Department of Electrical and Computer Engineering

Laboratory Manual
for
CENG460 Communications Networks
By
Ruonan Zhang, Emad Shihab, Zhe Yang, Xuan Wang, Lei Zheng,
and Lin Cai
Copyright 2012 University of Victoria. All rights reserved
The knowledge acquired in the CENG460 course and the labs should NOT
be misused under any circumstances. Please carefully read and follow the
Policy on Responsible Use for Information Technology Services (available
at http://www.uvic.ca/6030) for using Information technology services at
the University of Victoria.
Please refer to the CENG 460 lab web page for supplementary lab informa-
tion.
This lab manual has adopted several contents from the labs suggested in [1,
3, 4, 5].
Contents
1 Lab 1: Introduction to WireShark and Layered Protocol 2
1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.1 WireShark . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.2 Networking Tools . . . . . . . . . . . . . . . . . . . . . 4
1.1.3 Layered Protocol . . . . . . . . . . . . . . . . . . . . . 5
1.2 Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.1 Installation . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.2 Getting familiar with WireShark . . . . . . . . . . . . 5
1.2.3 Layered Protocol . . . . . . . . . . . . . . . . . . . . . 7
1.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3.1 Running WireShark . . . . . . . . . . . . . . . . . . . . 10
1.3.2 Networking Tools . . . . . . . . . . . . . . . . . . . . . 10
1.3.3 Layered Protocol . . . . . . . . . . . . . . . . . . . . . 10
2 Lab 2: Ethernet and IEEE 802.11 12
2.1 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.1 Ethernet . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.2 IEEE 802.11 . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3.1 Analyzing Ethernet frames . . . . . . . . . . . . . . . . 14
2.3.2 Exploring IEEE 802.11 functions . . . . . . . . . . . . 14
2.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4.1 Analyzing Ethernet frames . . . . . . . . . . . . . . . . 16
2.4.2 Exploring IEEE 802.11 functions . . . . . . . . . . . . 16
i
3 Lab 3: ARP, IP, and ICMP 18
3.1 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.1 Address Resolution Protocol (ARP) . . . . . . . . . . . 18
3.2.2 Internet Protocol (IP) . . . . . . . . . . . . . . . . . . 19
3.2.3 Internet Control Message Protocol (ICMP) . . . . . . . 19
3.3 Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3.1 Exploring ARP functions . . . . . . . . . . . . . . . . . 20
3.3.2 Analyzing IP frames . . . . . . . . . . . . . . . . . . . 20
3.3.3 Exploring ICMP functions . . . . . . . . . . . . . . . . 21
3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.4.1 Exploring ARP functions . . . . . . . . . . . . . . . . . 22
3.4.2 Analyzing IP frames . . . . . . . . . . . . . . . . . . . 23
3.4.3 Exploring ICMP functions . . . . . . . . . . . . . . . . 23
4 Lab 4: TCP 25
4.1 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.2.1 TCP Header Format . . . . . . . . . . . . . . . . . . . 27
4.2.2 TCP Connection Setup . . . . . . . . . . . . . . . . . . 27
4.2.3 TCP Data Flow . . . . . . . . . . . . . . . . . . . . . . 28
4.2.4 TCP Connection Release . . . . . . . . . . . . . . . . . 28
4.2.5 TCP Congestion Control (Optional) . . . . . . . . . . 29
4.2.6 TCP Flow Control (Optional) . . . . . . . . . . . . . . 30
4.2.7 Retransmission in TCP . . . . . . . . . . . . . . . . . . 30
4.3 Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.3.1 TCP Header Format . . . . . . . . . . . . . . . . . . . 31
4.3.2 TCP Connection Setup . . . . . . . . . . . . . . . . . . 32
4.3.3 TCP Data Flow . . . . . . . . . . . . . . . . . . . . . . 32
4.3.4 TCP Connection Release . . . . . . . . . . . . . . . . . 33
4.3.5 TCP Congestion Control . . . . . . . . . . . . . . . . . 33
4.3.6 TCP Flow Control . . . . . . . . . . . . . . . . . . . . 33
4.3.7 Retransmission in TCP (Optional) . . . . . . . . . . . 33
4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.4.1 TCP Header Format . . . . . . . . . . . . . . . . . . . 34
4.4.2 TCP Connection Setup . . . . . . . . . . . . . . . . . . 34
4.4.3 TCP Data Flow . . . . . . . . . . . . . . . . . . . . . . 35
4.4.4 TCP Connection Release . . . . . . . . . . . . . . . . . 36
ii
4.4.5 TCP Congestion Control . . . . . . . . . . . . . . . . . 37
4.4.6 TCP Flow Control . . . . . . . . . . . . . . . . . . . . 37
4.4.7 Retransmission in TCP (Optional) . . . . . . . . . . . 37
1
Chapter 1
Lab 1: Introduction to
WireShark and Layered
Protocol
The labs for this course were designed to help students better understand
the ideas learned in the classes through hands-on experiments.
A better way to understand network protocols is to observe how they
actually work. A basic tool for observing the messages exchanged between
executing protocol entities is the packet snier, which is an essential part of
network protocol analyzer. WireShark is a free and open-source network
protocol analyzer that runs on various operating systems including Linux,
Unix, Mac, and Windows. We will give a brief overview of it in the following
section.
This lab has three parts. The rst part includes simple tasks that let you
get familiar with the basic operations of WireShark. The second part will
introduce some handy networking tools, which will be used in the following
labs. The third part will focus on how protocols and layering are represented
in packets by exploring the snied packet traces.
2
1.1 Overview
1.1.1 WireShark
WireShark (previously called Ethereal) is one of the most widely used net-
work protocol analyzer. It passively snis packets that are sent from or
received by a designated network interface, but never sends packets itself. It
receives a copy of packets that are sent from or received by the applications
and protocols executing on the end-system (e.g., your computer). WireShark
also has a graphical front-end to display the packets that it snis.
Figure 1.1: Network protocol analyzer structure
Fig. 1.1 [1] shows the structure of a network protocol analyzer. At the
right of the gure shows the protocol stack and applications (such as a web
browser or an FTP client) that normally run on your computer. The network
protocol analyzer, shown within the dashed rectangle, has two parts, the
packet capture and the packet analyzer. The packet capture library receives
a copy of every link-layer frame that is sent from or received by a designated
network interface. Recall that messages exchanged by higher layer protocols
such as HTTP, FTP, TCP, UDP, DNS, or IP all are eventually encapsulated
in link-layer frames that are transmitted over physical media such as an
Ethernet cable. In Fig. 1.1, the assumed physical media is an Ethernet, and
so all upper layer protocols headers are eventually encapsulated within an
Ethernet frame. Capturing all link-layer frames thus gives you all messages
sent from or received by all protocols and applications executing in your
computer.
The second component is the packet analyzer, which displays the contents
3
of all elds within a link-layer frame. In order to do so, the packet analyzer
must understand the structure of messages exchanged by the protocols. For
example, we are interested in displaying the various elds in messages ex-
changed by the HTTP protocol in Fig. 1.1. The packet analyzer understands
the format of Ethernet frames, and so it can identify the IP datagram within
an Ethernet frame. It also understands the IP datagram format, so it can
extract the TCP segment within the IP datagram. It understands the TCP
segment structure, so it can extract the HTTP message contained in the TCP
segment. Finally, it understands the HTTP protocol and so, for example,
knows that an HTTP message may contain the string of GET, POST,
or HEAD.
1.1.2 Networking Tools
ping
The ping program in the source host sends a packet to the target IP address;
if the target is alive, the ping program in the target host responds by sending
a packet back to the source host. Both of these ping packets carry ICMP
messages. Try ping -h to nd out its usage.
ifcong
The ifcong is a tool to congure a network interface, for instance, setting an
interfaces IP address and netmask, disabling or enabling a given interface.
Try ifcong --help to nd out its usage.
netstat
The netstat is a tool that displays network connections, routing tables, and
network interface statistics. It is used for nding problems in the network
and to determine the amount of trac on the network as a performance
measurement. Try netstat --help to nd its usage.
wget
wget is a command-line program that let you fetch a URL. Unlike a web
browser, which fetches and executes the entire pages, wget give you the
control over exactly which URLs you fetch and when you fetch them. wget
4
has many options (try wget --help to see them) but a URL can be fetched
simply with wget URL.
1.1.3 Layered Protocol
Two reference models are used to describe the network architecture, the
OSI/ISO reference model and the TCP/IP reference model. The OSI/ISO
model divides the network into seven layers and the TCP/IP model divides
the network into four layers. No matter which model is used, the basic
principle of the layered architecture is that each layer performs some services
for the layer above it.
1.2 Procedures
1.2.1 Installation
WireShark is free to download at http://www.wireshark.org/. How to build
and install WireShark onto machines with dierent operating systems can be
referred to http://wiki.wireshark.org/BuildingAndInstalling.
1.2.2 Getting familiar with WireShark
A. Starting WireShark
When you run WireShark, you will see the graphical user interface (GUI) as
shown in Fig. 1.2. There are four main elds:
Filter eld: Used to lter out uninteresting packets with the entered
specications, so you can choose which packets should (not) be shown
on the screen;
Captured packets: Lists the packets captured by the selected inter-
face;
Details of selected packet: Lists information about the packet that
is selected in the captured packets window;
Content of packet in hex/ASCII: Displays the content of the cap-
tured packet, in hex and ASCII.
5
Filter filed
Captured
packets
Details of
selected
packet
Content
of packet
in hex and
ASCII
Figure 1.2: WireShark graphical user interface
B. Capture Trace
Use the following procedure to capture the trace.
Pick a URL and fetch it by wget. For example, open a console, type
wget http://www.google.ca, you will obtain the fetched resource
written in a le. A successful example is shown in Fig. 1.3. The ex-
pected response is 200 OK.
Close web browser. By minimizing browser activity you will stop your
computer from fetching unnecessary web content, and avoid incidental
trac in the trace.
Launch Wireshark. Choose the network interface that we would like to
capture the packets on. To do this, select Capture Options from
the command menu. A window similar to the one shown in Fig. 1.4
should pop up. Select the interface you are using. Uncheck Capture
6
Figure 1.3: Wget URL
packets in promiscuous mode. This mode is useful to overhear packets
sent to/from other computers on broadcast networks. We only want
to record packets sent to/from your computer. Use capture lter tcp
port 80. This lter will record only standard web trac and not other
kinds of packets that your computer may send. Click Start to start
the packet capture process.
When the capture is started, repeat the web fetch using wget above.
This time, the packets will be recorded by Wireshark as the content is
transferred.
After the fetch is successful, return to Wireshark and use the menus or
buttons to stop the trace (Capture Stop). If you have succeeded,
the upper Wireshark window will show multiple packets. How many
packets being captured will depend on the size of the web page, but
there should be at least 8 packets in the trace. An example is shown
in Fig. 1.5.
1.2.3 Layered Protocol
By inspecting the captured trace, or the provided trace (lab1-wget-trace.pacp)
to understand the layered protocol.
7
Select
interface
Figure 1.4: Capture options window
Select an HTTP GET packet. This packet carries the HTTP request
sent from your computer to the server.
The protocol layers being used in web fetching are shown in Fig. 1.6.
HTTP is the application layer web protocol used to fetch URLs. It
runs on top of the TCP/IP transport and network layer protocols. The
link layer protocol shown in the gure is Ethernet. It may be other
protocol, depends on your network.
Click on one HTTP packet, and turn to the middle panel with details of
the packet. The rst block is Frame. This is a record that describes
overall information about the packet, including when it was captured
and how many bits long it is. The second block is Ethernet (You
may have taken trace in a computer with 802.11, but still you will see
an Ethernet block. This is because Wireshark capture trac in Ether-
net format determined on the capture options. See Link-layer header
type.). Then we can see IP, TCP, and HTTP. This is a bottom-up or-
der, because as packets are passed down the protocol stack, the header
of the lower layer protocol is added to the front of the information from
8
Figure 1.5: Packet trace
Figure 1.6: Protocol stack for a web fetch
the higher layer protocol. That is, the lower layer protocols come rst
in the packet.
9
When an Ethernet frame arrives at a computer, the Ethernet layer
must hand the packet that it contains to the next higher layer to be
processed. In order to do this, the protocol use information in its header
to determine the higher layer data unit encapsulated. Which eld is
used here?
Draw a gure of an HTTP GET packet that shows the position and
size in bytes of the TCP, IP and Ethernet protocol headers. On this
drawing, show the range of header and payload of each layer.
1.3 Discussion
1.3.1 Running WireShark
1. Capture a trace without lter.
2. List at least 3 dierent protocols that appear in the protocol column
in the unltered packet-listing window.
3. How long did it take from the HTTP GET message being sent to the
HTTP OK reply being received?
1.3.2 Networking Tools
Explore the usage of ifcong, ping, netstat, and answer the following
questions.
1. How many Ethernet interfaces are in your computer, how to determine
it?
2. How to turn down/up an Ethernet interface?
3. Ping 10 packets to two websites. Compare the statistic results (packet
loss, avg rtt).
1.3.3 Layered Protocol
1. Draw the structure of a HTTP GET packet.
10
2. In the provided trace (lab1-wget-trace.pacp), calculate the overall
overhead of all the packets from the server to the client (in percentage).
(Hint: overhead is the size of all headers in one packet over the total
size).
3. Which Ethernet header eld tells the next higher layer protocol is IP?
What value it used?
4. Which IP header eld tells the next higher layer protocol is TCP? What
value it used?
11
Chapter 2
Lab 2: Ethernet and IEEE
802.11
2.1 Objective
In this lab, we will investigate the link layer protocols, including the Ethernet
and IEEE 802.11. The rst part of this lab is mainly about the Ethernet
frame format. The second part of the lab focuses on analyzing IEEE 802.11
frames.
2.2 Introduction
2.2.1 Ethernet
Ethernet stations communicate by sending each other data frames. As with
other IEEE 802 LANs, each Ethernet station is given a single 48-bit MAC
address, which is used to specify the destination and the source of each data
frame. Network interface cards (NICs) or chips normally do not accept frames
addressed to other Ethernet stations. Adapters are generally programmed
with a globally unique MAC address, but this can be overridden, either to
avoid an address change when an adapter is replaced, or to use locally ad-
ministered addresses.
All generations of Ethernet (except very early experimental versions) share
the same frame formats (and hence the same interface for higher layers), and
12
can be readily (and in most cases, cheaply) interconnected.
Due to the ubiquity of Ethernet, the ever-decreasing hardware cost of it,
and the reduced panel space needed by twisted pair Ethernet, most manu-
facturers now build the functionality of an Ethernet card directly into PC
motherboards, eliminating the need for installation of a separate network
card.
2.2.2 IEEE 802.11
In this part, we are going to explore the link layer, and management func-
tions of 802.11. Generally speaking, there are three types 802.11 frames,
the Data frame (Type 2), the Control frame (Type 1), and the Management
frame (Type 0). For each type of frame, there are also dierent subtypes.
Typically, Data frame is the longest, which can be up to 1500 bytes, while
Management frames are much shorter, and Control frames are very short.
As the Data and Control frames have been illustrated in the text book, here
we introduce some important types of Management frames.
Beacon frame Beacon frames are sent out periodically by an AP to
advertise its existence and capabilities to nearby computers. Beacon is
an IEEE 802.11 wireless LAN Management frame. In a Beacon frame,
there are a series parameters, including the SSID name of the AP, the
data rates it supports, and the channel on which it is operating.
Association A computer has to associate with the AP after it learned
an AP via a Beacon or otherwise and before it can send or receive data
from the AP. Possibly, authentication process will be involved during
the association. If the Association Request is successful received by AP,
the AP will return an Association Response, and then the computer
will acknowledge the association response. The Association Request
and Response carry information that describes the capabilities of the
AP and computer. Thus, both endpoints can know the others abilities.
Probe Request/Response In addition to nd AP by waiting to learn
about an AP from Beacons, a computer may also probe for specic APs.
A Probe Request is sent by a computer to test whether an AP with a
specic SSID is nearby. If the AP is nearby, it will reply with a Probe
13
Response. Like Beacon and Association frames, each of these frames
carries information describing the capabilities of the computer and AP.
2.3 Procedures
2.3.1 Analyzing Ethernet frames
Download and open the trace named ethernet-trace-1.
Because we are interested in Ethernet, we need not to consider any
IP or higher layer protocols. Change WireSharks settings so that it
shows information only about protocols below IP: select Analyze
Enabled Protocols and unchecked IP. The resulting gure should
look like Figure 2.1.
Find the HTTP GET message that was sent from the web browser to
gaia.cs.umass.edu (should be packet No.10) and answer question (1)-
(4) in section 2.4.1.
Find the Ethernet frame containing the rst byte of the HTTP response
message and answer question (5)-(8) in section 2.4.1.
2.3.2 Exploring IEEE 802.11 functions
Download and open the trace named wlan-trace-1 [4]. Note that it
may be dicult to gather your own trace using windows system. The
main issue is that Windows system made 802.11 frames appear to come
via a wired Ethernet. However, it is possible to use Mac or Linux to
gather 802.11 frames directly, without this conversion.
Select a Data packet. The packet detail can show four layers informa-
tion: 1) Frame, which is a record added by Wireshark with information
about the time and length of the frame; 2) Radiotap, which is also a
record of captured physical layer parameters, such as the strength of
the signal and the modulation; 3) IEEE 802.11, which is the bits of
the 802.11 Data frame;4) Data, which is a record containing the frame
payload data. Answer the related questions in section 2.4.2.
14
Figure 2.1: Screen shot of WireShark
Inspect dierent packets to see the values for dierent types of frames.
You can use lter to see only one type frames by entering the expression
wlan.fc.type==n into the Filter box above the list of frames in the top
panel. For example, n=2 is for data frames, n=1 is for control
frames, and n=0 is for management frames. Answer the related
questions in section 2.4.2.
Inspect the packet transmission reliability. Use lter expressions to
nd the number of Data frames that are originals and retransmissions.
For example, wlan.fc.type==2 && wlan.fc.retry==0 will nd original
Data frames. Answer the related questions in section 2.4.2.
Inspect the Management frame. Use lter to help you nd these frame,
and answer the related questions in section 2.4.2.
15
2.4 Discussion
2.4.1 Analyzing Ethernet frames
For trace le ethernet-trace-1, answer the following questions.
1. What is the 48-bit Ethernet address of the client computer?
2. What is the 48-bit destination address in the Ethernet frame? Is this
the Ethernet address of gaia.cs.umass.edu? (Hint: the answer is no).
What device has this as its Ethernet address?
3. Give the hexadecimal value for the two-byte Frame type eld.
4. What do the bit(s) whose value is 1 mean within the ag eld?
5. What is the value of the Ethernet source address? Is this the address
of your computer, or of gaia.cs.umass.edu (Hint: the answer is no).
What device has this as its Ethernet address?
6. What is the destination address in the Ethernet frame? Is this supposed
to be the Ethernet address of your computer?
7. Find the hexadecimal value for the two-byte Frame type eld.
8. What do the bit(s) whose value is 1 mean within the ag eld?
2.4.2 Exploring IEEE 802.11 functions
Answer the following questions based on the trace le wlan-trace-1.
1. Which AP is the most active one? what is its BSS ID?
2. How many Data frames are in the trace, how many subtypes, and what
is the most common subtype of Data frame?
3. How many subtypes of Control frames are in the trace, what are they?
and what is the most common subtype?
4. How many subtype of Management frames are in the trace, what are
they and what is the most common subtype?
16
5. Give an estimate of the retransmission rate as the number of retransmis-
sions over the number of original transmissions. Show your calculation.
6. What are the Type and Subtype values for the Association Request/Association
Response frames, the Probe Request/Probe Response frames?
17
Chapter 3
Lab 3: ARP, IP, and ICMP
3.1 Objective
In this lab, we will investigate the Address Resolution Protocol (ARP), the
IP (Internet Protocol), and Internet Control Message Protocol (ICMP). The
rst part of this lab is mainly about the ARP protocol. We will study the
operation of the protocol based on the elds that it sets in the Ethernet frame
containing the ARP message. The second part of the lab focuses on analyzing
IP frames, by observing and interpreting the elds in the IP frame. The last
part of this lab focuses on the format and content of the ICMP messages.
3.2 Introduction
3.2.1 Address Resolution Protocol (ARP)
ARP is the standard method for nding a hosts hardware address when
only its network layer address is known. It can be used to resolve many
dierent network-layer protocol addresses to hardware addresses. Due to
the overwhelming prevalence of IPv4 and Ethernet, ARP is primarily used
to translate IP addresses to Ethernet MAC addresses. ARP is used in the
following four cases when two hosts communicate:
1. When two hosts are on the same network and one desires to send a
packet to the other.
18
2. When two hosts are on dierent networks and must use a gateway/router
to reach the other host.
3. When a router needs to forward a packet for one host through another
router.
4. When a router needs to forward a packet from one host to the destina-
tion host on the same network.
The rst case is used when two hosts are on the same physical network (that
is, they can directly communicate without going through a router). The last
three cases are the most used over the Internet as two computers on the
Internet are typically separated by several hops.
3.2.2 Internet Protocol (IP)
Above the link layer, there is the network layer, which is responsible for relay
the date between the transport layer and the link layer. The network proto-
col in network layer is called the Internet Protocol, or more commonly, the IP
Protocol. The IP protocol performs two basic functions, the addressing (IP
address) and the packet routing. Note that the internet layer is agnostic of
application data structures as the transport layer, and it also does not dis-
tinguish between operation of the various transport layer protocols. Thus,
IP protocol can carry data for a variety of dierent upper layer protocols by
dierent protocol numbers, such as TCP, UDP and ICMP.
There are currently two versions of the IP protocol, IPv4 and IPv6. In this
section we examine IPv4, which the most widespread version. With given
trace les, were going to learn about the details of IP frame.
3.2.3 Internet Control Message Protocol (ICMP)
The Internet Control Message Protocol (ICMP) is one of the core protocols
for network management in the Internet. It is mainly used by networked com-
puters operating systems to send error messages indicating, for instance,
that a requested service is not available or that a host or router could not be
reached. It has been used in network trouble-shooting and analyzing appli-
cations such as ping and traceroute.
19
ICMP uses the basic support of IP as if it were a higher level protocol; how-
ever, ICMP is actually an integral part of IP, and must be implemented by
every IP module. ICMP messages are sent in several situations: for example,
when a datagram cannot reach its destination, when the gateway does not
have the buering capacity to forward a datagram, and when the gateway
can direct the host to send trac on a shorter route [RFC792].
In this part of the lab, we use two network tools. One is ping, which is used to
test whether a particular host is reachable across an IP network, to self-test
the network interface card of the computer, or to measure latency. The other
one is traceroute, used to determine the route taken by packets across an IP
network. We can understand the functions of ICMP by using these tools.
3.3 Procedures
3.3.1 Exploring ARP functions
Download and open the trace named ethernet-trace-1.
This trace basically emulates retrieving a long document.
The ARP protocol typically maintains a cache of IP-to-Ethernet ad-
dress translation pairs on your computer.
Find the ARP request message and answer questions (1)-(6) in sec-
tion 3.4.1.
Find the ARP reply that was sent in response to the ARP request and
answer questions (7)-(12) in section 3.4.1.
3.3.2 Analyzing IP frames
Using the same trace le as above.
Select any packet with the HTTP GET message in the trace and expand
the IP header elds (using the + expander or icon) to see the details.
You can simply click on a packet to select it (in the top panel). And the
details of its structure (in the middle panel) and the bytes that make
20
up the packet (in the bottom panel). Our interest is the IP header,
and you may ignore the other higher and lower layer protocols.
Select the the packet with HTTP GET message (the No.10 packet) and
answer questions (1)-(3) in section 3.4.2.
Observe all the packets and answer questions (4)-(6) in section 3.4.2.
3.3.3 Exploring ICMP functions
A. Ping The ping program in the source host sends a packet to the target IP
address; if the target is alive, the ping program in the target host responds
by sending a packet back to the source host. Both of these ping packets carry
ICMP messages.
The following procedures describe how to capture the traces of ping messages.
Start up the WireShark and begin packet capture.
Type the command ping www.engr.uvic.ca -n 10 in the command line.
The argument -n 10 indicates that ten ping messages should be sent.
When the ping program terminates, stop the packet capture.
Download and open ping-trace-1 in WireShark. Use the display lter to list
the ICMP messages only, as shown in Figure 3.1 and answer questions (1)-(4)
in section 3.4.3.
B. Traceroute The traceroute program is used to gure out the path a packet
takes from the source to the destination. The following procedures describe
how to capture the packets of traceroute messages.
Start up the WireShark and begin packet capture.
Type the command traceroute www.engr.uvic.ca in command line.
When the traceroute program terminates, stop the packet capture.
Download and open tracert-trace-2 in WireShark, and set the display lter
as icmp. Then answer the questions (5)-(8) in section 3.4.3 based on the trace.
21
Figure 3.1: Capture of ping packet with ICMP display lter
3.4 Discussion
3.4.1 Exploring ARP functions
Answer the following questions based on the trace le ethernet-trace-1.
1. What are the hexadecimal values for the source and destination ad-
dresses in the Ethernet frame containing the ARP request message?
2. Find the hexadecimal value for the two-byte Ethernet Frame type eld.
3. What do the bit(s) whose value is 1 mean within the ag eld?
4. Where the ARP opcode eld is located?
5. What is the value of the opcode eld within the ARP-payload part of
the Ethernet frame, in which an ARP request is made?
6. Does the ARP message contain the IP address of the sender?
22
7. Which eld contains the MAC address of the host whose corresponding
IP address is being queried, and what is it?
8. Where the ARP opcode eld is located? I.e., how many bytes from
the very beginning of the Ethernet frame does the ARP opcode eld
begin?
9. What is the value of the opcode eld within the ARP-payload part of
the Ethernet frame in which an ARP response is made?
10. What is the MAC address answered to the earlier ARP query?
11. What are the hexadecimal values for the source and destination ad-
dresses in the Ethernet frame containing the ARP reply message?
12. Why there is no ARP reply for the second ARP query (packet No.6)?
3.4.2 Analyzing IP frames
Answer following questions based on ethernet-trace-1.
1. Sketch a gure of the packet showing the position and size in bytes of
the IP header elds, as well as the values in hexadecimal. Your gure
can simply show the frame as a long, thin rectangle.
2. What are the IP and MAC addresses of the source and destination?
3. How does the value of the Identication eld change or stay the same
for dierent packets? (Hint: No.) Is there any pattern if the value does
change?
4. How can you tell from looking at a packet that it has not been frag-
mented?
3.4.3 Exploring ICMP functions
Answer following question based on ping-trace-1 and tracert-trace-2.
1. What is the IP address of the source host? What is the IP address of
the destination host?
23
2. Can you get the average RTT (Round Trip Time)? Whats that?
3. Examine one of the ping request packets. What are the ICMP type and
code numbers? What other elds does this ICMP packet have? How
many bytes are the checksum, sequence number and identier elds?
4. Examine the corresponding ping reply packet. What are the ICMP
type and code numbers? What other elds does this ICMP packet
have? How many bytes are the checksum, sequence number and iden-
tier elds?
5. What is the IP address of the source host? What is the IP address of
the destination host.
6. Is there any dierence between the request and reply packets from
ping-trace-1 and tracert-trace-2? Why so?
7. How many routers are between the source and the destination (www.engr.uvic.ca),
for the trace le? What are the sequences of these routers?
8. Can you get the average RTT times between the source host and each
router? What are they?
24
Chapter 4
Lab 4: TCP
4.1 Objective
In this lab, we rst get familiar with the format of TCP header, then
study the TCP 3-Way Handshake and reliable data transfer, followed by the
congestion control algorithm and retransmission scheme.
4.2 Introduction
TCP is the dominant transport layer protocol in the Internet. It provides
a reliable, in-order stream of data between two end-points, even if they are
connected by a network that may drop, re-order, or corrupt the packets.
TCP provides the reliable data streaming service by detecting if packets are
lost, delayed, or corrupted during transmission.
In this Lab and the following Lab, we investigate the behavior of TCP
in detail, by analyzing a trace of the TCP segments sent and received in
transferring a 300 KB le from a local computer (the client, IP address:
10.0.1.5) to a remote Web server (http://gaia.cs.umass.edu/, IP address:
128.119.245.12). The le, named alice.txt (which contains two copies of the
text of Alice in Wonderland) is stored on the client computer and is uploaded
to the server using the HTTP POST method. Here the POST method is
used in order to transfer a large amount of data from a computer to another
computer.
The procedure to transfer this le is as follows:
25
Start up Web browser on the client computer and go to http://gaia.
cs.umass.edu/ethereal-labs/TCP-ethereal-file1.html. The screen
looks like Figure 4.1.
Figure 4.1: Upload page
Use the Browse button to enter the full path name of alice.txt on the
client computer, and then press the Upload alice.txt le button to up-
load the le to the server gaia.cs.umass.edu.
Once the le has been uploaded, a new web page, which is a short
congratulation, will be transferred from the Web server to the client
and displayed in the web browser, as shown in Figure 4.2.
To transfer alice.txt and the congratulation page without error, a TCP
connection between the client and the server is established. The TCP connec-
tion completes the four operations in this real-world application as follows:
TCP connection setup.
Transfer the HTTP POST command and the le alice.txt, from the
client computer to the server gaia.cs.umass.edu.
Transfer the congratulation page from the server to the client.
26
Figure 4.2: Congratulation Page
TCP connection release.
WireShark is run on the client computer to capture the trace of the TCP
segments sent/received from/to the client computer while the le is being
transferred. We saved the trace from the TCP stream in the le tcp-trace-
1.cap. The trace tracked all of the above four actions of TCP. We use this
trace to study the TCP behaviors in this lab.
4.2.1 TCP Header Format
Every TCP segment consists of a header followed by an optional data
portion. The format of the header is dened in RFC 793, including Source
Port (16 bits), Destination Port (16 bits), Sequence Number (32 bits), ACK
(32 bits), .... In this lab, we will see the details of the TCP headers used in
this application.
4.2.2 TCP Connection Setup
Before transferring data, a TCP connection is established between the two
end systems, typically with three messages, called the three-way handshake:
27
SYN SYNACK ACK. The handshake is also used to negotiate certain
properties of the connection, e.g., the Maximum Segment Size (MSS) that
the client and server can accept, and the Selective Acknowledgement (SACK)
option. In this lab, we will see the three-way handshake procedure in the
trace tcp-trace-1.cap.
4.2.3 TCP Data Flow
Once the connection is established, the TCP sender partitions the message
from the application into segments. The MSS is used to determine how
to partition the single message so the underlying network can encapsulate
each segment into a packet without further fragmentation. The sequence
number and ACK number are used to detect packet loss, duplication, re-
order in transmission and to deliver the segments correctly and in-order to
the application in the destination host.
In this real-world application, after the connection was established, the
client computer wrote about 300KB into the data stream using the HTTP
POST command. From the applications perspective, this was sent as one
unit, or one message. However, the underlying network cannot support pack-
ets large enough to hold all 300KB of data. We will see that TCP broke this
single logical transmission into multiple segments according to MSS.
In the trace le tcp-trace-1.cap, the rst three segments are used to
establish the connection. Starting from the No.4 TCP segment, the client
began to transfer the application layer message to the server. The 4th seg-
ment contains the HTTP POST command (we will dig into the packet con-
tent eld and see this HTTP command). This segment is actually used to
transfer this HTTP command. The text le is transferred by the following
TCP segments. Here we regard both the HTTP POST command and the
le (alice.txt) together as the whole message. Therefore, we consider the 4th
TCP segment as the rst segment in the TCP connection to transfer the
message from the client to the server.
4.2.4 TCP Connection Release
The TCP connection is closed when the two end systems exchange TCP
segments with FIN bit set and ACKed by the other side. The FIN bit literally
means that no additional new data will be sent on that side of the connection.
28
The sequence of two FINs and their corresponding ACKs is the preferred
way to gracefully terminate a TCP connection. However, TCP connections
can also be terminated by setting the RESET bit. Although the RESET was
designed to be used for unrecoverable errors, it is often used in practice for
fast termination that avoids the formalities of the FIN-ACK exchanges.
In the trace le tcp-trace-1.cap, after the client acknowledged the data
of the congratulation page, the server sent a FIN indicating that it would
not be sending any additional data. The client acknowledged this FIN by
sending back the ACK. Therefore the ow in the direction from the serve to
the client is closed. The client computer could also terminate its ow to the
server by sending the FIN segments. Alternatively, the client computer sent
a RESET segment to the server to release the connection.
4.2.5 TCP Congestion Control
In TCP, congestion control provides the ability to limit the sending rate
in response to signs of network congestion. Congestion control helps the
network to recover from congestion by shrinking senders outgoing trac
and therefore avoids network congestion collapse, and at the same time tries
to achieve throughput as high as possible.
Congestion control is realized by setting the size of congestion window,
according to two strategies, the slow start and the congestion avoidance.
During the slow start phase, the congestion window increases one MSS with
each acknowledgment, and subsequently the window size is doubled every
round-trip time (RTT). During congestion avoidance, each acknowledgment
increases the congestion window by MSS
2
/congestion window size (if the
receiver sends ACK for each received packet without delay), and subsequently
the congestion window size is increased by one MSS every RTT. Slow start
phase changes to congestion avoidance phase when congestion window ex-
ceeds the slow-start threshold.
We will use the TCP segment trace le, tcp-trace-1.cap to investigate
TCP congestion control. In particular, we look at how the congestion window
evolved since the beginning of transferring the HTTP POST command and
the alice.txt le.
29
4.2.6 TCP Flow Control
TCP also provides ow control or the ability to limit the sending rate
to avoid a fast sender over-running a slow receiver. To provide a reliable
service, a TCP receiver cannot deliver data that it received out of order
to the waiting application. Therefore, the TCP receiver typically allocates a
xed amount of buer space to store both out-of-order data and data waiting
for the application to fetch. If the TCP receiver runs out of buer space to
hold the incoming data, then it has no choice but to drop the out-of-order
data packet even if it is error-free.
The receiver advertises its available buer in each acknowledgment. The
receivers advertised window eld is used to inform the sender how much
room is left for the incoming data. Then in the sliding-window based ow
control, the sender chooses the minimum of the receiver window and the
congestion window to be the size of the sliding window in order to make sure
that the receiver will not run out of buer space.
In this subsection, we will still use the TCP segment trace le, tcp-trace-
1.cap, to exam TCP ow control. We will see that the receiver window took
eect and throttled the sender even though the congestion window continued
to grow.
4.2.7 Retransmission in TCP (Optional)
We learned that TCP provides reliable data transmission over an unre-
liable network by relying on feedback from the receiver to detect loss and
responding to packet loss with retransmissions. TCP uses two kinds of in-
dications of packet losses, timeout and duplicated acknowledgment (which
is regarded as an early indication of packet loss and causes the fast retrans-
mission instead of waiting until timeout). The TCP sender must maintain a
copy of the data it sent in case retransmission is needed, so it must store the
data until the corresponding acknowledgment is received.
However, in the trace tcp-trace-1.cap, all the packets were received
correctly the rst time and thus there was no retransmission. In order to
investigate the TCP retransmission scheme, we are going to analyze another
trace of TCP connections, tcp-trace-retransmission.cap [3], in which re-
transmission does occur.
The trace, tcp-trace-retransmission.cap, was taken on a private net-
work [3]. A desktop PC and a laptop were connected via a wireless router.
30
The laptop was connected via a wireless interface and specically placed so
as to interfere with a strong signal. The IP addresses of the desktop and the
laptop are, respectively, 192.168.0.100 and 192.168.0.102. The desktop sent
a le (about 40K bytes) to the laptop using TCP. The TCP port number for
the desktop is 4480, and 5001 for the laptop. The experiment conguration
is shown in Figure 4.3. WireShark was run on the sender, i.e., the desktop,
while the le was being transferred to capture the TCP segments exchanged.
The TCP connection trace was saved in le tcp-trace-retransmission.cap.
Figure 4.3: Network Conguration
In this lab, we will take a look at both the fast retransmission and the
timeout retransmission using this trace le.
4.3 Procedures
Note: You will answer a set of questions by exploring the trace le tcp-
trace-1.cap and tcp.analysis.retransmission.cap. Whenever possible,
when answering a question you should provide the information of the packet(s)
within the trace that you used to answer the question asked. The informa-
tion includes the Packet No., the name(s) and value(s) of the packet eld(s)
that you use to answer the questions.
4.3.1 TCP Header Format
Download the traces folder from the lab website.
31
Open the captured trace le named tcp-trace-1.cap with WireShark.
Now what you should see is a series of TCP segments sent between the
client and the server gaia.cs.umass.edu.
Since this lab is about TCP rather than HTTP, change WireSharks
Packet List Pane window so that it shows information about the TCP
segments containing the HTTP messages. To do this, in WireShark,
select Analyze Enabled Protocols. Then uncheck the HTTP box and
select OK.
Select the rst packet and explore the details of the TCP segment using
the packet details pane and the packet bytes pane.
Select the Transmission Control Protocol item in the Packet Details
Pane then the content of the header is highlighted in the Packet Bytes
Pane.
Answer the related questions in section 4.4.1.
4.3.2 TCP Connection Setup
Find the initial three-way handshake in the trace le. (Hint: You
should see the SYN segment sent from the client to gaia.cs.umass.
edu, and also the SYNACK segment being returned.)
Answer the related questions in section 4.4.2.
4.3.3 TCP Data Flow
Check the HTTP POST command. Select the 4th segment in the
Packet List Pane. Select the Data item in the Packet Details Pane
and the content of the data carried by this segment is highlighted in
the Packet Bytes Pane. You should nd a POST and other HTTP
command information within its Date eld.
Set time reference. In order to make the following analysis easier, set
time reference to the 4th packet. Choose the Time Reference items in
the Edit menu, or from the pop-up menu of the Packet List Pane.
32
Note: Now the 4th packet becomes the starting point for all subse-
quent packets. The time values of all the following packets are calcu-
lated relative to this packet.
Set the time display format as microseconds. Choose the Time Display
Format in the View menu. Then select Seconds Since Beginning of
Capture and Microseconds.
Answer the related questions in section 4.4.3.
4.3.4 TCP Connection Release
Find the segments used to release the connection between the client
and the server.
Answer the related questions in section 4.4.4.
4.3.5 TCP Congestion Control
Open the captured trace le named tcp-trace-1.cap with WireShark.
Answer the related questions in section 4.4.5.
4.3.6 TCP Flow Control
Open the captured trace le named tcp-trace-1.cap with WireShark.
Answer the related questions in section 4.4.6.
4.3.7 Retransmission in TCP (Optional)
Open the captured trace le named tcp-trace-retransmission.cap
with WireShark.
List retransmissions. Search for retransmissions with the display l-
ter tcp.analysis.retransmission. Applying this lter, you should see 9
retransmissions in the trace.
Answer the related questions in section 4.4.7.
33
4.4 Discussion
4.4.1 TCP Header Format
1. Write down the TCP header content in hexadecimal format (in the
packet bytes pane). Dissect the TCP header and indicate the value of
each eld in the header. Annotate the hexadecimal content to explain
your answer.
2. What are TCP port numbers used by the client computer (source) and
the server (destination) when transferring the le to gaia.cs.umass.
edu? How did the client computer determine the port numbers when
it wanted to set up a TCP connection to the server?
3. What is the maximum header length? Given the value of the Header
Length eld, how to calculate the length of the head in unit of bytes?
Verify your answer using the rst TCP segment in the trace le.
4. (Optional) How does TCP calculate the Checksum eld? What is the
pseudoheader format? Write down the pseudoheader of the ow from
the client to the server in hexadecimal format. Verify the Checksum
value in the rst TCP segment in the trace le.
4.4.2 TCP Connection Setup
1. Which segments are the initial three-way handshake in the trace le?
How do you determine this?
2. What is the actual initial sequence number in each direction (in hex-
adecimal format)? How did the client and the server determine these
values?
Note: WireShark displays the relative sequence number. You should
select the Sequence Number eld in the header, the actual value is
highlighted in the Packet Bytes Pane.
3. What is the value of the acknowledgment number in the SYNACK
segment? How did gaia.cs.umass.edu determine that value?
34
4. What are the values of the sequence number and the acknowledgment
number in the third ACK segments in the three-way handshake? How
did the client computer determine these values?
5. How did the client and the server announce the maximum TCP payload
size that they were willing to accept? What are the values and why
did they choose these values?
6. Is there data sent in the SYN, SYNACK, and ACK segment? How do
you determine this?
4.4.3 TCP Data Flow
1. Beginning with the 4th segment, what are the sequence number, ac-
knowledgment number, data length, and the time of the segment sent/received
from/to the client computer of the 4th, 5th, 6th, ..., 15th segments in
the TCP connection? Fill out Table 4.1 for the data ow from the
client computer to the server.(Note: list both the actual value and
relative value of the sequence number and acknowledgment number.)
Data Segments
10.0.1.5 --> 128.119.245.12
ACK Segments
128.119.245.12 --> 10.0.1.5 Packet
No.
Seq. No./
Relative Seq. No.
Data
Length
Time (s)
Ack. No./
Relative Ack. No.
Data
Length
Time (s)
4
5
6
7
8
9
10
11
12
13
14
15
Table 4.1: TCP segment exchange table
2. What are the segments acknowledged by the packet 6, 9, 12, and 15, re-
spectively? (Hint: acknowledgment number is the next byte expected,
35
so it actually acknowledges the bytes before the acknowledgment num-
ber.)
3. Given the dierence between when each TCP segment was sent, and
when its acknowledgment was received, what is the RTT value for each
of the segments which have been acknowledged before the 15th seg-
ment?
4. (Optional) What is the Estimated RTT value after the receipt of each
ACK? Assume that the value of the Estimated RTT is equal to the
measured RTT for the rst segment, and then is computed using the
Estimated RTT equation for all subsequent segments. (Hint: Com-
pare your calculation with the statistics analysis of TCP stream by
WireShark. See Section 3.5).
5. In the trace le, how did the sequence number of the packets from
the server to the client change? Why? (Hint: When transferring the
alice.txt le, the server was only a receiver and did not send any data
to the client.)
6. (Optional) At the end of the trace le, nd the TCP segments used
by the server to transfer the congratulation webpage to the client com-
puter? How do you determine this?
7. (Optional) Are there any retransmitted segments in the trace le?
What do you check for (in the trace) in order to answer this question?
4.4.4 TCP Connection Release
1. Which packets were used to close the data ow from the server to the
client? How do you determine this? (Hint: two segments are involved
in the FIN-ACK sequence.)
2. Which packets were used to close the data ow from the client to the
server? How do you determine this?
3. (Optional) In the FIN segment, what is the sequence number? In the
corresponding ACK segment, what is the acknowledgment number?
How did the client determine this number?
36
4.4.5 TCP Congestion Control
1. Exam the 4th to 15th TCP segments and take a reference to the Table
in Question 1 of Section 4.4.3. Can you nd a pattern of the number of
segments sent from the client and from the server gaia.cs.umass.edu?
Why did the TCP data ow have such a pattern?
2. What is the initial size of congestion window? How do you determine
this? What is the size of congestion window when the segment 5, 8, 11
and 14 were sent out?
3. In the lecture we have learned that the congestion window doubles
its size in every RTT in the slow start phase. Beginning with the 4th
packet, what is the size of the congestion window and which packet were
inside the congestion window (i.e., these packets could be sent) during
the rst RTT? What is the size of the congestion window and which
packet were inside the congestion window during the second RTT? How
about the third RTT? Give the segment numbers.
4. When did the senders congestion control change from the slow start
phase to the congestion avoidance phase? Give the segment number
and the time. How do you determine this?
5. What is the threshold between the slow start and congestion avoidance?
(Hint: the size of congestion window when TCP change from slow start
phase to congestion avoidance phase.)
4.4.6 TCP Flow Control
1. Exam the 179th segment in the trace le, why did the sender stop
sending more segments? What is the size of receiver window advertised
by the receiver at this moment? How do you determine this?
4.4.7 Retransmission in TCP (Optional)
1. Segment 12 is the rst retransmission. What is it in the segment that
identies the segment as a retransmission? (Hint: the sequence number
has been used by a previous packet.) Which segment was segment 12
retransmitted for?
37
2. Segment 12 is a fast retransmission, which should be triggered by triple-
duplicated-acknowledgment. Find the three acknowledgments which
triggered the fast retransmission of segment12. (Hint: in order to trig-
ger a fast retransmission, the duplicated acknowledgments should ac-
knowledge the same acknowledgment number, which is the sequence
number of the fast retransmission.)
3. Is segment 44 a fast retransmission or timeout retransmission? How
do you determine this? (Hint: whether the sequence number in the
segment has been acknowledged for three times or not.)
38
Bibliography
[1] Ethereal Labs, http://www-net.cs.umass.edu/ethereal-labs
[2] Wikipedia.org, http://en.wikipedia.org/wiki/HTTP
[3] Jeanna Matthews, Computer Networking: Internet Protocols in Action,
John Wiley & Sons, Inc., Dec. 2004.
[4] Andrew Tanenbaum and David Wetherall, Computer Networks 5/E,
Prentice Hall, Oct. 2010
[5] James F. Kurose and Keith W. Ross. 2009. Computer Networking: A
Top-Down Approach (5th ed.). Addison-Wesley Publishing Company,
USA.
39