Lecture 11

ECE 452 Communication Networks

Reference

“Computer Networks” by Andrew S. Tanenbaum

Can also be found in other books (mentioned in the syllabus) but we will follow the order of this
book.

Lecture Contents

Why do we need a network?

Types of networks

1
Footnotes are relevant page numbers from the reference (Tanenbaum). Page numbers refer to the hardcopy.
For the softcopy add around 4 to each page number.

1
Today is a general introduction about Networks.

Why do we need a network?

To share information between more than two end points. (If they are just two it is not a network).

Figure 1: A generic network

Also these locations can be far away from each other.

Types of networks

There is usually something called a core network covering a large area. For example in a company it
covers the whole company.

ISP stands for Internet Service Provider. This provides a connection to the internet. Like TE-Data.

What are the methods of accessing the Internet?

1. Dial-up, wireless
2. Cable
3. DSL
4. T1
5. T3
6. E1

2
 7. E3
8. Fiber.

Cable is not available in Egypt, but in some other countries. It means that there is a certain network
that provides: TV, telephone and internet.

Service Bandwidth Range

Fiber is usually not used by a single user, but shared by many in a close-by location.

The Router

The router is put in a place that looks like a large closet (‫)دوالب‬, with certain openings for air-cooling.

A router or switch is formed of a processor and a certain number of cards, called interface cards.

The router2 can be considered as a computer having a processor, memory and hard-disk. Does it
have a keyboard and screen? It has some “port” either serial/Ethernet or something else, can be
connected to a PC. So, now we have a complete computer and we can configure the router.

If I want to change the configuration of the router, I can log in the router and configure it remotely
from the control center without accessing it directly (physically).

Core and edge3

If the network should do a lot of services,

The job of the core network is to connect with a large bandwidth. Regardless of the number of
interfaces (don’t have to be large).

Like the fiber that was cut between Egypt and Europe 2 years ago.

The )‫ (سنترال‬can either be edge or core.

*I’m arranging for a trip to El-Masreyya lel-itisalat to see the routers and switches, etc. But they only
need 80 students.]

2
Definition of a router: page 19.
3
Couldn’t find this classification in the reference.

3
Networks can be classified according to:

1. Physical distance
a. LAN: Local Area Network4
b. WAN: Wide Area Network5
2. Connection (Communication)
a. Client/Server6
b. Pear-to-pear7
3. Service
a. Core (backbone)
i. ATM
ii. Frame Relay (FR)
iii. Internet (IP)
b. Edge
i. DSL
ii. VoIP
iii. E1, E3, …
iv. T1, T3, …
v. Fiber
4. Wired/Wireless
a. Wired
b. Wireless8
i. WLAN
1. Infrastructure (Access point)
2. Pear-to-pear (Bluetooth)
ii. WWAN
1. Infrastructure (cell phone)
2. Ad Hoc

Later we’ll see the difference between the Hub, the switch, the router and the gateway.

Here the first classification is according to the distance. The LAN is in a small area (around 100m),
while the WAN is in a wide area (>100m).

Why do we classify according to distance?

The distance above 100m needs different protocols and trying to connect a LAN that distance may
not work.

4
LANs: pages 16 (partially) and 17.
5
WANs: pages 19, 20 and 21 (partially).
6
Client/Server model: pages 4 and 5 (partially).
7
Pear-to-pear communication: pages 5 and 6 (partially).
8
Wireless networks: pages 21 (partially), 22 and 23.

4
 Figure 2

If A is sending to C and then D starts to send, there will be collision.

There is a method such that the device checks first if there is a signal on the line before it sends.

But if both devices send at the same time, or with difference so small such that the signal didn’t
reach the second device. So, in that case a collision will happen.

As the distance increases, the probability that a collision may happen increases. This is because the
propagation times between devices increase.

Therefore this protocol9 (for example) can’t be used for wide networks.

What is the difference between a client and a server?

The client always starts the request, then the server replies, not the other way around.

In case of chatting, it is a peer-to-peer case.

What do client and server mean?!

They are the software that reply and request, so the machine itself is not the server/client.

If – for example – you have certain software for chat. Part of it “client” you send request, and the
reply comes from the other point and the “server” part that replies to a request send by another
point.

The most famous core network is the Internet (IP).

We can also classify according to wireless or wired.

Wireless means the connection between the two points doesn’t use a wire. The “mobile” has a
different definition because it can move.10

Can something be fixed and wireless?

9
Definition of the word protocol: page 27
10
For the definition of wireless/mobile see pages 9 (partially) and 10.

5
A PC connected to USB modem. Also, wireless cards to connect far away PCs.

Can something be mobile and wired?

Dial-up can be considered “magazan” mobile. You can access it anywhere; all you need is a
telephone line.

Why do we classify wireless networks into large and small areas?

We do so because there are different protocols. The attenuation of the signal is large in air. This
means that the Wireless WAN needs higher power in transmission than the wireless LAN (for
example).

Devices connected to a wireless LAN have the same frequency and collision may happen. While for a
wireless WAN, multiple frequencies are used.

For both wireless LAN and WAN both can be classified into client/server.

If you look at an access point with some 10 computers connected to it. This is called Infrastructure
and is shown in figure 1

Peer to peer is like Bluetooth.

Also for wireless WAN we have Infrastructure and peer-to-peer. An example for the first is the cell
phones. (Up till the base station)

Soldiers in the marines wearing helmet with an antenna. These helmets are wireless, mobile routers.

So if one soldier has broken his helmet, others can still communicate.

5. Destination11
a. Point-to-point
b. Multi-cast
c. Broadcast
6. Switching
a. Circuit switching
b. Packet switching
i. Virtual circuit switching
ii. Data gram
7. Connection type
a. Connection-oriented
b. Connectionless
8. Reliability
a. Reliable

11
Broadcast, multicast and point-to-point networks: page 15 (partially)

6
 b. Unreliable

Are you sending to one destination (for example one: e-mail or chat).

Broadcast sends to everyone. Like radio and TV.

Multicast sends to certain users, not everyone, but more than one defined users. Also the satellite.
Not everyone receives it.

The router can also do that and send to only some users.

Switching is very important, there are two types, circuit and packet switching. We’ll talk about it in
more details during the course. The telephone for example when you talk to someone, the channel
is dedicated to you, no sharing. Sharing happens in packet switching.

The difference between connection-oriented, when you chat, the user sets the connection, sends
the data and then ends the connection.

In broad-cast for example, data is sent directly.

Reliable means two things:

1. Data arrives in order

It is valid to send packet “2” then “1”. But the connection is reliable only if the receiver is able to re-
order it.

Voice over IP (VoIP) means chatting with voice over the Internet. This is an unreliable connection,
because there is no rigid restriction on the quality of the voice.