1

Tecnologie e Protocolli per Internet 1

Prof. Stefano Salsanoe-mail: stefano.salsano@uniroma2.it

AA2011/12 – Blocco 3

2

Introduction to 802.11 Wireless LANsIntroduction to 802.11 Wireless LANs

Quote from Matthew Gast - 802.11® Wireless Networks The Definitive Guide – apr. 2005, 2nd edition

At this point, there is no way to prevent the spread of Wi-Fi.

In the years since the first edition of [his] book, wireless networking has gone from an interesting toy to a must-have

technology.

[…] [Wireless networking] seems poised to continue its march towards the standard method of network connection,

replacing "Where's the network jack?" with "Do you have Wi-Fi?" as the question to ask about network access.

3

WLAN historyWLAN history

• Original goal: » Deploy “wireless Ethernet”

» First generation proprietary solutions (end ’80, begin ’90):

» WaveLAN (AT&T)

» HomeRF (Proxim)

» Abandoned by major chip makers (e.g. Intel: dismissed HomeRF in april 2001)

• IEEE 802.11 Committee formed in 1990» Charter: specification of MAC and PHY for WLAN

» First standard: june 1997

» 1 and 2 Mbps operation

» Reference standard: september 1999

» Multiple Physical Layers

» Two operative Industrial, Scientific & Medical (ISM) shared unlicensed band � 2.4 GHz: Legacy; 802.11b/g

� 5 GHz: 802.11a

• 1999: Wireless Ethernet Compatibility Alliance (WECA) certification» Later on named Wi-Fi

» Boosted 802.11 deployment!!

4

WLAN data ratesWLAN data rates

• Legacy 802.11» Work started in 1990; standardized in 1997

» 1 mbps & 2 mbps

• The 1999 revolution: PHY layer impressive achievements» 802.11a: PHY for 5 GHz

» published in 1999

» Products available since early 2002

» 802.11b: higher rate PHY for 2.4 GHz

» Published in 1999

» Products available since 1999

» Interoperability tested (wifi)

• 2003: extend 802.11b» 802.11g: OFDM for 2.4 GHz – 54mbps

» Published in june 2003

» Backward compatibility with 802.11b Wi-Fi

• 2009: 802.11n» Launched in september 2003

» Published October 2009 (pre-standard products from 2007)

» Uses “MIMO” multiple input multiple output Up to 600 Mbps (with 4 antennas each side)

1, 2, 5.5, 11; 6, 9, 12, 18, 24, 36, 48, 54

2.4 GHzDSSS, HR-DSSS, OFDM

802.11g

6, 9, 12, 18, 24, 36, 48, 54

5.2, 5.5 GHz

OFDM802.11a

1, 2, 5.5, 11,

22, 33, 44

2.4 GHzDSSS, HR-DSSS, (PBCC)

"802.11b+" non-standard

1, 2, 5.5, 11

2.4 GHzDSSS, HR-DSSS

802.11b

1, 22.4 GHz, IR

FHSS, DSSS, IR

802.11 legacy

Data RatesMbps

Freq. Band

Transfer Method

Standard

5

Why multiple rates?“Adaptive” (?) coding/modulation

Why multiple rates?“Adaptive” (?) coding/modulation

Example: 802.11a case

6

PHY distance/rate tradeoffsPHY distance/rate tradeoffs

0.0

20.0

40.0

60.0

80.0

100.0

120.0

140.0

Dis

tan

ce

(m

)

54Mbps36Mbps24Mbps12Mbps6Mbps 11Mbps5.5Mbps1Mbps

2.4 GHz OFDM (.11g)

5 GHz OFDM (.11a)

2.4 GHz (.11b)

7

Coverage performance Cisco Aironet 350 Access Point

Coverage performance Cisco Aironet 350 Access Point

11 Mb/s DSSfrom ~30 to ~45 mt

5.5 Mb/s DSSfrom ~45 to ~76 mt

2 Mb/s DSSfrom ~76 to ~107 mt

Configurable TX power:

50, 30, 20, 5, 1 mW(100 mW outside Europe)

Greater TX power, faster battery consumptions!

Question: how to select transmission rate?

(STA does not explicitly know its distance from AP)

More later (implementation-dependent ☺☺☺☺)

8

WLAN NIC addressesWLAN NIC addresses

• Same as Ethernet NIC

» 48 bits = 2 + 46

• Ethernet & WLAN addresses do coexist

» undistinguishable, in a same (Layer-2) network

» role of typical AP = bridge� (to be precise: when the AP act as “portal” in 802.11 nomenclature)

802 IEEE

48 bit addresses

1 bit = individual/group

1 bit = universal/local

46 bit address

AP

192.168.1.32

00:0a:e6:f8:03:ad

AP

192.168.1.43

00:06:6e:00:32:1a 192.168.1.52

00:82:00:11:22:33

C:>arp -a

192.168.1.32 00-0a-e6-f8-03-ad dinamico

192.168.1.43 00-06-6e-00-32-1a dinamico

192.168.1.52 00-82-00-11-22-33 dinamico

9

Protocol stackProtocol stack

802.2 Logical Link Control

802.3

MAC

802.3

PHY

DATA LINK LAYERLLC sublayer

LLC

MAC

802

overview

&

architecture

802.1

management

&

bridging

802.11 MAC…

…

…

…

…802.11

FHSS PHY

802.11DSSS PHY

802.11aOFDM PHY

802.11bHR-DSSS

PHY

802.11gExtended

Rate PHY

DATA LINK LAYERMAC sublayer

PHYSICAL LAYER

802.11: “just” another 802 link layer ☺☺☺☺

10

802.11 MAC Data Frame802.11 MAC Data Frame

PHY IEEE 802.11 Data 0 - 2312 FCS

Protocol

versionType Sub Type info

2 2 12

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

Frame

Control

Duration

/ IDAddress 1 Address 2 Address 3

Sequence

ControlAddress 4 Data

Framecheck

sequence

2 2 2 40-23126666

Fragment

numberSequence number

4 12

MAC overhead:

- 28 bytes (24 header + 4 FCS) or

- 34 bytes (30 header + 4 FCS)

DETAILS AND EXPLANATION LATER ON

11

EncapsulationEncapsulation

IdenticalTo 802.3/LLC encapsulation

802.11 MAC frame: no “type” field (such as Ethernet II)!!LLC encapsulation mandatory

12

Crossing a wireless bridge / 1Crossing a wireless bridge / 1

ETH/802.11

bridge

802.11/ETH

bridge

typeSRCDEST P

typeSRCDEST P

AA AA 03 00.00.00802.11 MAC header Type P

13

Crossing a wireless bridge / 2Crossing a wireless bridge / 2

ETH/802.11

bridge

802.11/ETH

bridge

AA AA 03 00.00.00LenSRCDEST Type P

typeSRCDEST P

AA AA 03 00.00.00802.11 MAC header Type PType P

14

Why Ethernet Tunnel?(just needed in very special cases: IPX, AARP)

Why Ethernet Tunnel?(just needed in very special cases: IPX, AARP)

AA AA 03 00.00.00LenSRCDESC Type

ETH/802.11

bridge

802.11/ETH

bridge

P

AA AA 03 00.00.00LenSRCDESC Type P

TypeSRCDESC P ?????Some protocols

MUST have this

Encapsulation:

-Novell IPX

(Type 0x8137)

- Apple-Talk ARP

(Type 0x80F3)

15

Handling 802.11 framesHandling 802.11 frames

Ethernet-like driver interfacesupports virtually all protocol stacks

Maximum Data limited to 1500 octets

Frame translationIEEE Std 802.1H

IEEE 802.3 frames: translated to 802.11

Ethernet Types 8137 (Novell IPX) and 80F3 (AARP) encapsulated via Ethernet Tunnel

In general, any protocol listed in the “selective translation table” of the bridge

All other Ethernet Types: encapsulated via RFC 1042 SNAP

Transparent bridging to Ethernet

Platform Computer

Platform

Computer

PC-Card

Hardware

PC-Card

HardwareRadio

Hardware

Radio Hardware

WMAC controller with

Station Firmware

(WNIC-STA)

WMAC controller with

Station Firmware(WNIC-STA)

Driver Software

(STADr)

Driver

Software

(STADr)

802.11 frame format

802.3 frame format

Ethernet V2.0 / 802.3

frame format

Protocol StackProtocol Stack

BridgeSoftware

BridgeSoftware

PC-Card

Hardware

PC-Card Hardware

Radio

Hardware

Radio

Hardware

WMAC controller with

Access Point Firmware

(WNIC-AP)

WMAC controller with

Access Point Firmware

(WNIC-AP)

Driver

Software

(APDr)

Driver

Software

(APDr)

802.11 frame format

802.3 frame format

Ethernet V2.0 / 802.3frame format

Kernel Software (APK)Kernel Software (APK)

Bridge

Hardware

Bridge

HardwareEthernet

Interface

Ethernet

Interface

STA AP

16

http://www.ieee802.org/1/files/public/docs2009/H-nolish-draftintro-0109-v01.ppt

Ethernet/802.3 is a data link standard. A data link may support multiple

network layer protocols, such as IP, IPX, ARP, whatever. There needs to be a

mechanism that demultiplexes a frame’s contents so that its contents are directed

to the proper protocol handler upon reception at a station. These mechanisms are

different for different variants of Ethernet and 802.n standards. 802.1H describes

how the protocol selection function is modified when transferring a frame

between two different 802.n data links.

RFC 1042 was the first attempt to solve the Ethernet/802.3 interworking issue

Frames transiting to an 802.3 network would have an RFC 1042 header added to

them. Frames transiting an Ethernet network would have the RFC 1042

encapsulation removed

17

Certain protocols are interpreted differently on an end station depending upon if

they are raw EtherType or LLC/SNAP encapsulated. Thus a frame transiting

between two Ethernet LANS via an intervening 802.3 LAN would always lose its

LLC/SNAP encapsulation, thus changing the meaning of the data flow between

the end stations.

802.1H Defines the Bridge Tunnel Service. The basic concept is that certain

protocols are encapsulated using the Bridge-Tunnel encapsulation for transit

across an 802.3 network.

When bridged to a non-802.3 network, the frames are de-encapsulated before

transiting to the non-802.3 network. However, frames with Bridge-Tunnel

encapsulation are not de-encapsulated but are sent using the RFC 1042

encapsulation.

This preserves the distinction between LLC/SNAP and raw Ethertype frames

across the 802.3 network.

18

802.11 Network ArchitectureAnd related addressing

19

Basic Service Set (BSS)Basic Service Set (BSS)

• Infrastructure BSS» or, simply, BSS

» Stations connected through AP

» Typically interconnetted to a (wired) network infrastructure

• Independent BSS (IBSS)» Stations communicate directly

with each other

» Smallest possible IBSS: 2 STA

» IBSS set up for a specificpurpose and for short time (e.g. meeting)

» That’s why they are alsocalled “ad hoc networks”

AP

Network infrastructure

BSS: group of stations that can communicate with each other

20

Frame Forwarding in a BSSFrame Forwarding in a BSS

AP

Network infrastructure

BSS: AP = relay functionNo direct communication allowed!

IBSS: direct communicationbetween all pairs of STAs

21

Why AP = relay function?Why AP = relay function?

• Management:» Mobile stations do NOT neet to maintain neighbohr relationship with other

MS in the area

» But only need to make sure they remain properly associated to the AP

» Association = get connected to (equivalent to plug-in a wire to a bridge ☺☺☺☺)

• Power Saving:» APs may assist MS in their power saving functions

» by buffering frames dedicated to a (sleeping) MS when it is in PS mode

• Security:» AP may manage security and authenticate users

• Obvious disadvantage: use channel bandwidth twice…

22

Addressing in IBSS (ad hoc)Addressing in IBSS (ad hoc)

Frame

Control

Duration

/ ID

Address 1

DA

Address 2

SA

Address 3

BSSID

Sequence

ControlData FCS

SA = Source AddressDA = Destination Address

BSSID = Basic Service Set IDentifierused for filtering frames at reception (does the frame belong to OUR cell?)format: 6 bytes random MAC address with Universal/Local bit set to 1

SA

DA

23

Addressing in a BSS?Addressing in a BSS?

X

AP

DA

SA

24

Addressing in a BSS!Addressing in a BSS!

AP

Distribution system

Frame must carry following info:1) Destined to DA2) But through the APWhat is the most general addressing structure?

DASA

25

Addressing in a BSS (to AP)Addressing in a BSS (to AP)

Frame

Control

Duration

/ ID

Address 1

BSSID

Address 2

SA

Address 3

DA

Sequence

ControlData FCS

AP

Distribution system

DASA

BSSID

Protocol

versionType

2 2

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

1 0

Address 2 = wireless TxAddress 1 = wireless Rx

Address 3 = dest

BSSID = AP MAC address

26

Addressing in a BSS (from AP)Addressing in a BSS (from AP)

Frame

Control

Duration

/ ID

Address 1

DA

Address 2

BSSID

Address 3

SA

Sequence

ControlData FCS

AP

Distribution system

DASA

BSSID

Protocol

versionType

2 2

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

0 1

Address 2 = wireless TxAddress 1 = wireless Rx

Address 3 = src

27

From AP: do we really need 3 addresses?From AP: do we really need 3 addresses?

AP

Distribution system

DASA

BSSID

DA correctly receives frame, and send 802.11 ACK to … BSSID (wireless transmitted)

DA correctly receives frame, and send higher level ACK to … SA (actual transmitter)

28

ESS - Extended Service SetESS - Extended Service Set

AP1

AP2 AP3 AP4

BSS1

BSS2 BSS3 BSS4

ESS: created by merging different BSS through a network infrastructure(possibly overlapping BSS – to offer a continuous coverage area)

Stations within ESS MAY communicate each other via Layer 2 proceduresAPs acting as bridgesMUST be on a same LAN or switched LAN or VLAN (no routers in between)

29

Service Set IDentifier (SSID)Service Set IDentifier (SSID)

• name of the WLAN network » Plain text (ascii), up to 32 char

• Assigned by the network administrator

» All BSS in a same ESS have same SSID

• Typically (but not necessarily) istransmitted in periodic management frames (beacon)

» Disabling SSID transmission = a (poor!) security mechanism

» Typical: 1 broadcast beacon every 100 ms (configurable by sysadm)

» Beacon may transmit a LOT of other info (see example – a simple one!)

IEEE 802.11 wireless LAN management frame

Fixed parameters (12 bytes)

Timestamp: 0x00000109EAB69185

Beacon Interval: 0,102400 [Seconds]

Capability Information: 0x0015

.... .... .... ...1 = ESS capabilities: Transmitter is an AP

.... .... .... ..0. = IBSS status: Transmitter belongs to a BSS

.... .... .... 01.. = CFP participation capabilities: Point coordinator at

AP for delivery and polling (0x0001)

.... .... ...1 .... = Privacy: AP/STA can support WEP

.... .... ..0. .... = Short Preamble: Short preamble not allowed

.... .... .0.. .... = PBCC: PBCC modulation not allowed

.... .... 0... .... = Channel Agility: Channel agility not in use

.... .0.. .... .... = Short Slot Time: Short slot time not in use

..0. .... .... .... = DSSS-OFDM: DSSS-OFDM modulation not allowed

Tagged parameters

Tag Number: 0 (SSID parameter set)

Tag length: 4

Tag interpretation: WLAN

Tag Number: 1 (Supported Rates)

Tag length: 4

Tag interpretation: Supported rates: 1,0(B) 2,0(B) 5,5 11,0 [Mbit/sec]

Tag Number: 6 (IBSS Parameter set)

Tag length: 1

Tag interpretation: ATIM window 0x2

Tag Number: 5 ((TIM) Traffic Indication Map)

Tag length: 4

Tag interpretation: DTIM count 0, DTIM period 1,

Bitmap control 0x0, (Bitmap suppressed)

30

The concept of Distribution SystemThe concept of Distribution System

“Logical” architecture componentProvides a “service”

DSS = Distribution System Service

Standard does NOT say how it is implemented

Specified only which functions it provides

Association

Disassociation

Reassociation

Integration

Distribution

Association/disassociationRegistration/de-registration of a STA to an APEquivalent to “plugging/unplugging the wire” to a switchDS uses this information to determine which AP send

frames to

Reassociationi.e. handling STA mobility in a same ESS!

DistributionAn AP receives a frame on its air interface (e.g. STA 2)It gives the message to the distribution service (DSS) of the

DSThe DSS has the duty to deliver the frame to the proper

destination (AP)

IntegrationMust allow the connection to non 802.11 LANs

Though, in practice, non 802.11 LANs are Ethernet and no “real portals” are deployed

31

DS, againDS, again

AP1 AP2 AP3

Association

IAPP/proprietary IAPP/proprietary

Distribution system (physical connectivity + logical service support)

MSs in a same ESS need to1) communicate each other2) move through the ESS

Typical implementation (media)Switched Ethernet BackboneBut alternative “Distribution Medium” are

possibleE.g. Wireless Distribution System (WDS)

Implementation dutiesan AP must inform other APs of associated

MSs MAC addresses

StandardizationFrom 1997: tentative to standardize an IAPPFinalized as “working practice standard” in 802.11F

(june 2003)Nobody cared!

Plenty of proprietary solutionsMust use APs from same vendor in whole ESS

Current trends (2004+):Centralized solutions (see Aruba, Cisco, Colubris)

Include centralized management, too!Current attempt: convergence to CAPWAP?

32

Addressing in an ESSAddressing in an ESS

AP

Distribution System

DA

SA

BSSID#1

Frame

Control

Duration

/ ID

Address 1

BSSID#1

Address 2

SA

Address 3

DA

Sequence

ControlData FCS

Protocol

versionType

2 2

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

1 0

AP

DA

idea: DS will be able to forward frame to dest(either if fixed or wireless MAC)

Same approach! Works in general, even if DA in different BSS

33

Addressing in an ESSAddressing in an ESS

Same approach! Works in general, even if DA in different BSS

AP

Distribution System

DASA

BSSID#2

AP

DA

Frame

Control

Duration

/ ID

Address 1

DA

Address 2

BSSID#2

Address 3

SA

Sequence

ControlData FCS

Protocol

versionType

2 2

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

0 1

34

Wireless Distribution SystemWireless Distribution System

AP1 AP2 AP3

DS medium:- not necessarily an ethernet backbone!- could be the 802.11 technology itself

Resulting AP = wireless bridge

35

Addressing within a WDSAddressing within a WDS

AP

Wireless Distribution System

SA

TA

AP

DA

Frame

Control

Duration

/ ID

Address 1

RA

Address 2

TA

Address 3

DA

Sequence

Control

Address 4

SAData FCS

Protocol

versionType

2 2

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

1 1

RA

Address 4: initially forgotten? ☺

36

Addressing: summaryAddressing: summary

Wireless DS

To AP

From AP

IBSS

Function

SADATARA11

N/ADASARA = BSSID01

N/ASABSSIDRA = DA10

N/ABSSIDSARA = DA00

Address 4Address 3Address 2Address 1From DSTo DS

Receiver Transmitter

BSS Identifier (BSSID)unique identifier for a particular BSS. In an infrastructure BSSID it is the MAC address of the AP.

In IBSS, it is random and locally administered by the starting station. (uniqueness)

Transmitter Address (TA)MAC address of the station that transmit the frame to the wireless medium. Always an individual

address.

Receiver Address (RA)to which the frame is sent over wireless medium. Individual or Group.

Source Address (SA) MAC address of the station who originated the frame. Always individual address. May not match TA because of the indirection performed by DS of an IEEE 802.11 WLAN. SA field

is considered by higher layers.

Destination Address (DA)Final destination . Individual or Group. May not match RA because of the indirection.

37

802.11 MACCSMA/CA Distributed Coordination

Function

Carrier Sense Multiple AccessCarrier Sense Multiple Access

With Collision AvoidanceWith Collision Avoidance

38

Wireless Medium UnreliabilityWireless Medium Unreliability

11 Mbps 802.11b outdoor measurements - Roma 2 Campus - roof nodes

39

Must rely on explicit ACKsMust rely on explicit ACKs

• Successful DATA transmission:» ONLY IF an ACK is received

• ACK transmission provided by MAC layer» Immediate retransmission

� Don’t get confused with higher layer rtx

• DATA-ACK exchange:» Also called two-way handshake

» Or Basic Access Mechanism

SENDER RECEIVER

DATA

ACK

40

Possible errorsPossible errors

• Three causes of insuccess» PHY Error

» Receiver cannot synchronize with transmitted frame� preamble + SFD needed

» or cannot properly read Physical Layer Control Protocol (PLCP) header� PLCP header contains the essential information on employed rate

� Without it receiver cannot know how to demodulate/decode received frame!

» CRC32 error

» MAC frame (MAC Header + Payload) CRC failures� The greater the rate, the higher the SNR required to correctly transmit

» ACK Error

» Transmitter does not receive ACK� ACK corrupted by PHY or CRC32 errors

» It IS an error: though data frame was correctly received, transmitted does not know

� Introduce issue of duplicated frames at the receiver

PHY MAC header Payload FCS

Preamble SFD PLCP hdr

41

Wireless errors

11 Mbps 802.11b/g OUTDOOR measurements - Roma 2 Campus - roof nodes

PHY errors CANNOT be reduced through automatic rate fallback mechanisms

An (apparent) paradox: 802.11b@11mbps outdoor outperforms 802.11g@6mbps !!!but it is NOT a paradox ☺☺☺☺ since most 802.11g errors are PHY (unrelated with rate)…

802.11b@11Mbps 802.11g@6Mbps

42

Must forget Collision Detection!Must forget Collision Detection!

• One single RF circuitry» Either TX or RX…

» Half-duplex

• Even if two simultaneous TX+RX: large difference (100+ dB!) in TX/RX signal power

» Impossible to receive while transmitting

» On a same channel, of course

• Collision detection at sender: meaningless in wireless!

» Ethernet = collision detection at sender

» Wireless = large difference in the interference power between sender & receiver!

» Collision OCCURS AT THE RECEIVER

STA

tx

rx

CA B

A detects a very low interference

(C is far)no “collision”

B detects a disructive interference(C is near)

collision occurs

43

Distributed Coordination Function Basics

Distributed Coordination Function Basics

44

802.11 MAC802.11 MAC

DISTRIBUTED COORDINATION FUNCTION

DCF(CSMA/CA)

POINT

COORDINATION FUNCTION

PCF(polling)

Intended forContention-Free

ServicesUsed for all other services,

and used as basis for PCF

PCF: baiscally never user / supported!!

45

802.11 MAC evolution(802.11e, finalized in december 2005)

802.11 MAC evolution(802.11e, finalized in december 2005)

DCF

PCF(polling)

Intended for

Contention-FreeServices

Used for service

differentiation(priorities)

All enhancements rely on DCF basic operation!

Dead ☺

HCF ControlledChannel Access

HCCA(scheduling)

Enhanced DistributedChannelAccess

EDCA(prioritized CSMA)

Legacy

HYBRID COORDINATION FUNCTION

HCF

46

Carrier Sense Multiple AccessCarrier Sense Multiple Access

• Station may transmit ONLY IF senses channel IDLE for a DIFS time» DIFS = Distributed Inter Frame Space

• Key idea: ACK replied after a SIFS < DIFS» SIFS = Short Inter Frame Space

• Other stations will NOT be able to access the channel during thehandshake

» Provides an atomic DATA-ACK transaction

DIFSDATA

SIFS ACK

TX

RX

Packetarrival

OTHER

STA

DIFS

Packetarrival

Must measure

a whole DIFS

OK!

47

DATA/ACK frame formatDATA/ACK frame format

Frame

Control

Duration

/ IDAddress 1 Address 2 Address 3

Sequence

ControlAddress 4 Data

Framecheck

sequence

2 2 2 40-23126666

Frame

Control

Duration

/ IDAddress (RA)

Framecheck

sequence

2 2 46

DATA frame: 28 (or 34) bytes + payload

Protocol

versionType

2 2

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

ACK frame: 14 bytes – No need for TA address (the station receiving the ACK knows who’s this from)!!

Protocol

versionType

2 2

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

0 1Type = Control (01)

SubType = ACK (1101)1 1 0 1

Type = Data (10)SubType = Data (0000)

1 0 0 0 0 00 0

0 0 0 0 0 0 0 00 x

x x x x x xx x

48

Grasping wi-fi (802.11b) numbersGrasping wi-fi (802.11b) numbers

• DIFS = 50 µµµµs» Rationale: 1 SIFS + 2 slot-times

» Slot time = 20 µµµµs, more later

PHY MAC header 24 (30) Payload FCS

• SIFS = 10 µµµµs» Rationale: RX_TX turnaround time

» The shortest possible!

• DATA frame: TX time = f(rate)• Impressive PHY overhead!

» 192 µµµµs per every single frame

• Total data frame time (1500 bytes)» @1 Mbps: 192+12288= 12480 µµµµs

� PHY+MAC overhead = 3.3%

» @11 Mbps: 192+ 1117.1 = 1309.1 µµµµs� PHY+MAC overhead = 16.%

» Overhead increases for small frames!

• ACK frame: TX at basic rate» Typically 1 mbps but 2 mbps possible…

» ACK frame duration (1mbps): 304 µµµµs

Preamble SFD PLCP hdr

128 16 48

1 mbps DBPSK

192 µs

(28+payload) [bytes] x 8 / TX_rate [mbps] = µs

PHY ACK 14

192 µs

DATA

ACK

112 µs

49

And when an ACK is “hidden”?And when an ACK is “hidden”?

SENDER RECEIVERSTA

1)Sender TX

Receiver RX

STA defers

BUSY DETECT (DATA)

SENDER RECEIVERSTA

2)Receiver ACKs

(after SIFS)STA cannot hear…

SIFSACK

STASTA TX!DIFS

SENDER RECEIVERSTA

3)STA tranmits

And destroys ACK!

50

The Duration FieldThe Duration Field

Frame

Control

Duration

/ IDAddress 1 Address 2 Address 3

Sequence

ControlAddress 4 Data

Framecheck

sequence

2 2 2 40-23126666

0# microseconds

1514131211109876543210

When bit 15 = 1 � NOT used as duration(used by power-saving frames to specify station ID)

DIFSDATA

SIFS ACK

OTHER

STA

Physical carrier sensing

NAV (data)

• Allows “Virtual Carrier Sensing”» Other than physically sensing the channel, each station keeps a Network

Allocation Vector (NAV)» Continuously updates the NAV according to information read in the duration

field of other frames

Virtual carrier sensing

51

Issues with “duration” readingIssues with “duration” reading

RX

TXC

• “Duration” field in MAC header» Coded at same rate as payload» Must receive whole MAC frame correctly

11 Mbps tx

11 mbps

range5.5 mbps

range

2 mbps

range

1 mbps

range

• C cannot read TX frame» No way to know duration value

52

ACK may be hidden once again!ACK may be hidden once again!

RX

TXC

• C hidden from RX» Carrier sense remains IDLE during RX����TX ACK

» NAV could not be updated» May transmit after a DIFS» Destroying ACK!

ACK transm

Se si verifica la situazione mostrata nella slide precedente la stazione C non riuscirebbead allocare correttamente il NAV e la sua trasmissione si può sovrapporre all’ACK

53

EIFS = protect ACKEIFS = protect ACK

RX

TXC

• C cannot read data frame» CRC32 error» Most of PHY errors

11 Mbps tx

11 mbpsrange

5.5 mbps

range

2 mbpsrange

1 mbpsrange

• If planning to transmit:» No more after a DIFS» But after a LONGER interval of time

» Sufficiently long to protect ACK transmission

Quindi nel caso in cui la stazione C abbia un problema sulla ricezione di una trama, imposta un timer più lungo, chiamato EIFS prima di iniziare a trasmettere

54

EIFSEIFS

Data ACKNAV

SourcestationDestination stationOther stations receiving Data frame correctlyOther stationsreceiving Data frameincorectly

DIFSSIFS

EIFSBack- offBack- offBack- off

55

And when a terminal is “hidden”?And when a terminal is “hidden”?

RECEIVERSENDER STA

… this can be “solved” by increasing the sensitiveness of the Carrier Sense…Quite stupid, though (LOTS of side effects – out of the goals of this lecture)

SENDER STA

… this can’t be “solved”

by any means!

RECEIVER

• The Hidden Terminal Problem» SENDER and STA cannot hear each

other» SENDER transmits to RECEIVER» STA wants to send a frame

» Not necessarily to RECEIVER…» STA senses the channel IDLE

» Carrier Sense failure» Collision occurs at RECEIVER

• Destroys a possibly very long TX!!

56

DIFSDATA

SIFS ACK

TX

RX

Packetarrival

RTS

SIFS CTS SIFS

The RTS/CTS solutionThe RTS/CTS solution

TX

RX

hidden

others

RTS

NAV (RTS)

RTS/CTS: carry the amount of time the channel

will be BUSY. Other stations may update a

Network Allocation Vector, and defer TX

even if they sense the channel idle

(Virtual Carrier Sensing)

CTS CTS

NAV (CTS)

(Update NAV)

data

57

RTS/CTS framesRTS/CTS frames

Frame

Control

Duration

/ IDAddress (RA)

Framecheck

sequence

2 2 46

CTS frame: 14 bytes (same as ACK)

Protocol

versionType

2 2

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

0 1Type = Control (01)

SubType = CTS (1100)1 1 0 0

Frame

Control

Duration

/ IDAddress 1 (RA)

Framecheck

sequence

2 2 46

RTS frame: 20 bytes

Protocol

versionType

2 2

Sub TypeTo

DS

From

DS

More

FragRetry

Pwr

MNG

More

DataWEP Order

4 1 1 1 1 1 1 1 1

Type = Control (01)SubType = RTS (1011)

Address 2 (TA)

6

0 0 0 0 0 0 0 00 x

0 1 1 0 1 10 0 0 0 0 0 0 00 x

58

RTS/CTS and performanceRTS/CTS and performance

RTS/CTS cons: larger overhead

RTS/CTS pros: reduced collision duration

ESPECIALLY FOR LONG PACKETSLong ���� packet > RTS_Threshold (configurable)

TODAY higher rates � No more significant

59

Why backoff?Why backoff?

DIFSDATA

SIFS ACK

STA1

STA2

STA3

DIFS

Collision!

RULE: when the channel is initially sensed BUSY, station defers transmission;

THEN,when channel sensed IDLE again for a DIFS, defer transmission of a

further random time (Collision Avoidance)

60

Slotted BackoffSlotted Backoff

STA2

STA3

DIFS

Extract random number

in range (0, W-1)

Decrement every slot-time σ

w=7

w=5

Note: slot times are not physically delimited on the channel!

Rather, they are logically identified by every STA

Slot-time values: 20µs for DSSS (wi-fi)Accounts for: 1) RX_TX turnaround time

2) busy detect time

3) propagation delay

In 802.11 DCF the backoff counter is decremented at the end of the time slot

61

Backoff freezingBackoff freezing

• When STA is in backoff stage:» It freezes the backoff counter as long as the channel is sensed

BUSY

» It restarts decrementing the backoff as the channel is sensed IDLE for a DIFS period

DIFS DATA

SIFS ACK

STATION 1

DIFS

SIFS ACK 6 5

DIFS

Frozen slot-time 4

BUSY medium

STATION 2

DIFS

3 2 1

62

Why backoff betweenconsecutive tx?

Why backoff betweenconsecutive tx?

• A listening station would never find a slot-time after the DIFS (necessary to decrement the backoff counter)

• Thus, it would remain stuck to the current backoff counter valueforever!!

DIFS DATA

SIFS ACK

S 1

DIFS

6 5

DIFS

Frozen slot-time 4

BUSY medium

S 2

DIFS

3

DATA

SIFS ACK

DIFS

BUSY medium DIFS

63

Backoff rulesBackoff rules

• First backoff value:» Extract a uniform random number in range (0,CWmin)

• If unsuccessful TX:» Extract a uniform random number in range (0,2×(CWmin+1)-1)

• If unsuccessful TX:» Extract a uniform random number in range (0,22×(CWmin+1)-1)

• Etc up to 2m×(CWmin+1)-1

Exponential Backoff!

For 802.11b:

CWmin = 31

CWmax = 1023 (m=5)

64

Further backoff rulesFurther backoff rules

• Truncated exponential backoff

» After a number of attempts, transmission fails and frame is dropped

» Backoff process for new frame restarts from CWmin

» Protects against cannel capture

» unlikely when stations are in visibility, but may occur in the case of hidden stations

• Two retry limits suggested:

» Short retry limit (4), apply to frames below a given threshold

» Long retry limit (7), apply to frames above given threshold

» (loose) rationale: short frames are most likely generated by real time stations

» Of course not true in general; e.g. what about 40 bytes TCP ACKs?

65

DCF OverheadDCF Overhead

66

802.11b parameters (summary)802.11b parameters (summary)

PIFS used by Point Coordination Function

- Time-bounded services

- Polling schemePCF Never deployed

102331205010802.11b PHY

CWmaxCWminSlot Time

(µµµµsec)DIFS (µµµµsec)

SIFS

(µµµµsec)Parameters

67

DCF overheadDCF overhead

min_

[ ]

[ ] / 2station

Frame Tx

E payload

E T DIFS CWS =

+ +

_Frame Tx MPDU ACKT T SIFS T= + +

TxCTSPLCPCTS

TxRTSPLCPRTS

TxACKPLCPACK

TxMPDUPLCPMPDU

RTT

RTT

RTT

RLTT

_

_

_

_

/148

/208

/148

/)28(8

⋅+=

⋅+=

⋅+=

+⋅+=

_Frame Tx RTS CTS MPDU ACKT T SIFS T SIFS T SIFS T= + + + + + +

68

Example: maximum achievable throughput for802.11b

Example: maximum achievable throughput for802.11b

DATA SIFS

ACK

DIFS

backoff

DATA

Cycle time

• Data Rate = 11 mbps; ACK rate = 1 mbps• Payload = 1500 bytes

MbpsThr

BackoffE

DIFSSIFS

T

T

ACK

MPDU

07.631050304101303

81500

310202

31][

50;10

3041/148192

130311/)150028(8192

=++++

×=

=×=

==

=⋅+=

≈+⋅+=

• Data Rate = 11 mbps; ACK rate = 1 mbps• Payload = 576 bytes

MbpsThr

BackoffE

DIFSSIFS

T

T

ACK

MPDU

53.33105030410631

8576

310202

31][

50;10

3041/148192

63111/)57628(8192

=++++

×=

=×=

==

=⋅+=

≈+⋅+=

REPEAT RESULTS FOR RTS/CTS ���� Not viable (way too much overhead) at high rates!

69

DCF overhead (802.11b)DCF overhead (802.11b)

0 2000 4000 6000 8000

Transmssion Time (use c )

Basic

RTS/CTS

Basic

RTS/CTS

DIFS Ave Backoff RTS+SIFS CTS+SIFS Payload+SIFS ACK

70

DCF overhead (802.11b)DCF overhead (802.11b)

0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1

100 300 500 700 900 1100 1300 1500 1700 1900 2100 2300Payload Size (Bytes)

Normalized Throughput

BAS-2M bps

RTS-2M bps

BAS-11M bps

RTS-11M bps

71

802.11 multi rate operations and “performance anomaly”

802.11 multi rate operations and “performance anomaly”

72

Multi-rate operationMulti-rate operation

• Rate selection: proprietary mechanism!» Result: different chipsets operate widely different

• Two basic approaches» Adjust rate according to measured link quality (SNR estimate)

» How link quality is computed is again proprietary!

» Adjust rate according to frame loss

» How many retries? Step used for rate reduction? Proprietary!

» Problem: large amount of collisions (interpreted as frame loss) forces rate adaptation

73

Performance AnomalyPerformance Anomaly

• Question 1:» Assume that throughput measured for a single 11 mbps greedy

station is approx 6 mbps. What is per-STA throughput when two 11 mbps greedy stations compete?

• Answer 1:» Approx 3 mbps (easy ☺☺☺☺)

• Question 2:» Assume that throughput measured for a single 2 mbps greedy station

is approx 1.7 mbps. What is per-STA throughput when two 2 mbpsgreedy stations compete?

• Answer 2:» Approx 0.85 mbps (easy ☺☺☺☺)

• Question 3:» What is per-STA throughput when one 11 mbps greedy station

compete with one 2 mbps greedy station?

• Answer 3:» ...

74

Understanding Answers 1&2(neclect collision – indeed rare – just slightly reduce computed value)

STA 1 SIFS

ACK

DIFS

backoff

Cycle time

STA 2 SIFS

ACK

DIFS

Frozen backoff

• Data Rate = 11 mbps; ACK rate = 1 mbps

• Payload = 1500 bytes

]_[]2[]1[

81500

][

][]2[]1[

backofftotEDIFSACKSIFSTDIFSACKSIFSTtimecycleE

payloadEThrThr

MPDUMPDU ++++++++

×===

MbpsThr

backofftotE

DIFSSIFS

T

T

ACK

MPDU

3.3310)50304101303(2

81500

310202

31]_[

50;10

3041/148192

130311/)150028(8192

=++++×

×=

=×=

==

=⋅+=

≈+⋅+=

• Data Rate = 2 mbps; ACK rate = 1 mbps

• Payload = 1500 bytes

MbpsThr

backofftotE

DIFSSIFS

T

T

ACK

MPDU

88.0310)50304106304(2

81500

310202

31]_[

50;10

3041/148192

63042/)150028(8192

=++++×

×=

=×=

==

=⋅+=

≈+⋅+=

75

E[tot_backoff] indicato al denominatore deve essere considerato come il totale dei

tempi medi di backoff che si hanno dopo la trasmissione di ogni pacchetto nel

ciclo. Questi tempi di backoff dipendono dal numero di stazioni che stanno

competendo. Possiamo chiamare E[backoff(N)] la durata media dell’intervallo

tra la fine di un pacchetto e l’inizio del successivo, dove N è il numero di stazioni

che stanno competendo.

Quindi nella slide precedente E[tot_backoff] = E[backoff(2)]+ E[backoff(2)]

Come valutare il tempo di backoff quando ci sono più stazioni che competono?

nel caso in cui c'e' una sola stazione, ovviamente di ha:

E[backoff(1)]=SlotTime*CWmin/2=SlotTime*31/2

76

Si supponga che le stazioni A e B si alternino sul canale radio (ed in effetti in

media sia ha che la metà dei pacchetti è di una stazione e metà dell'altra).

La stazione A estrae il suo backoff e inizia a contare all'indietro, prima dello

scadere del backoff B trasmette (dato che abbiamo ipotizzato l'alternanza nelle

trasmissioni) e A sospende il conteggio all'indietro. A riprende il conteggio

all'indietro appena B finisce di trasmettere, finisce il conteggio senza essere

interrrotta (sempre per l'ipotesi dell'alternanza) e poi inizia a trasmettere il

nuovo pacchetto chiudendo il ciclo.

Se guardiamo quello che è successo in un ciclo, dall'inizio della trasmissione di A

all'inizio della trasmissione successiva di A, ci sono stati due periodi di backoff, la

cui durata media è la metà del periodo medio di backoff che la stazione A

avrebbe se trasmettesse in isolamento.

Quindi E[backoff(2)] = E[backoff(1)]/2 e nel caso in esame:

E[tot_backoff]=E[backoff(2)]+E[backoff(2)]=E[backoff(1)]=SlotTime*31/2

Il tutto, ricordiamocelo, nell'ipotesi di trascurare le collisioni !!!

77

Emerging “problem”: long-term fairness!

Emerging “problem”: long-term fairness!

If you have understood the previous example, you easilyrealize that 802.11 provides FAIR access to stations in terms of EQUAL NUMBER of transmissionopportunities in the long term!

But this is INDEPENDENT OF transmission speed!

STA1 STA2 STA1 STA1STA2 STA2

78

Computing answer 3Computing answer 3

STA (2mbps) SIFS

ACK

DIFS

Cycle time

STA 11 SIFS

ACK

DIFS

Frozen backoff

RESULT: SAME THROUGHPUT (in the long term)!!

!!!!!!39.1310)5030410(213036304

81500

]_[]2[]1[

81500

][

][]2[]1[

Mbps

backofftotEDIFSACKSIFSTDIFSACKSIFST

timecycleE

payloadEThrThr

MPDUMPDU

=+++++

×=

=++++++++

×=

===

DRAMATIC CONSEQUENCE: per-station throughput is limited by

STA with slowest rate (lower that the maximum throughputachievable by the slow station)!!

79

Performance anomaly into actionPerformance anomaly into action

Why the network is

soooo slow today? We’re so

Close, we have a 54 mbps and“excellent” channel, and we get

Less than 1 mbps …

Hahahahahah!!

Poor channel, Rate-fallbacked @ 1mbps ☺