packet radio: aloha

24
CS 536 Park Packet radio: ALOHA Base Station Stationary F1 F1’ F1’ F1 F1 F1’ ALOHA Stationary Stationary Stationary Stationary −→ downlink broadcast channel F 1 −→ shared uplink channel F 1 Ex.: ALOHANET data network over radio frequency Univ. of Hawaii, 1971; 4 islands, 7 campuses

Upload: others

Post on 31-Dec-2021

11 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Packet radio: ALOHA

CS 536 Park

Packet radio: ALOHA

Base Station

Stationary

F1

F1’

F1’

F1F1

F1’

ALOHA

Stationary

Stationary

Stationary

Stationary

−→ downlink broadcast channel F1

−→ shared uplink channel F1′

Ex.: ALOHANET

• data network over radio frequency

• Univ. of Hawaii, 1971; 4 islands, 7 campuses

Page 2: Packet radio: ALOHA

CS 536 Park

• Norm Abramson

→ precursor to Ethernet

→ parallel to wired packet switching technology

• carrier frequency

→ uplink: 407.35 MHz; downlink: 413.475 MHz

• bit rate: 9.6 kb/s

• contention-based multiple access: MA

→ plain and simple

→ needs explicit ACK frames (stop-and-wait)

Page 3: Packet radio: ALOHA

CS 536 Park

Wireless LAN (WLAN): infrastructure mode

Access Point

F1

F1

F1

F1F1

F1

MobileMobile

Mobile

Mobile

Mobile

Mobile

Mobile

WLAN: Infrastructure Network

−→ shared uplink & downlink channel F1

• basic service set (BSS)

→ “hot spot”

• SSID (service set identifier): name/label of BSS

• base station: access point (AP)

• mobile stations must communicate through AP

Page 4: Packet radio: ALOHA

CS 536 Park

WLAN: ad hoc mode

F1

F1

F1

F1F1

F1

MobileMobile

Mobile

Mobile

Mobile

Mobile

MobileMobile

F1

F1

F1

F1

F1

WLAN: Ad Hoc Network

−→ homogeneous: no base station

−→ everyone is the same

−→ share forwarding responsibility

• independent basic service set (IBSS)

• mobile stations communicate peer-to-peer

→ also called peer-to-peer mode

Page 5: Packet radio: ALOHA

CS 536 Park

WLAN: internetworking

Access Point

F1

F1

F1

F1F1

F1

MobileMobile

Mobile

Mobile

Mobile

Mobile

Mobile

Access Point

F1

F1

F1

F1F1

F1

MobileMobile

Mobile

Mobile

Mobile

Mobile

Mobile

Access Point

F1

F1

F1

F1F1

F1

MobileMobile

Mobile

Mobile

Mobile

Mobile

Mobile

WLAN: Extended Service Set

Distribution System

−→ internetworking between BSS’s through APs

−→ mobility and handoff

• extended service set (ESS): shared SSID

• APs are connected by distribution system (DS)

→ typically: Ethernet switch

Page 6: Packet radio: ALOHA

CS 536 Park

How do APs and Ethernet switches know where to for-

ward frames?

→ spanning tree

→ IEEE 802.1 (Perlman’s algorithm)

Learning bridge: source address discovery

→ log source MAC address of incoming frames per inter-face

→ initially (or if unclear): broadcast

→ simple form of routing

→ adequate for small systems

Misconfiguration issues resulting in loops

→ modifications to spanning tree algorithm

Page 7: Packet radio: ALOHA

CS 536 Park

Additional headache: mobility

−→ roaming

−→ how to perform handoff

−→ mobility management at link vs. network layer

−→ link layer handoff dominant (vs. Mobile IP)

Mobility between BSS’s in an ESS

• Association

→ registration process

→ AP sends out periodic beacon frame

→ mobile station (MS) associates with one AP

• Disassociation

→ upon permanent departure: notification

Page 8: Packet radio: ALOHA

CS 536 Park

Handoff from old to new AP:

• Reassociation

→ movement of mobile from one AP to another

→ mobile initiated

→ e.g., AP’s signal strength is low

→ passive (beacon) or active (probe) scanning to findalternate AP

→ go through association process

• Handoff

→ inform new AP of old AP

→ forwarding of buffered frames from old to new APin ESS

Page 9: Packet radio: ALOHA

CS 536 Park

IEEE 802.11b/g WLAN spectrum 2.4–2.4835 GHz:

→ 11 channels (U.S.)

→ 2.412 GHz, 2.417 GHz, . . ., 2.462 GHz

→ unlicensed ISM (Industrial, Scientific, Medical) band

→ global: 2.4–2.4835 GHz

→ up to 14 channels (e.g., Japan)

IEEE 802.11a: 5.15–5.35 GHz and 5.725–5.825 GHz

→ UNNI (unlicensed National Information Infrastruc-ture)

→ 23 non-overlapping channels

Page 10: Packet radio: ALOHA

CS 536 Park

IEEE 802.11n: both 2.4 and 5 GHz

→ 2.4 GHz: backward compatible

→ also uses multiple antennae

→ called MIMO (multiple input multiple output)

IEEE 802.11ac: extension of n/g with more streams,wider bandwidth, 256-QAM

→ 802.11ax in the works with 1024-QAM, OFDMA

Page 11: Packet radio: ALOHA

CS 536 Park

Non-interference specification for 802.11b:

• each channel has 22 MHz bandwidth

• require 25 MHz channel separation

−→ thus, only 3 concurrent channels possible

−→ e.g., channels 1, 6 and 11

−→ 3-coloring. . .

Page 12: Packet radio: ALOHA

CS 536 Park

HAAS (Old CS Building):

First floor frequency reuse:

Page 13: Packet radio: ALOHA

CS 536 Park

Second floor frequency reuse:

Ground floor frequency reuse:

Page 14: Packet radio: ALOHA

CS 536 Park

IEEE 802.11 WLAN:

→ uses CSMA as MAC

However:

• 802.11a/g/n: use OFDM

→ single-user (not OFDMA)

→ e.g.: 801.11g uses 48 subcarriers

→ 312.5 KHz separation

→ 1st bit on subcarrier 1, 2nd bit on subcarrier 2, . . .

• 802.11b: uses DSSS CDMA

→ 11-bit chip sequence

→ single-user

Why not just use multi-user OFDMA or CDMA?

→ two main reasons

Page 15: Packet radio: ALOHA

CS 536 Park

Empty room:

→ large lecture room

→ no obstructions

→ 802.11 WLAN hot spot (infrastructure mode)

→ how does indoor signal reception look like?

Page 16: Packet radio: ALOHA

CS 536 Park

Signal strength reception at height 0.7 m:

→ approximately table height

→ carrier frequency: 5.32 GHz

-10 0 10 20 30 40

AP->stations (5.32GHz)

0 5 10 15 20 25 30 35 40 0

5

10

15

20

25

30

→ signal strength varies by distance

→ but also varies by location

→ connection to microwave oven?

Page 17: Packet radio: ALOHA

CS 536 Park

Multi-path fading:

→ EM waves interfere constructively and destructively

→ depends on phase

Locations of destructive interference:

→ bad signal reception even though close to AP

Locations of constructive interference:

→ good signal reception even though far from AP

What impact does it have?

Page 18: Packet radio: ALOHA

CS 536 Park

Throughput:

→ SNR and throughput along straight line from AP

0

20

40

0 5 10 15 20 25 30 35 40

SN

R [

dB

]

x-coordinate

y=14m

→ significant SNR variation

0

20

40

60

0 5 10 15 20 25 30 35 40

Rate

[Mbps]

x-coordinate

y=14m

→ good locations, bad locations

→ leads to unfairness, even starvation

Page 19: Packet radio: ALOHA

CS 536 Park

Indoor office 802.11 WLAN hot spot:

Page 20: Packet radio: ALOHA

CS 536 Park

Throughput share of 16 HP/Compaq pocket PCs:

→ uplink CSMA competition

0

0.1

0.2

0.3

0.4

0.5

0.6

4 4.5 5 5.5 6 6.5 7 7.5 8

Thr

ough

put (

Mb/

s)

Offered Load (Mb/s)

node 1node 2node 3node 4node 5node 6node 7node 8node 9

node 10node 11node 12node 13node 14node 15node 16

→ offered load: stress placed on the system

→ significant unfairness

→ what can be done?

Page 21: Packet radio: ALOHA

CS 536 Park

Changing carrier frequency:

→ 5.805 GHz (channel 12)

→ qualitatively similar to channel 8

→ quantitatively different

→ use OFDM to send bits on different subcarriers

→ combined with forward error correction

Page 22: Packet radio: ALOHA

CS 536 Park

IEEE 802.11 MAC:

−→ CSMA/CA with exponential backoff

−→ explicit positive ACK frame

−→ added optional feature: CA (collision avoidance)

Two modes for MAC operation:

• Distributed coordination function (DCF)

→ uses CSMA

• Point coordination function (PCF)

→ polling-based priority

→ telephony support

PCF, CA not used in typical real-world deployments.

Page 23: Packet radio: ALOHA

CS 536 Park

Additional issues with CSMA in wireless media:

Hidden station problem:

MobileMobile Mobile

A B C

Hidden Station Problem

(1) (2)

(3)

(1) A transmits to B

(2) C does not sense A; transmits to B

(3) interference occurs at B: i.e., collision

Page 24: Packet radio: ALOHA

CS 536 Park

Hidden station problem: RTS/CTS handshake “clears”

hidden area

MobileMobile Mobile

A B CRTS

CTS CTS

RTS/CTS Handshake

"clears the area"

RTS/CTS perform only if data > RTS threshold

−→ why not for small data?

. . . feature available but not used