1

CHAPTER

PRACTICE SET
Questions
Q1-1.

In this case, when a station sends a message to another station, the message is
received by all stations. All stations except the intended recipient need to drop
the message. This is an example of broadcast transmission (one to many).

Q1-2.

There is no need for a link-layer switch to have an address. The message is not
directly sent to the switch; it is sent to the intended recipient (Host 3). The
switch needs to have the intelligence to recognize the destination address and
send the message out of the appropriate connection.

Q1-3.

Each LAN should be connected to (n 1) LANs. This means that we have n

(n 1) connections. However, if each connection can be used in both directions, we need only [n (n 1)]/2 connections.

Q1-4.

Local telephone communication is mostly a circuit-switched network. When

the communication is established, the circuit is dedicated between the two
parties.

Q1-5.

The telephone company acts as an ISP. The connection from the resident to
the telephone company is a point-to-point access WAN that connects the
premises to the Internet. At the same time, the telephone company needs to
provide the necessary services such as e-mail.

Q1-6.

To make the communication bidirectional, each layer needs to be able to provide two opposite tasks, one in each direction.

Q1-7.

The link-layer switch is normally involved in the first two layers of the TCP/
IP protocol suite: the physical layer and the data-link layer.

Q1-8.

The router is involved in three physical layers, three data-link layers, and only
one network layer.

Q1-9.

The identical objects are the two messages: one sent and one received.

Q1-10.
a. At the application layer, the unit of data is a message.
b. At the network layer, the unit of data is a datagram.
c. At the data-link layer, the unit of data is a frame.
1

Q1-11.

A frame is a link-layer data unit. It encapsulates a data unit coming from the
network layer. In this case, the data unit is a datagram.

Q1-12.

A user datagram is a transport-layer data unit. It decapsulates a data unit going

to the application layer. In this case, the data unit is a message.

Q1-13.

The data unit should belong to layer 4. In this case, it is a user datagram.

Q1-14.

We mentioned HTTP, FTP, SMTP, SNMP, TELNET, SSH, and DNS.

Q1-15.

The transport-layer packet needs to include two port numbers: source and destination port numbers. The transport-layer header needs to be at least 32 bits
(four bytes) long, but we will see in Chapter 3 that the header size is normally
much longer because we need to include other pieces of information.

Q1-16.
a. At the application layer, we normally use a name to define the destination-

computer name and the name of the file we need to access. An example is
[email protected].
b. At the network layer, we use two logical addresses (source and destination)

to define the source and destination computers. These addresses are unique
universally.
c. At the data-link layer, we use two link-layer addresses (source and destina-

tion) to define the source and destination connections to the link.

Q1-17.

The answer is no. Multiplexing/demultiplexing at the transport layer does not

mean combining several upper-layer packets (from the same or different
applications) into one transport-layer packet. It only means that each of the
transport-layer protocols (such as TCP or UDP) can carry a packet from any
application-layer protocol that needs its service. However, a transport-layer
packet can carry one, and only one, packet from an application-layer protocol.
For example, UDP can carry a message from FTP in one user datagram and a
message from HTTP in another user datagram.

Q1-18.

The application layer is the top layer in the suite; it does not provide services
to any layer, which means multiplexing/demultiplexing does not exist for this
layer.

Q1-19.

We do not need a link-layer switch because the communication in this case is

automatically one-to-one. A link-layer switch is needed when we need to
change a one-to-many communication to a one-to-one.

Q1-20.

We do not need a router in this case because a router is needed when there is
more than one path between the two hosts; the router is responsible for choosing the best path at each moment.

Q1-21.

An Internet draft is a working document with no official status and a sixmonth lifetime; an Internet draft may become a proposed standard after six
months if it has received enough attention in the Internet community.

Q1-22.

A required RFC is one that needs to be implemented by all Internet systems. A

recommended RFC does not have to be implemented by all Internet systems;
it is recommended only for its usefulness.

Q1-23.

The IETF is responsible for identifying operational problems and recommending solutions; the IRTF focuses on long-term research topics.

Problems
P1-1.

The services provided in part a and part b are the opposite of each other.
a. Layer 1 takes the ciphertext from layer 2, inserts (encapsulates) it in an

envelope and sends it.

b. Layer 1 receives the mail, removes (decapsulates) the ciphertext from the

envelope and delivers it to layer 2.

P1-2.

The services provided in part a and part b are the opposite of each other.
a. Layer 2 takes the plaintext from layer 3, encrypts it, and delivers it to layer

1.
b. Layer 2 takes the ciphertext from layer 1, decrypts it, and delivers it to

layer 3.
P1-3.

In 10 years, the number of hosts becomes about six times (1.2010 6.19) the
number in 2010. This means the number of hosts connected to the Internet is
more than three billion.

P1-4.

The system transmits 150 bytes for a 100-byte message. The efficiency is 100/
150 or 66.66%.

P1-5.

The advantage of using large packets is less overhead. When using large packets, the number of packets to be sent for a huge file becomes small. Since we
are adding three headers to each packet, we are sending fewer extra bytes than
in the case in which the number of packets is large. The disadvantage manifests itself when a packet is lost or corrupted during the transmission; we need
to resend a large amount of data.

P1-6.
a. The network layer is responsible for route determination.
b. The physical layer is the only layer that is connected to the transmission

media.
c. The application layer provides services for the end users.

P1-7.
a. User datagrams are created at the transport layer.
b. The data-link layer is responsible for handling frames between adjacent

nodes.
c. The physical layer is responsible for transforming bits to electromagnetic

signals.
P1-8.

There should be an upper-layer identifier in the header of the IP protocol to

define to which upper-layer protocol the encapsulated packet belongs.

P1-9.

The following shows the situation. If we think about multiplexing as many-toone and demultiplexing as one-to-many, we have demultiplexing at the source
node and multiplexing at the destination node in the data-link layer. However,
some purists call these two inverse multiplexing and inverse demultiplexing.

FTP

HTTP

DNS

UDP

TCP

Multiplexing

Multiplexing /
Demultiplexing

L1

SNMP

L2

DNS

HTTP

IP

L3

L1

SNMP

UDP Demultiplexing

TCP

IP

a. At the source node

P1-10.

FTP

L2

Demultiplexing /
multiplexing

L3

Multiplexing

b. At the destination node

Every time any packet at any layer is encapsulated inside another packet at the
same layer, we can think of this as a new layer being added under that layer.
The following shows the new suite.

Application
Security
Transport
Network
Link
Physical

Application
Security
Transport
Network
Link
Physical

P1-11.

The following shows the layers. Note that we have not shown the security
checking that you need to pass through because it does not have the counterpart when you arrive. It must be included in baggage/checking layer.

Source/destination
airport

Source/destination
airport

Baggage
checking/claiming

Baggage
checking/claiming

Boarding/unboarding

Boarding/unboarding

Takeoff/Landing

Takeoff/Landing
Flying

P1-12.

Communication between hosts in an internetwork should take place using the

network-layer addresses. Assume a host at the west coast office sends a packet
to a host in the east cost office. Since the switch works only at the first two
layers, it cannot check the network-layer address in the packet. The switch
sends a frame with the address not in the west coast office to router R1. If R1
can recognize the network-layer address as belonging to one of the hosts at the
east coast office, it sends it to router R2. Otherwise, the packet is discarded.
The same is true about router R2. In other words, each router actually has
three connections (interfaces), two shown in the figure and one invisible (the
one from which the packet is discarded). This means that when the router
receives a packet from an input connection, it needs to choose the best output
connection out of the two (concept of routing).

P1-13.

The following shows the position of the presentation layer. The new layer is at
the same position as the presentation layer in the OSI model if we ignore the
session layer.

Application
Presentation
Transport

P1-14.

Application
New layer

Presentation
Transport

Network

Network

Data link

Data link

Physical

Physical

The only two layers that need to be changed are the data-link layer and the
physical layer. The new hardware and software need to be installed in all host,
routers, and link-layer switches. As long as the new data-link layer can encapsulate and decapsulate datagrams from the network layer, there is no need to
change any protocol in the upper three layers. This is one of the characteristics
of the protocol layering.

P1-15.

The reason for having several protocols in a layer is to provide different services to the upper-layer protocols. The services provided by UDP are different
from the services provided by TCP. When we write an application program,
we need to first define which transport-layer protocol is supposed to give services to this application program. The whole program is written based on the
availability of these services. Note that this does not violate the principle of
layer independence. The independency of a layer means that we can change a
protocol in a layer as long as the new one gives the same services as the old
one. This does not mean that we can replace UDP by TCP, because they provide different services.

P1-16.

The following shows the layers and the flow of data. Note that each host is
involved in five layers, each switch in two layers, and each router in three layers.

East coast office

West coast office

Point-to-point
WAN

R1

R2

Router

Router

Host

Switch

Router

Router

Switch

Host