1

Wireless Medium Access Control

Romit Roy ChoudhuryWireless Networking LecturesDuke University

2

Wired Vs Wireless Media Access

Both are on shared media.Then, what’s really the problem ?

3

The Channel Access Problem

Multiple nodes share a channel

Pairwise communication desired Simultaneous communication not possible

MAC Protocols Suggests a scheme to schedule communication

• Maximize number of communications• Ensure fairness among all transmitters

AA CCBB

4

The Trivial Solution

Transmit and pray Plenty of collisions --> poor throughput at high

load

AA CCBB

collision

5

The Simple Fix

Transmit and pray Plenty of collisions --> poor throughput at high

load

Listen before you talk Carrier sense multiple access (CSMA) Defer transmission when signal on channel

AA CCBB

Don’ttransmit

Don’ttransmit

Can collisions still occur?Can collisions still occur?

6

CSMA collisions

Collisions can still occur:Propagation delay non-zero between transmitters

When collision:Entire packet transmission time wasted

spatial layout of nodes

note:Role of distance & propagation delay in determining collision probability

7

CSMA/CD (Collision Detection)

Keep listening to channel While transmitting

If (Transmitted_Signal != Sensed_Signal) Sender knows it’s a Collision ABORT

8

2 Observations on CSMA/CD

Transmitter can send/listen concurrently If (Sensed - received = null)? Then success

The signal is identical at Tx and Rx Non-dispersive

The TRANSMITTER can detect if and when collision occurs

The TRANSMITTER can detect if and when collision occurs

9

Unfortunately …

Both observations do not hold for wireless

Because …

10

Wireless Medium Access Control

A B

C D

Distance

Signalpower

SINR threhold

11

Wireless Media Disperse Energy

A B

C D

Distance

Signalpower

SINR threhold

A cannot send and listen in parallel

Signal not same at different locations

12

Collision Detection Difficult

Signal reception based on SINR Transmitter can only hear itself Cannot determine signal quality at

receiver

A CD

B

13

Calculating SINR

AB

C

CB

CtransmitC

B

AB

AAB

d

PI

d

PSoI

NNoiseIceInterferen

SoIterestSignalOfInSINR

transmit

)()(

)(

CB

Ctransmit

AB

A

AB

d

PN

d

P

SINR

transmit

D

14

A B

C D

Distance

Signalpower

SINR threhold

Red signal >> Blue signal

X

Red < Blue = collision

15

A B

C D

Distance

Signalpower

SINR threhold

Important: C has not heard A, but can interfere at receiver B

X

C is the hidden terminal to A

16

A B

C D

Distance

Signalpower

SINR threhold

Important: X has heard A, but should not defer transmission to Y

X

X is the exposed terminal to AY

17

Any Questions at this point?

18

So, how do we cope withHidden/Exposed Terminals?

19

A B

C D

Distance

Signalpower

SINR threhold

X

How to prevent C from trasmitting?

20

A Project Idea!

A B

C D

A node decides to intelligently choose a Carrier sensing threshold (T)

The node senses channel If signal > T, then node does not transmit If signal < T, then don’t transmit

Possible to guarantee no collisions?

21

A B

C D

Distance

Signalpower

SINR threhold

X

A Project Idea!

Sensitivity threshold

22

A B

C D

Distance

Signalpower

SINR threhold

Sensitivity threshold

X

T

A Project Idea!Will this solve the wireless MAC problem? Do not transmit in this region

23

Whatever the answer …

This is an example of a good class project

If you came up with the idea,Showed that it’s a new idea,

And evaluated it to demo how it performs

24

The Emergence of MACA, MACAW, & 802.11

Wireless MAC proved to be non-trivial

1992 - research by Karn (MACA) 1994 - research by Bhargavan (MACAW)

Led to IEEE 802.11 committee The standard was ratified in 1999

25

CTS = Clear To Send

RTS = Request To Send

IEEE 802.11

D

Y

S

M

K

RTS

CTS

X

26

IEEE 802.11

D

Y

S

X

M

Ksilenced

silenced

silenced

silencedData

ACK

27

802.11 Steps

All backlogged nodes choose a random number R = rand (0, CW_min)

Each node counts down R Continue carrier sensing while counting down Once carrier busy, freeze countdown

Whoever reaches ZERO transmits RTS Neighbors freeze countdown, decode RTS RTS contains (CTS + DATA + ACK) duration = T_comm Neighbors set NAV = T_comm

• Remains silent for NAV time

28

802.11 Steps

Receiver replies with CTS Also contains (DATA + ACK) duration. Neighbors update NAV again

Tx sends DATA, Rx acknowledges with ACK After ACK, everyone initiates remaining countdown Tx chooses new R = rand (0, CW_min)

If RTS or DATA collides (i.e., no CTS/ACK returns) Indicates collision RTS chooses new random no. R1 = rand (0, 2*CW_min) Note Exponential Backoff Ri = rand (0, 2^i * CW_min) Once successful transmission, reset to rand(0, CW_min)

29

But is that enough?

30

RTS/CTS

Does it solve hidden terminals ? Assuming carrier sensing zone =

communication zone

C

F

A B

E

D

CTS

RTS

E does not receive CTS successfully Can later initiate transmission to D.Hidden terminal problem remains.

E does not receive CTS successfully Can later initiate transmission to D.Hidden terminal problem remains.

31

Hidden Terminal Problem

How about increasing carrier sense range ?? E will defer on sensing carrier no

collision !!!

CB DData

A

E

CTS

RTSF

32

Hidden Terminal Problem

But what if barriers/obstructions ?? E doesn’t hear C Carrier sensing does not

help

CB DData

A

EF

CTS

RTS

33

Exposed Terminal

B should be able to transmit to A RTS prevents this

CA B

E

D

CTSRTS

34

Exposed Terminal

B should be able to transmit to A Carrier sensing makes the situation worse

CA B

E

D

CTSRTS

35

Thoughts !

802.11 does not solve HT/ET completely Only alleviates the problem through RTS/CTS and

recommends larger CS zone

Large CS zone aggravates exposed terminals Spatial reuse reduces A tradeoff RTS/CTS packets also consume bandwidth Moreover, backing off mechanism is also wasteful

The search for the best MAC protocol is still on. However, 802.11 is being optimized too.Thus, wireless MAC research still alive

36

Takes on 802.11

Role of RTS/CTS Useful? No? Is it a one-fit-all? Where does it not fit?

Is ACK necessary? MACA said no ACKs. Let TCP recover from losses

Should Carrier Sensing replace RTS/CTS?

New opportunities may not need RTS/CTS Infratructured wireless networks (EWLAN)

37

MACA-BI [GerlaUCLA]

RTS/CTS/ACK are control overhead Needed to reduce it

Rx predicts trasmission from the Tx Traffic estimation (???)

If Rx thinks Tx has pending packets for Rx Rx transmits RTR to Tx Tx replies with Data

Improves MACA with no RTS/ACK improvement but not too much

38

DBTMA [HaasCornell98]

A B X Y

RTS CTS Signal X

Rx Busy tone

A B X Y

CTS RTS Signal X

Tx Busy toneTx Busy tone

39

Implicit MACKnowledgment

APs typically backlogged with traffic Persistent traffic possibility of optimzation

We propose an implicit ACK optimization Piggyback the CTS with ACK for previous

dialog802.11

Implicit ACK

Gain

40

Hybrid Channel Access

The optimization timeline

T RRTS

CTS

Data

ACK

RTS

CTS

Data

ACK

T RRTS

CTS

Data

RTS

CTS +ACK

Data

T RRTS

CTS

Data

Poll +ACK

Data

RTS

CTS +ACK

Bac

kof

f

Bac

kof

fB

ack

off

Bac

kof

f

Poll +ACK

Data

Bac

kof

fB

ack

off

802.11802.11 Implicit ACKImplicit ACK Hybrid Channel AccessHybrid Channel Access

41

Seedex [KumarUIUC03]

Forget channel reservation and backoff Instead, let nodes pick sequence of time slots

Decides to probably transmit in some, else listen Transmit slots chosen using a random seed Publishes the seed to 2-hop neighbors

When PT slots arrive, nodes transmit with Probability “p” “p” chosen as a function of overlapping neighbors

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.

42

Hot Research Topics

Power control increases spatial reuse Whisper in the room so that many people can talk

Rate control based on channel quality Expolit channel diversity

Utilize multiple channels to parallelize dialogs

Exploit spatial diversity Use directional antennas to interfere over smaller

region (next class)

… and many more topics

43

Questions ?

44

Announcements

Reviews: You are forgetting to appreciate the paper There is a reason why the paper was accepted

Please organize your papers/reviews Would be valuable later in career You never know what you will do after 5 years

Meet me with slides before you present Email for an appointment Don’t have to review if you are presenter

45

Announcements

Review Template: Problem definition? Why is it important? Validity of models and assumptions Solution Evaluation

Email review to TA (CC to me) Bring print out to class Name-date-subject in email subject Will post example reviews on webpage

Some of you still doing summaries.

46

Backup slides onIEEE 802.11

Read for more details

47

Today’s Discussions

IEEE 802.11 overview - some raw data Architecture PHY specifications – Spread Spectrum radios: FH & DS MAC specifications – DCF and PCF Synchronization, Power management, Roaming,

Scanning Security

Deliberations on 802.11 (DCF) MAC Hidden terminal & Exposed terminal issues Carrier sensing

Some other ideas & open challenges Could be interesting for the project

48

IEEE 802.11 – An overview

49

IEEE 802.11 in OSI Model

Wireless

50

802.11 Scope & Modules

MAC Sublayer

MAC LayerManagement

PLCP Sublayer

PMD Sublayer

PHY LayerManagement

LLC

MAC

PHY

To develop a MAC and PHY spec for wirelessconnectivity for fixed, portable and moving stationsin a local area

51

Applications

Single Hop Home networks Enterprise networks (e.g., offices, labs, etc.) Outdoor areas (e.g., cities, parks, etc.)

Multi-hops Adhoc network of small groups

(e.g.,aircrafts) Balloon networks (SpaceData Inc.) Mesh networks (e.g., routers on lamp-posts)

52

802.11 Architecture – Two modes

53

802.11 PHY Technologies

Two kinds of radios based on “Spread Spectrum” “Diffused Infrared”

Spread Spectrum radios based on Frequency hopping (FH) Direct sequence (DS)

Radio works in 2.4GHz ISM band --- license-free by FCC (USA), ETSI (Europe), and MKK (Japan) 1 Mbps and 2Mbps operation using FH 1, 2, 5.5, and 11Mbps operation using DSSS (FCC)

54

Why Spread Spectrum ?

C = B*log2(1+S/N) . . . [Shannon] To achieve the same channel capacity C

Large S/N, small B Small S/N, large B Increase S/N is inefficient due to the logarithmic relationship

power

B

signalnoise, interferences

power

signal

Bfrequency

e.g. B = 30 KHz e.g. B = 1.25 MHz

55

Spread Spectrum

Reduce effect of jamming Military scenarios

Reduce effect of other interferences More “secure”

Signal “merged” in noise and interference

Methods for spreading the bandwidth of the transmitted signal over a frequency band (spectrum)

which is wider than the minimum bandwidth required to transmit the signal.

56

Frequency Hopping SS (FHSS)

2.4GHz band divided into 75 1MHz subchannels

Sender and receive agree on a hopping pattern (pseudo random series). 22 hopping patterns defined

Different hopping sequences enable co-existence of multiple BSSs

Robust against narrow-band interferences

f f f f f f f f f f f

One possible pattern

57

FHSS due to [Lamarr1940]power

B

signalnoise, interferences

power

signal

Bfrequencyf f f f f f f f f f f

Invented by Hedy Lamarr (Hollywood film star) in 1940, at age of 27, with musician George Antheil

Simple radio design with FHSSData rates ~ 2 Mbps

58

Direct Sequence SS

Direct sequence (DS): most prevalent Signal is spread by a wide bandwidth

pseudorandom sequence (code sequence) Signals appear as wideband noise to

unintended receivers

Not for intra-cell multiple access Nodes in the same cell use same code

sequence

59

IEEE 802.11b DSSS

ISM unlicensed frequency band (2.4GHz)

Channel bandwidth: fhigh – flow = 22 MHz

1MHz guard band Direct sequence

spread spectrum in each channel

3 non-overlapping channels

Channel flow fhigh

1 2.401 2.423

2 2.404 2.428

3 2.411 2.433

4 2.416 2.438

5 2.421 2.443

6 2.426 2.448

7 2.431 2.453

8 2.436 2.458

9 2.441 2.463

10 2.446 2.468

11 2.451 2.473

60

Diffused Infrared

Wavelength range from 850 – 950 nm For indoor use only Line-of-sight and reflected transmission 1 – 2 Mbps

61

PHY Sublayers

Physical layer convergence protocol (PLCP) Provides common interface for MAC

• Offers carrier sense status & CCA (Clear channel assesment)

• Performs channel synchronization / training

Physical medium dependent sublayer (PMD) Functions based on underlying channel quality

and characteristics• E.g., Takes care of the wireless encoding

62

PLCP (802.11b)

longpreamble

192us

shortpreamble

96us(VoIP, video)

63

PLCP (802.11b)long

preamble192us

shortpreamble

96us(VoIP, video)

Note:

To send even one bit payloadreliably, you will have to form a packet with the PLCP preambleand the PLCP header.

This constraints protocol design

You cannot arbitrarily exchangecontrol messages.

What are the control messagesin IEEE 802.11 ?

64

IEEE 802.11 MAC

65

802.11 MAC (DCF)

CSMA/CA based protocol Listen before you talk CA = Collision avoidance (prevention is better than cure !!)

Robust for interference Explicit acknowledgment requested from receiverExplicit acknowledgment requested from receiver

• for unicast frames Only CSMA/CA for Broadcast frames

Optional RTS/CTS offers Virtual Carrier Sensing RTS/CTS includes duration of immediate dialog Addresses hidden terminal problems

66

802.11 MAC (DCF)

67

Physical Carrier Sense & Backoff

68

MAC Management Layer

Synchronization Finding and staying with a WLAN

• Uses TSF timers and beacons

Power Management Sleeping without missing any messages

• Periodic sleep, frame buffering, traffic indication map

Association and Reassociation Joining a network Roaming, moving from one AP to another Scanning

69

Synchronization

Timing Synchronization Function (TSF) Enables synchronous waking/sleeping Enables switching from DCF to PCF Enables frequency hopping in FHSS PHY

• Transmitter and receiver has identical dwell interval at each center frequency

Achieving TSF All stations maintain a local timer. AP periodically broadcasts beacons containing

timestamps, management info, roaming info, etc.• Not necessary to hear every beacon

Beacon synchronizes entire BSS• Applicable in infrastructure mode ONLY

Distributed TSF (for Independent BSS) more difficult

70

Power management Battery powered devices require power efficiency

LAN protocols assume idle nodes are always ON and thus ready to receive.

Idle-receive state key source of power wastage

Devices need to power off during idle periods Yet maintain an active session – tradeoff power Vs throughput

Achieving power conservation Allow idle stations to go to sleep periodically APs buffer packets for sleeping stations AP announces which stations have frames buffered

when all stations are awake – called Traffic Indication Map (TIM)

• TSF assures AP and Power Save stations are synchronized • TSF timer keeps running when stations are sleeping

Independent BSS also have Power Management Similar in concept, distributed approach

71

Roaming & Scanning

Stations switch (roam) to different AP When channel quality with current AP is poor

Scanning function used to find better AP Passive Scanning Listen for beacon from different Aps Active Scanning Exchange explicit beacons to determine

best AP

Station sends Reassociation Request to new AP If Reassociation Response successful Roaming

If AP accepts Reassociation Request AP indicates Reassociation to the Distribution System Distribution System information is updated Normally old AP is notified through Distribution System

72

MAC management frame

Beacon Timestamp, Beacon Interval, Capabilities, ESSID,

Supported Rates, parameters Traffic Indication Map

Probe ESSID, Capabilities, Supported Rates

Probe Response Timestamp, Beacon Interval, Capabilities, ESSID,

Supported Rates, parameters same for Beacon except for TIM

Association Request Capability, Listen Interval, ESSID, Supported Rates

Association Response Capability, Status Code, Station ID, Supported Rates

73

MAC Management Frame

Reassociation Request Capability, Listen Interval, ESSID, Supported Rates,

Current AP Address

Reassociation Response Capability, Status Code, Station ID, Supported Rates

Disassociation Reason code

Authentication Algorithm, Sequence, Status, Challenge Text

Deauthentication Reason

74

Security Range of attacks huge in wireless

Easy entry into the network Jamming, selfish behavior, spatial overhearing

Securing the network harder than wired networks Especially in distributed environments

WEP symmetric 40 or 128-bit encryption WPA: Wi-Fi protected access

Temporal key integrity protocol (TKIP) – better User authentication

IEEE 802.11i – Efforts toward higher security

75

PLCP

PLCP has two structures. All 802.11b systems have to support Long preamble. Short preamble option is provided to improve efficiency

when trasnmitting voice, VoIP, streaming video.

PLCP Frame format PLCP preamble

• SFD: start frame delimiter PLCP header

76

PLCP Header

8-bit signal or data rate (DR) indicates how fast data will be transmitted

8-bit service field reserved for future 16-bit length field indicating the length of

the ensuing MAC PDU (MAC sublayer’s Protocol Data Unit)

16-bit Cyclic Redundancy Code

77

Power management approach

Allow idle stations to go to sleep station’s power save mode stored in AP

APs buffer packets for sleeping stations. AP announces which stations have frames buffered Traffic Indication Map (TIM) sent with every Beacon

Power Saving stations wake up periodically listen for Beacons

TSF assures AP and Power Save stations are synchronized stations will wake up to hear a Beacon TSF timer keeps running when stations are sleeping synchronization allows extreme low power operation

Independent BSS also have Power Management similar in concept, distributed approach

78

Scanning

Scanning required for many functions. finding and joining a network finding a new AP while roaming initializing an Independent BSS (ad hoc) network

802.11 MAC uses a common mechanism for all PHY. single or multi channel passive or active scanning

Passive Scanning Find networks simply by listening for Beacons

Active Scanning On each channel Send a Probe, Wait for a Probe

Response Beacon or Probe Response contains information necessary

to join new network.

79

Active scanning example

80

Collision Detection

What is the aim of collision detection ?

It’s a transmitter’s job:

To determine if the packet was successfully received without explicitly asking the receiver

It’s a transmitter’s job:

To determine if the packet was successfully received without explicitly asking the receiver