Lecture 4:

Datalink layer
 Functionalities:
 Encapsulation, addressing
 Error detection and correction
 Flow control
 Media access control

1
Overview of Data link layer

2
Link layer: introduction
terminology: mobile network
 hosts and routers: nodes national or global ISP

 communication channels that

connect adjacent nodes along
communication path: links
• wired
• wireless
• LANs
 layer-2 packet: frame, datacenter
encapsulates datagram network

link layer has responsibility of

transferring datagram from one node enterprise
network
to physically adjacent node over a link

Link Layer: 6-3

Datalink layer in Layer
architecture
Application Media independent
sub-layer
Transport

Network LLC
(Logical Link Control)
Data-link
MAC
Physical (Media Access Control) Media dependent
sub-layer

802.2 LLC

802.3 802.4 802.5 802.11 ….. 802.16

Ethernet Token Bus Token Ring Wi-Fi Wi-Max
4
IEEE 802.x series
Overview of Datalink services
(functionalities)

Framing Flow control

Media Access Control

Addressing Error control

Datalink layer

5
Link layer: context
 datagram transferred by transportation analogy:
different link protocols over  trip from Princeton to Lausanne
different links: • limo: Princeton to JFK
• e.g., WiFi on first link, Ethernet • plane: JFK to Geneva
on next link • train: Geneva to Lausanne

 each link protocol provides  tourist = datagram

different services  transport segment =
• e.g., may or may not provide communication link
reliable data transfer over link  transportation mode = link-layer
protocol
 travel agent = routing algorithm

Link Layer: 6-6

Link layer: services
 framing, link access: …
• encapsulate datagram into frame, adding …
header, trailer
• channel access if shared medium
• “MAC” addresses in frame headers identify
source, destination (different from IP
address!)
 Media access control:
• If the nodes in the network share common
media, a Media access control protocol is
required

Link Layer: 6-7

Link layer: services (more)
 flow control: …
• pacing between adjacent sending and …
receiving nodes
 error detection:
• errors caused by signal attenuation, noise.
• receiver detects errors, signals
retransmission, or drops frame
 error correction:
• receiver identifies and corrects bit error(s)
without retransmission
 half-duplex and full-duplex:
• with half duplex, nodes at both ends of
link can transmit, but not at same time

Link Layer: 6-9

Where is the link layer
implemented?
 in each-and-every host
 link layer implemented in
network interface card (NIC) or
on a chip application
transport
cpu memory
• Ethernet, WiFi card or chip network
link

• implements link, physical layer host bus

(e.g., PCI)

 attaches into host’s system

controller
link
physical
buses physical

 combination of hardware,
software, firmware
network interface

Link Layer: 6-10

Interfaces communicating
application application
transport transport
cpu memory memory CPU
datagram network network
link link

linkh datagram controller controller

link
datagram
link
physical physical
physical physical

sending side: receiving side:

 encapsulates datagram in frame  looks for errors, reliable data
 adds error checking bits, reliable data transfer, flow control, etc.
transfer, flow control, etc.  extracts datagram, passes to
upper layer at receiving side

Link Layer: 6-11

Identifier: MAC address
MAC address: 48 bit, organized by IEEE
Each port is assigned one MAC
Cannot be changed
Physical address
No hierarchical system, flexible
MAC Address is unchanged when changing networks
Broadcast address in LAN:
FF-FF-FF-FF-FF-FF

12
Error control
Error detection
Error correction

13
Principle of error detection

EDC: error detection and correction bits (e.g., redundancy)

D: data protected by error checking, may include header fields

datagram datagram Error detection not 100%

otherwise reliable!
all  protocol may miss
bits in D’ N
OK
?
detected
error
some errors, but
d data bits rarely
D EDC D’ EDC’  larger EDC field yields
better detection and
bit-error prone link correction

Link Layer: 6-14

 Parity checking

single bit parity: two-dimensional bit parity:

 detect single bit  detect and correct single bit errors
errors
0111000110101011 1
row parity
d1,1 ... d1,j d1,j+1
d data bits d2,1 ... d2,j d2,j+
parity ... ... ... 1. . .
bit di,1 ... di,j di,j+1
column
parity di+1,1 ...
Even parity: set parity di+1,j di+1,j+
bit so there is an even 1
number of 1’s
no errors: 1 0 1 0 1 1 detected 10101 1
11110 0 and 10110 0 parity
error
correctable
01110 1 single-bit 01110 1
00101 0 error: 00101 0
* Check out the online interactive exercises for more parity
error
examples: http://gaia.cs.umass.edu/kurose_ross/interactive/

Link Layer: 6-15

Checksum
Goal: detect errors (i.e., flipped bits) in transmitted segment
sender: receiver:
 Divide data to n-bit  Divide data to n-bit segments
segments  Calculate the sums of segments. If
having overflow bits, add them to the
 Calculate the sums of results
segments. If having  Add the received checksum with the
overflow bits, add them to results
the results  Check the final outcome
 checksum: addition (one’s • Contains 0 - error detected
complement sum) of segment • Only 1 - no error detected. But maybe
content errors nonetheless?

Link Layer: 6-16

 checksum: Example
Data: 0011 0110 1000
checksum 4 bit:
0011
+ 0110
1000
Overflow 10001
bit 1
0010
Reverse bit 🡪 checksum: 1101
Transfering bits: 0011 0110 1000 1101

17
 checksum: receiving
Receiving bit: 0011 0110 1000 1101
Checking:
0011
0110
+ 1000
1101
Overflow
11110
bit 1
1111 🡪 No errors

18
Cyclic Redundancy Check (CRC)
 more powerful error-detection coding
 D: data bits (given, think of these as a binary number)
 G: bit pattern (generator), of r+1 bits (given)
r CRC bits
d data bits
D R bit pattern
<D,R> = D*2r XOR R formula for bit pattern

goal: choose r CRC bits, R, such that <D,R> exactly divisible by G (mod 2)
• receiver knows G, divides <D,R> by G. If non-zero remainder: error detected!
• can detect all burst errors less than r+1 bits
• widely used in practice (Ethernet, 802.11 WiFi)

Link Layer: 6-19

 Cyclic Redundancy Check (CRC):
example
We want: G 1 0 1 0 1 1
D.2r XOR R = nG 1 0 0 1 1 0 1 1 1 00 0 0
or equivalently: 1 0 0 1
D.2r = nG XOR R 1 0 1 D * 2r
0 0 0
or equivalently: 1 0 1 0
1 0 0 1
if we divide D.2r by G, want 1 1 0
remainder R to satisfy: 0 0 0
1 1 0 0
R = remainder [ D 2 ]
. r
1 0 0 1
G 1 0 1 0
1 0 0 1
0 1 1
R
* Check out the online interactive exercises for more examples: http://gaia.cs.umass.edu/kurose_ross/interactive/

Link Layer: 6-20

CRC under polynomial form

 1011 > x3 +x +1
 Example of some CRC using in the pratice:
 CRC-8 = x8 + x2 + x + 1
 CRC-12 = x12+x11+x3+x2+x
 CRC-16-CCITT = x16 + x12 + x5 + 1
 CRC-32 = x32 + x26 + x23 + x22 + x16 + x12 + x11 + x10 + x8 + x7 + x5
+ x4 + x2 + x + 1
 The longer G is, the more possible that CRC detects
errors.
 CRC is widely used in the practice
 Wi-fi, ATM, Ethernet…
 Operation XOR is implemented in hardware
 Capable to detect less than r+1 bits errors
22
 CRC – Example

Frame : 1101011011
Generator : G(x) = x4 + x + 1  P = 10011
Dividend : Fk = 11010110110000
R = Fk mod P = 1110
Send : 11010110111110

23
CRC – Example
11010110110000 10011
10011
010011
10011
0000010110
10011
0010100
10011
001110 🡪 Remainder:CRC
24
CRC – Check 11010110111110
11010110111110 10011
10011
010011
10011
0000010111
10011
0010011
10011
000000 🡪 No errors

25
CRC – Check 11010010111110
11010010111110 10011
10011
010010
10011
000011011
10011
010001
10011
00010110
10011
00101 🡪 not 0 🡪 errors
26
Reaction when errors detected
 Objective: assure that data  Popular techniques:
are transmitted correctly  Error detection (as we seen)
even though the chanel is  Acknowledgement/confirmation
 Retransmis after timeout
not realiable.
 Retransmis after a clear
 Condition confirmation that frame is not
arrived
 Data fram must be
transmitted correctly  ARQ technique: (automatic
repeat request). There are 3
 Negligible transmission versions:
delay.  Stop and Wait ARQ
 Possible errors  Go Back N ARQ
 Selective Reject ARQ
 Whole frame loss
 Similar to techniques used in
 Error frame flow control.
 Loss of error warning
message
27
Stop-and-wait ARQ
Normal
Sender Receiver

send pkt0

pkt0 is
OK
rcv ACK
send pkt1

pkt1 is
corrupted

rcv NAK
resend pkt1

Time

28
 Stop-and-wait ARQ
ACK/NAK errors
Sender Receiver

Resend send pkt0

Repetitive packets? pkt0 is
OK
Add Seq. No. rcv ACK
send pkt1

pkt1 is
OK

rcv sth corrupted!

resend pkt1 rcv pkt1
duplicate,
Time discard it

29
 Stop-and-wait ARQ
Sender Receiver
Not using NAK
send pkt0

pkt0 is
OK
rcv ACK0
send pkt1

pkt1 is
OK
rcv ACK1
send pkt2
pkt2 is corrupted

rcv ACK1
resend pkt2

Time

30
Stop-and-wait ARQ: Lost ACK?
Data and ACK can be lost
Can’t receive ACK?
How to resend data?
Mechanism
Packets will be numbered and sender automatically resends
packets until receiving ACK of packets
Resending after time-out
How long of timeout?
At least 1 RTT (Round Trip Time)
Each sent packets should have 1 timer
Received packets but lost ACK?
Duplicate packets
Receiver will remove duplicate ones

31
ARQ illustration

32
ARQ illustration (cont.)

33
Flow control

34
What is flow control
 Goal: Make sure that the sender does not overload the receiver
 Why overloading?
 The receiver stores data frame in buffer.
 Receiver performs some processing before deliver data to the upper level.
 Buffer could be full, leaving no space for receiving more frame  some data
fram must be dropped.
 Problem of errors in transmission is excluded
 All frames are transmitted to correct receiver without error
 Propagation time is small and could be ignored
 Solution
 Stop-and-wait mechanism
 Sliding window mechanism

35
Stop-and-wait
 Principles
 Transmitter sends a single frame

 Receiver receives the fram, process and then informs

the transmitter that it is ready to receives next frames
by a clear acknowledgement (ACK).
 Transmitter waits until reception of the ACK before
sending next frames.

36
Stop-and-wait
transmitter
Emetteur receiver
Récepteur

Packet
Paquets Paquets
Packet

frame
Trâme

Ack

Trâme
frame

frame
Trâme
Trâme

37
38
Stop-and-wait
 Advantage
 Simple, suitable for transmission of big size
frames
 Weakness
 When frames are small, the transmission chanel
are not used efficiently.
 Cannot use often for big size frame due to
 Limitation in buffer size
 Big size frame prones to bigger error probability
 In shared medium, it is not convenient to leave one
station using medium for long time
39
Sliding window: principle
 Transmitter sends more than one frame without waiting
in order to reduce waiting time
 Transmitted frame without ACK will still be stored in
buffer.
 Number of frame to be transmitted without ACK depends
on the size of buffer at transmitter
 When transmitter receives ACK, it realises the
succesfully transmitted frame from buffers
 Transmitter continues sending a number of frame
equivalent to the number of succesfully trasmitted
frames.
40
Sliding window: principle
 Assume that A and B are two stations connected by a full
duplex media
 B has a buffer size of n frame.
 B can receives n frame without sending ACK
 Acknowledgement
 In order to keep track of ACKed frames. It is neccessary to
number frames.
 B acknowledge a frame by telling A which fram B is waiting for
(by number of frame), implicitely saying that B receives well all
other frame before that.
 One ACK frame serves for acknowledes several frames.

41
 Sliding windows: principle

42

Window list the frames to transmit Window list the frames in waiting to receive
Sliding windows

43
Sliding windows
 Frame are numbered. The maximum number must
not be smaller than the size of the window.
 Frame are ACKed by another message with number
 Accumulated ACK: If frame 1,2,3,4 are well
receive, just send ACK 4
 ACK with number k means all frame k-1, k-2
…already well received.

44
Sliding windows
 Transmitter needs to manage some information:
 List of frames transmitted sucessfully
 List of frames transmitted without ACK
 List of frames to be sent immediatly
 List of frames NOT to be sent immediately
 Receiver keep tracks of
 List of frames well received
 List of frames expected to receive

45
Piggy backing
 A and B transmitte data in both sides
 When B needs to send an ACK while still needs to
send data, B attaches the ACK in the Data frame:
Piggybacking
 Otherwise, B can send an ACK frame separatly
 After ACK, if B sends some other data, it still put the
ACK information in data frame.
 Sliding window is much more efficient than Stop-
and-Wait
 More complicated in management.

46
Exercices
 Given a link with rate R=100Mbps
 We need to send a file over data link layer with file size L=100KB
 Assume that the size of a frame is: 1KB, header size is ignored
 Round trip time (RTT) between 2 ends of the link is 3ms
 An ACK message is sent back from receiver whenever a frame is
arrived. Size of ACK message is negligible
 What is the transmission time required if using Stop-and-wait
mechanism?
 Transmission time with sliding window if the window size is =7?
 Which size of window allow to obtain the fastest transmission?

47
Transmission time with Stop-
and-wait
transmitter
Emetteur receiver
Récepteur

Packet
Paquets Packet
Paquets

T transmit
frame
Trâme

RTT Ack

frame
Trâme

Trâme
frame
Trâme

48
Transmission time with Stop-
and-wait
 T total= Nb.frame * (T_transmit + RTT)
 T_transmit (F) = L(Frame)/ R
 Nb. frame = L/L(frame)

 With the given parametters

 Nb. frame =100 KB/1KB =100
 T_transmit (F) = 1KB/100 Mbps
=10^3*8/10^8 =8. 10^-5 (s)=0.08 (ms)
49
Sliding windows

50
Trasmission time with window
size 7
 T fastest= (T transmit 7 frames+ wait) * Nb.
Waiting time.
 1 waiting= (T transmit 1 frame+ RTT) – T
transmit 7 frames
 Nb. Waiting time= Nb frame /7

51
 Fastest transmission time with
sliding window
 Fastest transmission time
obtained if transmitter
receives ACK of the first
W gói
ACK frame when it finishes
transmitting the last frame
of the sliding window.
 Window size:W
 T transmit(W fram) >= T
transmit first frame + RTT

52
Fastest transmission time with
sliding window
 T transmit (W frame) = W * 1KB/R
 => (W-1)*1KB/R >= RTT
 => W >= RTT*R/1KB +1
 W>= 3ms * 100 Mbps/ 1KB + 1
 W>=38.5
 Smallest value of W = 39
 Time to transmit all data L = L/R + RTT =8
ms +3ms =11 ms
53
Media access control

54
Connection types
 Point-to-point
 ADSL
 Telephone modem
 Leased Line….
 Broadcast
 LAN using bus topology
 Wireless LAN
 HFC:
 …
 Broadcast networks need media access control
protocol in order to avoid collision when nodes try to
send data.
55
Multiple access links, protocols
two types of “links”:
 point-to-point
• point-to-point link between Ethernet switch, host
• PPP for dial-up access
 broadcast (shared wire or medium)
• old-fashioned Ethernet
• upstream HFC in cable-based access network
• 802.11 wireless LAN, 4G/4G. satellite

shared wire (e.g., humans at a cocktail party

cabled Ethernet) shared radio: 4G/5G shared radio: WiFi shared radio: satellite
(shared air, acoustical)

Link Layer: 6-56

Multiple access protocols
 single shared broadcast channel
 two or more simultaneous transmissions by nodes: interference
• collision if node receives two or more signals at the same time

multiple access protocol

 distributed algorithm that determines how nodes share channel,
i.e., determine when node can transmit
 communication about channel sharing must use channel itself!
• no out-of-band channel for coordination

Link Layer: 6-57

An ideal multiple access protocol
given: multiple access channel (MAC) of rate R bps
desiderata:
1. when one node wants to transmit, it can send at rate R.
2. when M nodes want to transmit, each can send at average
rate R/M
3. fully decentralized:
• no special node to coordinate transmissions
• no synchronization of clocks, slots
4. simple

Link Layer: 6-58

MAC protocols: taxonomy
three broad classes:
 channel partitioning
• divide channel into smaller “pieces” (time slots, frequency, code)
• allocate piece to node for exclusive use
• e.g. time - TDMA, frequency- FDMA, Code- CDMA
 random access
• channel not divided, allow collisions
• “recover” from collisions
• e.g. Pure Aloha, Slotted Aloha, CSMA/CD, CSMA/CA…
 “taking turns”
• nodes take turns, but nodes with more to send can take longer turns
• Token Ring, Token Bus

Link Layer: 6-59

Channel division
 FDMA: frequency division multiple access
 TDMA: time division multiple access
 CDMA: code division multiple access

60
 ex
TDMA và FDMA 4 stations
FDMA

frequency

time
TDMA:

frequency

time 61
CDMA
 Several senders can share the same frequency on a
single physical channel.
 Signals come from different senders are encoded by a
different random code
 Encrypted signals are mixed and then transmit on a
common frequency.
 The signals are recovered at the receiver by using the
same codes as at sender side.
 CDMA use the spread spectrum theory, CDMA shows
a lot of advantages that other technology cannot
achieve.
http://en.wikipedia.org/wiki/Spread_spectrum

62
 DS-CDMA System Overview
(Forward link)
CDMA is a multiple spread spectrum.

Freq. Freq. Freq. Freq.

Data A BPF BPF

Despreade Data A
r

Code A MS-A Code A

Freq. Freq. Freq.

Freq.

BPF BPF
Data B Despreade Data B
r

Code B
Code B MS-B

•••
BS
•••

Difference between each communication path is only the spreading code

 Random access: Pure Aloha
 Aloha is used in mobile network of
1G, 2.5G, 3G using GSM
technology .
 Pure Aloha:
 When one sender has data to send, just
sends it
 If while sending, the senders receive
data from other stations  there is
collision. All stations need to resend their
data.
 There are possibility to have collision
when retransmit.
 Problem: Sender does not check to see
if the chanel is free before sending data
 Grey package are having overlap in
time causing collision 64
Random access: Slotted Aloha
 Times axe is divided
into equal slots.
 Each station sends
data only at the
beginning of a time
slot.
  Collision possibility
is reduced
 Still have collision in
grey package 65
Random access: CSMA
 CSMA: Carrier Sense Multiple Access
 CSMA idea is similar to what happens in a meeting.
 CSMA:
 The sender “Listen before talk”
 If the channel is busy, wait
 If the chanel is free, transmit

66
CSMA

 CSMA: Sender listens before transmission:

 If the channel is free, send all the data
 If the channel is busy, wait.
 Why there are still collision?
 Due to propagation delay

67
Collision in CSMA
 Assume that there are 4
nodes in the channel
 The propagation of the
signal from one node to
the other requires a
certain delay.
 Ex:
 Transmissions from B and D
cause collision

68
CSMA/CA
 CSMA/CA is used WIFI standard IEEE 802.11
 If two stations discover that the channel is busy,
and both wait then it is possible that they will try to
resend data in the same time.
  collision
 Solution CSMA/CA.
 Each station wait for a random period  reduce the
collision possibility

69
CSMA/CD
 Used in Ethernet
 CSMA with Collision Detection:
 “Listen while talk”.
 A sender listen to the channel,
 If the channel is free then transmit data
 While a station transmit data, it listens to the channel. If it detects a
collision then transmits a short signal warning the collision then stop
 Do not continue the transmission even in collision as CSMA
 If the channel is busy, wait then transmit with
probability p
 Retransmit after a random waiting time.
70
Comparison between channel
division and random access
 Channel division
 Efficient, treat stations equally.
 Waste of resources if one station has much smaller data to
send than the others
 Random access
 When total load is small: Efficient since each station can
use the whole chanel
 When total load is large: Collision possibility increases.
 Token control: compromise between the two above
methods.
71
Token Ring
 A “tocken” is passed
T
from one node to the
other in a ring topo
 Only the token holder
can transmit data (nothing
 After finishing sending to send)
data, the token need to T
be passed to next
nodes.
 Some problem
 Time consuming in
passing token
 Loss of token due to data
72
some reasons
Summary on Media access control
mechanisms
 Channel division
 Random access
 Token
 What do you thinks about their advantages
and weaknesses

73