Very high

File transfer

Very long

Very low

Very low

Very high

Hi-resolution
graphics

Very long

Low to
moderate

High

Low

Packetized
voice

Very short

Very high

High

Low

CSE-550-T112

Lecture Notes - 1

15

Application Bandwidths
Transaction
Processing

100 Bytes

Few Kbps

Word Processing

100s Kbps

Few Mbps

File Transfers

Few Mbps

10s Mbps

Real-Time Imaging

10s Mbps

100s Mbps

CSE-550-T112

Lecture Notes - 1

16

A Look on Multimedia Networking

Video standard Bandwidth per
user

WAN services

Digital video
interactive

1.2 Mbps

DS1 lines ISDN

H11, Frame
Relay, ATM

Motion JPEG

10 to 240 Mbps

ATM 155 or 622

Mbps

MPEG-1

1.5 Mbps

DS1 lines ISDN

H11, Frame
Relay, ATM

MPEG-2

4~6 Mbps

DS2, DS3, ATM at

DS3 rate

CSE-550-T112

Lecture Notes - 1

17

Some Networking Issues

LAN, MAN and WAN

Switching and routing
Technologies: Ethernet, FDDI, ATM
Wireless/Mobile networking
Internetworking
Applications
Service quality
Security concerns

CSE-550-T112

Lecture Notes - 1

18

Network Design: Achievable?

Response Time

Business Growth

Reliability
CSE-550-T112

Cost

Lecture Notes - 1

19

Where to begin?
WWW
Addressing Traffic
Patterns
Access
Campus

Security

CSE-550-T112

Users

WAN

Lecture Notes - 1

Dial in
Users
Network
Management

20

Traditional Network Design Methodology

Many network design tools and methodologies that

have been used resemble the connect-the-dots

game

These tools let you place internetworking devices on

a palette and connect them with LAN or WAN media

Problem with this methodology:

It skips the steps of analyzing a customer's
requirements, and selecting devices and media based
on those requirements

CSE-550-T112

Lecture Notes - 1

21

Top-Down Network Design Methodology (1/2)

Good network design
Recognizes that a customers requirements embody
many business and technical goals
May specify a required level of network performance,
i.e., service level
Includes difficult network design choices and tradeoffs
that must be made when designing the logical network
before any physical devices or media are selected
When a customer expects a quick response to a

network design request

A bottom-up (connect-the-dots) network design

methodology can be used, if the customers
applications and goals are well known

CSE-550-T112

Lecture Notes - 1

22

Top-Down Network Design Methodology (2/2)

Network designers often think they understand a

customers applications and requirements.

However, after the network installation, they may
discover that:

They did not capture the customer's most important

needs
Unexpected scalability and performance problems
appear as the number of network users increases

CSE-550-T112

Lecture Notes - 1

23

Top-Down Network Design Process (1/2)

Begins at the upper layers of the OSI reference

model before moving to the lower layers

Focuses on applications, sessions, and data transport

before the selection of routers, switches, and media
that operate at the lower layers

Explores divisional structures to find the people:

For whom the network will provide services, and
From whom to get valuable information to make the
design succeed

CSE-550-T112

Lecture Notes - 1

24

Top-Down Network Design Process (2/2)

It is an iterative process:
It is important to first get an overall view of a
customer's requirements
More detail can be gathered later on protocol behavior,
scalability requirements, technology preferences, etc.
Recognizes that the logical model and the physical

design may change as more information is gathered

A top-down approach lets a network designer get
the big picture first and then spiral downward into
detailed technical requirements and specifications

CSE-550-T112

Lecture Notes - 1

25

Network Development Life Cycle

Analysis
Management/
Planning

Design

Simulation/
Prototyping

Monitoring/
Maintenance
Implementation

CSE-550-T112

Lecture Notes - 1

26

Network Design and Implementation Cycle

CSE-550-T112

Lecture Notes - 1

27

Network Design and Implementation Cycle (1/3)

Analyze requirements:
Interviews with users and technical personnel
Understand business and technical goals for a
new or enhanced system
Characterize the existing network: logical and
physical topology, and network performance
Analyze current and future network traffic,
including traffic flow and load, protocol
behavior, and QoS requirements

CSE-550-T112

Lecture Notes - 1

28

Network Design and Implementation Cycle (2/3)

Develop the logical design:
Deals with a logical topology for the new or
enhanced network
Network layer addressing and naming
Switching and routing protocols
Security planning
Network management design
Initial investigation into which service
providers can meet WAN and remote access
requirements
CSE-550-T112

Lecture Notes - 1

29

Network Design and Implementation Cycle (3/3)

Develop the physical design:
Specific technologies and products to realize the
logical design are selected
The investigation into service providers must be
completed during this phase
Test, optimize, and document the design:
Write and implement a test plan
Build a prototype or pilot
Optimize the network design
Document your work with a network design proposal
CSE-550-T112

Lecture Notes - 1

30

Another Perspective
Data collection
Traffic
Costs
Constraints
Design process
Performance analysis
Fine tuning
A painstaking iterative process

CSE-550-T112

Lecture Notes - 1

31

PDIOO Network Life Cycle (1/3)

(Cisco)
Plan:

Network requirements are identified in this phase

Analysis of areas where the network will be installed
Identification of users who will require network services

Design:

Accomplish the logical and physical design, according

to requirements gathered during the Plan phase

Implement:

Network is built according to the Design specifications

Implementation also serves to verify the design

CSE-550-T112

Lecture Notes - 1

32

PDIOO Network Life Cycle (2/3)

(Cisco)
Operate:

Operation is the final test of the effectiveness of the design

The network is monitored during this phase for performance
problems and any faults, to provide input into the Optimize
phase

Optimize:

Based on proactive network management which identifies

and resolves problems before network disruptions arise
The optimize phase may lead to a network redesign
if too many problems arise due to design errors, or
as network performance degrades over time as actual
use and capabilities diverge
Redesign may also be required when requirements change
significantly

CSE-550-T112

Lecture Notes - 1

33

PDIOO Network Life Cycle (3/3)

(Cisco)
Retire:

When the network, or a part of the network, is out-of-date, it

may be taken out of production
Although Retire is not incorporated into the name of the life
cycle (PDIOO), it is nonetheless an important phase

CSE-550-T112

Lecture Notes - 1

34

One More Look

Business
Planning

Network
Design

Implement
Network

Operations

Define Objectives
and Requirements

Develop
Architecture

Create
Implementation Plan

Develop Operations
Policies and
Capabilities

Create Initial
Solution

Develop Detailed
Design

Procure Resources
and Facilities

Fault
Management

Define Deployment
Strategy

Create Build
Documentation

Stage and Install

Configuration
Management

Review and
Approve

Review and Verify

Design

Certify and Hand-off

to Operations

Change
Management

Performance
Management

CSE-550-T112

Lecture Notes - 1

35

Information Flows between Network

Analysis, Architecture, and Design

CSE-550-T112

Lecture Notes - 1

36

Network Analysis and Design Methodology

- Overall Characteristics Requirements (business, application, and data)

definition is required prior to network design activities

Expected compliance with requirements in a Request
For Proposal (RFP) by both in-house personnel and
outside consultants
Activities from various stages often take place
simultaneously and backtrack to previous activities is
sometimes needed
This methodology is an overall guideline to the
network development process rather than cookbook
instructions
CSE-550-T112

Lecture Notes - 1

37

Network Analysis and Design Methodology

- Critical Success Factors of the NDLC (1/3) Identification of all potential customers and

constituencies

All groups must be consulted

Political awareness:

Corporate culture: hierarchical, distributed, or open

Backroom politics can play a role in systems design
Find ways to ensure objectivity of the analysis and design
process (e.g., measurable goals)

Buy-in:

Reach consensus on the acceptability of results of each stage

Approved results of one stage become the foundation or
starting point for the next stage
Makes the final presentation smoother

CSE-550-T112

Lecture Notes - 1

38

Network Analysis and Design Methodology

- Critical Success Factors of the NDLC (2/3) Communication:
With all groups
Write memos, communicate with key people in person,
etc.
Detailed project documentation:

Prepare agendas
Take meeting minutes
Action items
Use a project binder for all the above

CSE-550-T112

Lecture Notes - 1

39

Network Analysis and Design Methodology

- Critical Success Factors of the NDLC (3/3) Process/Product awareness:
Stay focused: what is the process/product at each
stage?
Keep meeting on track: no off-subject discussions
Be honest with yourself:

Be your own harshest critic (no one else knows the

potential weaknesses or areas for improvement in your
proposal better than you)
Use peer reviews
Not all weaknesses can be corrected (e.g., financial or
time constraints)

CSE-550-T112

Lecture Notes - 1

40

Network Analysis and Design Methodology

- Overall Guidelines Start with a clearly defined problem:
Identify affected parties and representatives
Held brainstorming sessions to define problems and
requirements of a solution
Understand strategic business objectives defined by

senior management
Collect baseline data from customer groups about
the current status of the system and network

This is used to measure eventual impact of the

installed network

Perform a feasibility study: problem definition and

associated alternative recommendations for further

study

CSE-550-T112

Lecture Notes - 1

41

Customers Requirements
- Understanding the Customer A good network design must recognize the customers

requirements - need to make sure your design meets

THEIR needs and not just YOURS!
The Customer may be your own firm, the who you are
designing the network for
Need an overview of a customers requirements
The best designed network will fail miserably without
the support of people

CSE-550-T112

Lecture Notes - 1

42

Customers Requirements
- Users Needs What do the users want?
Services
What do the users need?
What dont they know but they need?
Organize and Prioritize Requirement

CSE-550-T112

Lecture Notes - 1

43

Customers Requirements
- How they are used User Requirements

Performance Requirements

CSE-550-T112

Timeliness
Interactivity
Reliability
Quality
Security
Affordability
User Numbers
User Locations
User Growth
Lecture Notes - 1

Delay
Reliability

Capacity

44

Analysis and Design Processes

Set and achieve goals
Maximizing performance
Minimizing cost
Optimization with trade-offs
Recognizing trade-offs
No single best answer
Hierarchies
Provide structure in the network
Redundancy
Provides availability & reliability

CSE-550-T112

Lecture Notes - 1

45

Approaches Used for Design

Heuristic by using various algorithms
Exact by working out mathematical

solutions based on linear programming, etc.,

minimizing certain cost functions
Simulation often used when no exact

analytical form exists. Experiments are

conducted on simplified models to see the
performance of a network
CSE-550-T112

Lecture Notes - 1

46

Design and Study of a System

CSE-550-T112

Lecture Notes - 1

47

Art or Science?
The Art of Network Design
Technology choices
Relations to business goals

The Science of Network Design

Understanding of network technologies
Analysis of capacity, redundancy, delay

CSE-550-T112

Lecture Notes - 1

48

Types of Network Design

New network design
Re-engineering a network design
Network expansion design

CSE-550-T112

Lecture Notes - 1

49

New Network Design

Actually starting from scratch
No legacy networks to accommodate
Major driver is the budget, no compatibility

issues to worry about

Getting harder to find these situations

CSE-550-T112

Lecture Notes - 1

50

Re-engineering a Network Design

Modifications to an existing network to

compensate for original design problems

Sometimes required when network users

change existing applications or functionality

More of the type of problems seen today

CSE-550-T112

Lecture Notes - 1

51

Network Expansion Design

Network designs that expand network

capacity
Technology upgrades
Adding more users or networked equipment

CSE-550-T112

Lecture Notes - 1

52

This Whole Thing is Messy

This Whole Thing is Messy

Ambiguous Requirements

The network will only transport IP

The application requires Novell IPX

CSE-550-T112

Lecture Notes - 1

54

This Whole Thing is Messy

Conflicting Requirements

Keep costs down

High performance costs money

CSE-550-T112

Lecture Notes - 1

55

This Whole Thing is Messy

Lack of Design Tools
Lack of Management Tools
Lack of Vendor Interoperability

CSE-550-T112

Lecture Notes - 1

56

This Whole Thing is Messy

Lack of Documentation

Existing network

How things should be done (e.g., wiring)

Vendor information

CSE-550-T112

Lecture Notes - 1

57

This Whole Thing is Messy

Network Management

More management uses more bandwidth

Every vendor has their own management tools

Vendor tools may conflict with each other

CSE-550-T112

Lecture Notes - 1

58

This Whole Thing is Messy

Security
What is enough security?
What is too much security?

Security and management can not be dealt

with as afterthoughts. It is not an add-on
feature, it has to be integrated within.
Firewall
200Kbs

10Mb/s
Ethernet

CSE-550-T112

10Mb/s
Ethernet

T1 1.5Mb/s

Lecture Notes - 1

59

This Whole Thing is Messy

Evolving Network Technologies
Everything is a moving target
Products are put onto the market before
standards are approved
Everyone is a computer expert

CSE-550-T112

Lecture Notes - 1

60

OAM&P
Operations, Administration, Maintenance, Provisioning
Network
Management

Network
Provisioning

Network
Operations

Network
Maintenance

Planning

Fault Management / Service Restoration

Fault Management

Design

Configuration Management

Trouble Ticket
Administration

Performance Management / Traffic Management

Network Installation
Security Management
Network Repairs
Accounting Management
Facilities Installation
& Maintenance

Reports Management

Routine Network
Inventory Management

Tests

Data Gathering & Analyses

CSE-550-T112

Lecture Notes - 1

Figure 1.21 Network Management Functional Groupings

61

Functional Flow Chart

Network
Users

Configuration Data

Management
Decision

TT Restoration

Performance & Traffic Data

New
Technology

Engineering Group

Operations Group
NOC

- Network Planning &

Design

I & M Group
-Network Installation &
Maintenance

- Network Operations

Fault TT
Installation
CSE-550-T112

Lecture Notes - 1

Figure 1.22. Network Management Functional Flow Chart

62

Network Management Tools

SNMP
Application-level protocol
Part of TCP/IP protocol suite
Runs over UDP
From management station, three types of SNMP

messages issued

GetRequest, GetNextRequest, and SetRequest

Port 161

Agent replies with GetResponse

Agent may issue trap message in response to event

that affects MIB and underlying managed

Port 162

SNMP v2 (1)
Framework on which network management

applications can be built

e.g fault management, performance

monitoring, accounting

Protocol used to exchange management

information
Each player maintains local MIB

Structure defined in standard

At least one system responsible for

management

Houses management applications

SNPM v2 (2)
Support central or distributed management
In distributes system, some elements operate

as manager and agent

Exchanges use SNMP v2 protocol
Simple request/response protocol
Typically uses UDP

Ongoing reliable connection not required

Reduces management overhead

Figure 17.5 SNMPv2 Managed Configuration

SNMP v3
Addresses security issues of SNMP v1/2
RFC 2570-2575
Proposed standard January 1998
Defines overall architecture and security

capability
To be used with SNMP v2

SNMP v3 additional Services

(security)
Authentication

Part of User-Based Security (UBS)

Assures that message:

Came from identified source

Has not been altered
Has not been delayed or replayed

Privacy

Encrypted messages using DES

Access control

Can configure agents to provide a number of levels of

access to MIB
Access to information
Limit operations

Structure of
Management Information
SMI
Defines general framework with which MIB

defined and constructed

Identifies data types
How resources are represented and named
Encourages simplicity and extensibility
Scalars and two dimensional arrays of
scalars (tables) only

Network Diagnostic Tools

To assist a network support person, a number

of diagnostic tools are available:

Electrical testers
Cable testers
Network testers: monitor network health
Protocol/network analyzers: capture and
check packet contents

Data Communications and

Computer Networks: A
Business User's Approach,

70

Network Analysis and Sniffing

Process of capturing, decoding, and

analyzing network traffic

Why is the network slow
What is the network traffic pattern
How is the traffic being shared between nodes

Known as traffic analysis, protocol analysis,

sniffing, packet analysis, eavesdropping*, etc.

See wireshark screen captures next slides

*Listen secretly to what is said in private!

Wireshark Window
Menu Bar
Tool Bar

Filter Bar

Info
Field

Summary
Window

Protocol Tree Window

Data View Window

Disp.
Info field

Protocol Tree
Window:
Details of the
selected
packet (#8)
Raw data (content of
packet # 8)

Packet
number 8
BGP
(Boarder
Gateway
Prot)

Filtering
BGP
packets
only

Summary
When creating new network or adding to existing

network, there are many potential pitfalls and

opportunities for inaccurate and incomplete
assessments
NDLC is one of the most popular techniques used to
guide analysts through difficult decision-making
process of network design
An important part of NDLC is conducting of one or
more feasibility studies
Payback analysis is one possible financial analysis
technique that can be applied to proposed computer
network system to determine systems costs and
benefits
75
Data Communications and Computer Networks: A Business User's Approach, Sixth Edition

Summary (continued)
Network performance statistics should always be

monitored
A large variety of diagnostic tools are available that can
aid network personnel in troubleshooting and
maintaining common complex computer networks
The Simple Network Management Protocol (SNMP)
helps network support personnel monitor network
performance, find network problems, and then solve
those problems without physically touching the affected
device

76 A Business User's Approach, Sixth Edition

Data Communications and Computer Networks:

References

P. Oppenheimer, Top-Down Network Design, Cisco Press, 3 rd

edition, 2010
J. McCabe, Network Analysis, Architecture, and Design
Morgan Kaufmann Publishers, Inc., 3rd edition, 2007
J. E. Goldman, Applied Data Communications - A BusinessOriented Approach, 1998
Mani Subramanian, Network Management Principles and
Practice by, Pearson, Second Edition, 2010.

CSE-550-T112

Lecture Notes - 1

77