You are on page 1of 12

9/7/2009

Chapter 9 € The history and future of Ethernet (9.1 & 9.2)


€ Ethernet frames and error detection (9.3)
SCR 2213 Network Communication
€ Addressing (9.3, 9.7)
2009/2010 Sem
2009/2010 Sem I
By Marina Binti Md Arshad € Media Access Control (9.4)
Universiti Teknologi Malaysia
€ Physical cabling (9.5)
€ Hubs and switches (9.6)

1 Marina MA. SKK, FSKSM, UTM. 2009 2

€ Performs the DLL functions of framing, addressing,


€ Ethernet is a combination of Data Link layer media access control, and error detection.
software and Physical layer hardware.
€ Has physical hardware (media and devices) that
€ Data Link layer controls the Physical layer.
layer perform the functions of the Physical layer.

€ Ethernet is a standardized protocol, and therefore,


has well-defined rules for how it operates and well-
defined rules for the structure of the Data Link
layer frames and Physical layer signals it uses.

€ Ethernet devices are called “stations”, which is


synonymous with “node” from the Data Link layer
chapter or “host” from the TCP/IP chapters. These
terms are often used interchangeably.

Marina MA. SKK, FSKSM, UTM. 2009 3 Marina MA. SKK, FSKSM, UTM. 2009 4

1
9/7/2009

€ Because Ethernet is so popular, there are many € Has two main parts:
different types of physical cabling and network
ƒ LLC sublayer links Ethernet to the upper layers
devices (hubs and switches) that it can use.
ƒ MAC sublayer controls the hardware

Marina MA. SKK, FSKSM, UTM. 2009 5 Marina MA. SKK, FSKSM, UTM. 2009 6

KEY CONCEPT #1
€ The predecessor of Ethernet had its roots in
Alohanet in Hawaii in 1970 (a radio-based network).

€ Ethernet was developed in 1973 by Robert Metcalfe.

Metcalfe’s
Ethernet
Sketch

Marina MA. SKK, FSKSM, UTM. 2009 7 Marina MA. SKK, FSKSM, UTM. 2009 8

2
9/7/2009

€ Was published as an open standard in 1980 by an


industry consortium and then adopted as a formal
€ Several new versions of Ethernet are being
standard by IEEE as 802.3.
developed.
€ The original
g version used coaxial cable but todayy it € The current versions of 100 Mbps
Mbps, 1 Gbps,
Gbps and 10
uses twisted pair and fiber optic cable. Gbps are well-established and new versions that run
at 40 Gbps, 100 Gbps and 160 Gbps are under
development.
€ There used to be several competing protocols
used in LANs (e.g., Appletalk, Token Ring). € They will challenge the current technologies used in
WANs because they will be much faster and easier
€ Today,
Today for all practical purposes,
purposes there is only to integrate into LANs and backbone networks that
Ethernet. already use Ethernet.
€ Many experts think that within 10 years, Ethernet
€ Why do you think Ethernet won? will be dominant networking technologies in all
parts of the network.
Marina MA. SKK, FSKSM, UTM. 2009 9 Marina MA. SKK, FSKSM, UTM. 2009 10

Source: “Metro Ethernet Case Study” www.cisco.com, 2006.

“Until 2005, the Cisco San Jose campus used four separate SONET
OC-3 circuits to connect to the Internet. As demand grew and
business Cisco
as one of its Internet providers went out of business,
decided to move to a metro Ethernet WAN. Cisco wanted an
KEY CONCEPT #2
initial bandwidth of 200 Mbps, with incremental upgrades to
1000 Mbps. They installed 1000 Mbps Ethernet over fiber into
AT&T’s Metro Ethernet service. The circuit uses two separate
cables to ensure redundancy in then event that one fails.
Although the circuit is capable of 1000 Mbps, Cisco only wanted
to purchase a 200 Mbps connection initially, so they installed a
Cisco switch to limit traffic so that it does not exceed 200
Mbps. The cost was less than the original configuration, and
provided better speed. Traffic has continued to grow, but the
use of Ethernet enabled Cisco to purchase small incremental
increases in service over the same circuit, without having to lay
entirely new OC-3 or OC- 12 cable, which would have been
needed if it had continued to use an OC-3 solution.”
Marina MA. SKK, FSKSM, UTM. 2009 11 Marina MA. SKK, FSKSM, UTM. 2009 12

3
9/7/2009

€ Ethernet frame is the most commonly used Data


Link layer protocol.

€ Three major versions of Ethernet frames in use


today
1. 7 byte Frame Preamble
€ Class Discussion: 2. 1 byte Start of Frame Delimiter
ƒ What are the three major parts of an Ethernet 3. 6 byte Destination MAC Address
frame?
4
4. 6 byte Source MAC Address
ƒ What are the major parts of the header and trailer? 5. 2 byte Frame Length or Encapsulated Protocol Type
What fields do they contain?
6. 46 to 1500 byte Data (Encapsulated Packet) plus Padding if
ƒ What is the preamble used for? required
7. 4 byte Frame Check Sequence (CRC Checksum)
Marina MA. SKK, FSKSM, UTM. 2009 13 Marina MA. SKK, FSKSM, UTM. 2009 14

€ Ethernet uses a technique called Cyclic Redundancy


Check (CRC) to calculate FCS, which essentially
€ Itis important in “fragile” environments such as
divides the data by a preset number (using binary
wireless LANs, but is less important in LAN division) and uses the remainder as the 32-bit FCS.
environments because the probability of an
error is unlikely. € If it finds an error, Ethernet simply discards the
frame, so it is up to TCP to recognize that data was
€ How does Ethernet performs error checking? lost and re-send it.

€ Sender performs a calculation on the frame and


sends the result of the calculation in the FCS in
the message trailer; the receiver performs the
same calculation and compares the result to the
value in the FCS.

Marina MA. SKK, FSKSM, UTM. 2009 15 Marina MA. SKK, FSKSM, UTM. 2009 16

4
9/7/2009

€ Uses a multi-access shared media topology, so there


are several nodes (i.e., devices or computers) on
the same physical circuit.
KEY CONCEPT #3 € Each node must have a unique address.
€ When a device on a multi-access physical circuit
transmits, every node on the circuit receives the
message and looks at the destination address to see
if the frame is addressed to it.
€ If the frame is addressed to the node, the node
processes it; if not, the node discards it.
€ Analogy: having a discussion using a radio in taxis,
police car etc.

Marina MA. SKK, FSKSM, UTM. 2009 17 Marina MA. SKK, FSKSM, UTM. 2009 18

€ How long are Ethernet addresses?


€ For historical reasons,
Ethernet addresses are
usually written in
hexadecimal (base 16)
with each byte expressed
as a two-letter code.

€ Therefore, 6 bytes are


written as a set of six
two-letter codes.
€ What are the implications for security?
€ Wireshark demonstration.

Marina MA. SKK, FSKSM, UTM. 2009 19 Marina MA. SKK, FSKSM, UTM. 2009 20

5
9/7/2009

€ Most common type of addressing is called unicast


€ Used when there is a group of computers on the
addressing; the frame is addresses to just one node.
same physical circuit that should receive the same
data (e.g., online gaming, video conferencing).
€ All nodes on the shared media circuit receive everyy
frame and examine the address. The nodes choose € Analogy – mailing list, where a user can send one
to discard frames that are not addresses to them. email to one address and it is sent to all
subscribers.
€ A group of address is defined and all computers that
join the group process frames addressed to the
group address as well as their own unique address.
€ Reminder: on a shared-media circuit every frame is
received by every node and each node is responsible
to determining whether or not the frame is
addressed to it.
Marina MA. SKK, FSKSM, UTM. 2009 21 Marina MA. SKK, FSKSM, UTM. 2009 22

€ One frame is sent to all nodes on the physical circuit.

€ The broadcast address is FF-FF-FF-FF-FF-FF-FF (i.e,


ll 1’s)
all 1’ )

€ Used for special purposes such as finding addresses.

Marina MA. SKK, FSKSM, UTM. 2009 23 Marina MA. SKK, FSKSM, UTM. 2009 24

6
9/7/2009

€ Imagine that you are a router, standing on the


front of a house and have message to deliver to
€ Ethernet address resolution is the process of finding
an Ethernet MAC address given an IPv4 address.
one the occupants. You know the name (IP
address), but not the who the name belongs to
€ Locall address
L dd resolution
l i means fi
finding
di an E Ethernet
h (Eth
(Ethernett MAC address).
dd )
MAC address on the same physical circuit that your
computer is on.
How would you find the letter’s owner?
€ When devices are first turned on, they do not know
the Ethernet MAC addresses of any other computers €A node that needs to know the Ethernet MAC
or devices on their circuit. address of another node sends a broadcast
message – a message addressed to all computers
€ When they need to send a message to another – using a special protocol called Address
computer or device, they need to find its MAC Resolution Protocol (ARP).
address because there could be many computers on
the same circuit.
Marina MA. SKK, FSKSM, UTM. 2009 25 Marina MA. SKK, FSKSM, UTM. 2009 26

€ The ARP request says, “Whoever has this IP address, I need the I heard that broadcast.
please send me your Ethernet MAC address.” Ethernet address The message is for me.
of 176.16.3.2. Here is my Ethernet
€ All nodes receive this message and pass it to their address.
ARP software.

172.16.3.1 172.16.3.2
€ If they do not match, the ARP software takes no
action.
IP: 172.16.3.2 = ???

€ If the
h IP address
dd iin the
h ARP request matches
h the
h IP: 172.16.3.2
Ethernet: 0800.0020.1111
node’s IP address, the ARP software sends an
IP MAC
ARP response that says, “My IP address is ____
and my Ethernet MAC address is ____.”

Marina MA. SKK, FSKSM, UTM. 2009 27 Marina MA. SKK, FSKSM, UTM. 2009 28

7
9/7/2009

€ How does the node that receives the ARP


request knows the Ethernet MAC address of the
node sending the ARP request so it can send the
ARP response to it?

€ How efficient is it to send an ARP request every


time a node needs to know an Ethernet MAC 1. arp –a
address? 2. ipconfig/all
3. ping 10.60.80.1 
€ How would you make this more efficient? (gateway’s IP add)
4. arp ‐a

Marina MA. SKK, FSKSM, UTM. 2009 29 Marina MA. SKK, FSKSM, UTM. 2009 30

€ It is the process of finding addresses on physical


circuits outside of the physical circuit the node is on.
€ It is uncommon and usually occurs in older
implementations
p or in cases of improper
p p configuration
g .
KEY CONCEPT #4

Marina MA. SKK, FSKSM, UTM. 2009 31 Marina MA. SKK, FSKSM, UTM. 2009 32

8
9/7/2009

€A network has both a € Because Ethernet uses a shared multi-access


topology, there must be a way to control when
physical and logical devices transmit.
topology.
€ The process Ethernet uses is CSMA/CD.
€ Ethernet has a € Four steps:
logical bus topology
1. Listen to ensure no other device is transmitting.
and a physical star 2. If a device is transmitting, wait until it finished [plus
topology. another length of time called an interframe gap (the
time required to send 96 bits)]; otherwise if no other
device is transmitting.
transmitting Then transmit
transmit.
€ Terms
3. Listen while transmitting to ensure there are no
ƒ Bus = multi-access collisions.
ƒ Star = point-to-point 4. If there is a collision, then transmit a jamming signal
(to indicate a collision has occurred), then wait a
random amount of time before returning to step 1.

Marina MA. SKK, FSKSM, UTM. 2009 33 Marina MA. SKK, FSKSM, UTM. 2009 34

€ There is a minimum length of an Ethernet frame


depending upon the version of Ethernet in use.
€ Electricity moves very quickly through the physical
cable, but that it is not instantaneous.
€ When a device transmits, the electricity takes time
to move through the cable and does not instantly
reach the other devices on the cable; but it travels
very, very fast.
€ Note: Devices listen while they transmit to
determine if a collision occurs,, but once theyy stop
p
transmit, they stop listening for collisions.
€ We must ensure that there is time for the electricity
involved in the collision to make its way from the
site of the collision to the sending computer before
the sending computer stops sending.
Marina MA. SKK, FSKSM, UTM. 2009 35 Marina MA. SKK, FSKSM, UTM. 2009 36

9
9/7/2009

€ The minimum length of a frame depends upon how


fast devices can transmit, because faster devices
transmit frames faster than slower devices and thus
need longer slot times to ensure that collisions far
away can be detected before transmission ends.
KEY CONCEPT #5

Marina MA. SKK, FSKSM, UTM. 2009 37 Marina MA. SKK, FSKSM, UTM. 2009 38

€ There are many types of Ethernet cables that


run at different speeds.

€ Someversions of Ethernet are half-duplex and


some are full-duplex (2 x 500 Mps = 1 Gbps)

€ Different
types of cable provide different
maximum distances, which is one of the
important reasons for choosing between fiber
cables and twisted pair cables (UTP).

STORY: 100mb versus 1000mb Ethernet

Marina MA. SKK, FSKSM, UTM. 2009 39 Marina MA. SKK, FSKSM, UTM. 2009 40

10
9/7/2009

€ Legacy Ethernet used hubs to provide a shared


multi-access physical circuit, which meant that the
total capacity of the circuit was shared among all
the devices connected to it.
KEY CONCEPT #6
€ When a hub receives a frame, it sends it to all of
the devices connected to it.
€ The collisions were common and transmission was
slow because devices had to wait for each other.
€ For example, in a 10-computer network using a 100
Mbps
b h hub,
b all
ll devices
d would
ld share
h the
h same 100
00
Mbps circuit.
€ Each device would have average only 10 Mbps if all
devices wanted to transmit or receive.

Marina MA. SKK, FSKSM, UTM. 2009 41 Marina MA. SKK, FSKSM, UTM. 2009 42

€ Most Ethernet networks use switches, not hubs.


€ Switches provide each device with a point-to-point
circuit from the device to the switch.
€ When a switch receives a frame, it reads the MAC add on
frame and forwards it just to the one device to which the
frame is addressed.
Marina MA. SKK, FSKSM, UTM. 2009 43 Marina MA. SKK, FSKSM, UTM. 2009 44

11
9/7/2009

€ The frame is briefly stored in the switch’s memory


(so it can process the error check information) and
then retransmitted to the destination devices € If all the computers are sending messages
(“store and forward”) through the switch to a router connected to the
€ If that circuit is in use then it stores the frame in switch by a single 100 Mbps circuit (e.g.,
(e g to the
memory and forwards it once the circuit is free. Internet), then the switch provides no
advantage over the hub because the single 100
€ This means that the time it takes for the switch to
transmit the frame (the latency through the switch) Mbps circuit to the router must still be shared
is relative to the frame size – the larger the frame among all computers.
size, the longer the delay through the switch.
€ The bottleneck has moved from the hub to the
€ Therefore, in the same 10-computer network using
Therefore circuit-leaving
circuit leaving switch.
switch
a 100 Mbps switch, each computer receives the
same full 100 Mbps capacity. € Nonetheless, having a switch is still an
€ If the connection is full duplex, then the capacity is advantage for traffic that originates and
effectively doubled. terminates in the Local Area Network.

Marina MA. SKK, FSKSM, UTM. 2009 45 Marina MA. SKK, FSKSM, UTM. 2009 46

€ When switches receive a frame on a circuit, they € Ethernet is the most commonly used LAN technology
check it for errors and if there are no errors, the and is moving into the WEN.
switch reads the MAC address in the frame and then
€ Ethernet can be used with many different types of
forwards the frame to the device. physical cable.
€ How do switches know which device is connected on GROUP DISCUSSION
which circuit?
€ Imagine you have just turned on your computer. You
€ How do switches learn and make entries in the MAC open a Web browser and connect to your home
table all byy themselves? p g
page.

€ What does a switch do if it receives a frame with an € Discuss the series of messages (HTTP, IP, ARP, etc.)
address not in its MAC table? that flow to and from their computer in order to
display their home page. Your group should produce
€ HUB AND SWITCH ACTIVITY the sequence of messages.

Marina MA. SKK, FSKSM, UTM. 2009 47 Marina MA. SKK, FSKSM, UTM. 2009 48

12

You might also like