Cross layer design for Wireless networks

Kavé Salamatian

LIP6-UPMC

Future Wireless Systems

Nth Generation CellularWireless Internet AccessWireless Video/Music Wireless Ad Hoc NetworksSensor Networks Smart Homes/AppliancesAutomated Vehicle NetworksAll this and more…

Ubiquitous Communication Among People and Devices

Next generation network architecture

MobilityServices

Layer

RadioAccessLayer

MobileTerminal

Layer

InternetworkingLayer

LocalServiceLayer

NetworkServiceLayer

AccessManagement

Layer

AccessInterface

Layer

WirelessInterface

Layer

MobileApplication

Layer

Internet Wireless PSTN

Radio Access Network

Radio Access NetworkMobile User Equipment(e.g. Win9X, Palm OS)

Mobile User Equipment(e.g. Win9X, Palm OS)

Application

Network Server(e.g. WinNT, Unix)

Network Server(e.g. WinNT, Unix)

IP Transport(TCP, UDP, RTP)

Internet Protocol(IP)

RadioAccess

RadioAccessModemModemEthernetEthernet

Application

IP Transport(TCP, UDP, RTP)

Internet Protocol(IP)

EthernetEthernet ATMATMInternet

TransportAgents

TransportAgents

IP

RadioResourceMgmt

RadioL1

AccessL1

RadioL2

AccessL2

AccessL1

CoreL1

AccessL2

CoreL2

Radio-Optimized IP Networking• Transparent to TCP/IP protocols• Enables deployment of IP-based consumer applications in next generation wireless systems

Separation principles

Application, transport and physical layer can be separated if : No errors at physical layer No losses and delays at

transport layer No fluctuations in applications

rate Each layer being perfect from

the point of view of other layers

Application

Transport

Physical

Signal

Packet

Bits

Challenges

Wireless channels are a difficult and capacity-limited broadcast communications medium

Traffic patterns, user locations, and network conditions are constantly changing

Applications are heterogeneous with hard constraints that must be met by the network

Energy and delay constraints change design principles across all layers of the protocol stack

These challenges apply to all wireless networks, but are amplified in ad hoc/sensor networks

Why is Wireless Hard?The Wireless Channel Fundamentally Low Capacity: R< B log(1+SINR) bps

Spectrum scarce and expensive Received power diminishes with distance Self-interference due to multipath Channel changes as users move around Signal blocked by objects (cars, people, etc.) Broadcast medium – everyone interferes

d

…And The Wireless Network

Link characteristics are dynamic Network access is unpredictable and hard to coordinate Routing often multihop over multiple wireless/wired

channels Network topology is dynamic Different applications have different requirements

Wireline Backbone

Design objective

Want to provide end-to-end Properties

The challenge for this system is dynamics Scheduling can help shape these dynamics

Adaptivity can compensate for or exploit these dynamics

Diversity provides robustness to unknown dynamics

Scheduling, adaptivity, and diversity are most powerful in the context of a crosslayer design

Energy must be allocated across all protocol layers

Multilayer Design

Hardware Power or hard energy constraints Size constraints

Link Design Time-varying low capacity channel

Multiple Access Resource allocation (power, rate, BW) Interference management

Networking. Routing, prioritization, and congestion control

Application Real time media and QOS support Hard delay/quality constraints

Multilayer Design

Crosslayer TechniquesAdaptive techniques

Link, MAC, network, and application adaptation Resource management and allocation (power control) Synergies with diversity and scheduling

Diversity techniques Link diversity (antennas, channels, etc.) Access diversity Route diversity Application diversity Content location/server diversity

Scheduling Application scheduling/data prioritization Resource reservation Access scheduling

Key Questions

What is the right framework for crosslayer design?What are the key crosslayer design synergies?How to manage its complexity?What information should be exchanged across layers,

and how should this information be used? How do the different timescales affect adaptivity? What are the diversity versus throughput

tradeoffs? What criterion should be used for scheduling? How to balance the needs of all

users/applications?

Single user example

1 2 3 4 5 6 7 8 9 1010

-6

10-5

10-4

10-3

10-2

10-1

100

WIFI : (171,133)

SNR

Packet

Err

or

Rate

Adaptive Modulation and Coding in Flat Fading

Adapt transmission to channel Parameters: power,rate,code,BER, etc. Capacity-achieving strategy

Recent Work Adaptive modulation for voice and data (to meet QOS) Adaptive turbo coded modulation (<1 db from capacity) Multiple degrees of freedom (only need exploit 1-2) Adaptive power, rate, and compression with hard deadlines

(t)

UncodedData Bits Buffer

M()-QAM ModulatorPower: S()

To Channel(t)

PointSelector

log2 M() Bits One of theM() Points

BSPK 4-QAM 16-QAM

Crosslayer design in multiuser systems

• Users in the system interact (interference, congestion)

• Resources in the network are shared• Adaptation becomes a “chicken and

egg” problem• Protocols must be distributed

Wireless networks

They are formed by nodes with radios There is no a priori notion of “links” Nodes simply radiate energy

Nodes Cooperation

Decode and forward Why not: Amplify and

Forward

Increase Signal for Receiver

Why not: Reduce Interference at Receiver

How should node cooperates ?

Some obvious choices Should nodes relay packets? Should they amplify and forward? Or should they decode and forward? Should they cancel interference for other nodes? Or should they boost each other’s signals? Should nodes simultaneously broadcast to a group of

nodes? Should those nodes then cooperatively broadcast to

others? What power should they use for any operation? …

Or should they use much more sophisticated unthought of strategies?

Example: Six Node Network

Capacity Regions (Goldsmith)

(a): Single hop, no simultaneous transmissions.(b): Multihop, no simultaneous transmissions. (c): Multihop, simultaneous transmissions.(d): Adding power control (e): Successive interference cancellation, no power control.

jiijRij ,34,12 ,0

Multiplehops

Spatial reuse

SIC

Optimal Routing

The point is achieved by the following scheduling :

MbpsRR 64.13412

Idea: Adapt transmission rate according to channel quality Change modulation to get higher rate if channel is good Could send multiple packets at higher rates (A

suggested cscheme)

Protocol details RTS/CTS and Broadcast packets sent at lowest rate Receiver measures strength of RTS Communicates rate to sender in CTS DATA and ACK at that rate

Adaptive Rate MAC (Kumar)

Interaction with Min Hop Routing Protocol

Most current routing protocols are min hop Consider DSDV for example Chooses long hops But long hops => low signal strength => low rates

Switching off adaptation is better

Routing based approach

Luigi & al.

Routing in wireless network

« Shortest path approche is not optimal » Physical channel is instable Each transmission inject interference in the

network Interference reduce capacity

Power management is needed Make use of multi-rate and power control on WIFI card

L’architecture en couches n’est pas optimale

Cross Layer approch

Maximise throughput

Gupta & Kumar

Throughput

Node number

Transmission range

Rate

To maximise throughput we have to maximise transmission rate and reduce interference generated by each packets

Capacity Constraints

Cross-Layer Approach

SIRInterface characteristics

Next-HopData-RateTransmission power

Routing metric Rate Interference Packet Error Rate

Interference

Measurement: unrealistic More neighbor => More interference More power => More interference Defining a interference replacement function I(P) Minimise I(P) => Minimise Real interference

Packet Error Rate (I)

MAC

Convolution Coder

Interleaver

Modulator &Scrambler

Channel

Rake Receiver

Deinterleaver

Viterbi Decoder

MAC

InterferenceNoise

(White or fading)

IP packet IP packet

Multiple AntennaMultiple Antenna

Single AntennaSingle Antenna

Packet Error Rate (III)

LEP

BERSIRPER

f

f

Routing Strategy

• Rate (Mbps)•Maximise

•Interference (mW)• Minimise

•PER• Minimise

•Power (mW)• trade off for optimising routing parameter

•NP-Complet Problème

Routingless approach

Ramin & al.

Ad-Hoc Network

Ad Hoc Networks function by multi-hop transport Nodes relay packets until they reach their destinations

Must of the traffic carried by the nodes is relay traffic The actual useful traffic per user pair is small

Intermediate nodes relay the same information Duplicated information might be received by the receiver

More intelligent relaying is needed

Which packet to relay Which information to relay

• The relay nodes must only send useful information

Coding for erasure channels MDS (Maximum Distance Separable) codes

Get k packets, generates n-k redundant packets Each combination of k packets out of n enable to retrieve

the initial packets Generating matrix

• Each submatrix of is invertible

Reed Solomon codes are MDS We suppose that sender generates m redundant

packets We suppose that relay generates l packets

How to choose m and l to achieve the bound

( )k k k n kC I B

( )k n kB

Achievability of the capacity bound for the more capable case

Choose a code length n. Knowing packet loss matix of the netwok R and can be determined. We chose then

The code is a MDS code The receiver is able to retrieve the k initial packets if it receives at

least k packets from sender and relay together This happen asymptotically with large n if the rate validate

the bound

10)( lkknkkk BBIC

optnlnRk ,

0)( knkkk BIWX

11 lkBWX

CXXW 1,

10)(

knkkk BIXW

WW p

1p 2p

opt

Comments & practical consideration

Relay send only useful side information over the channel

The relay load is chosen as the minimal value which maximize the global rate

Each sender and relay can derivate the number of needed redundant packets if it know the packet loss probability matrix

The proposed scheme can be implemented very easily in WiFi based wireless network Does not need any change to physical layer

Practical implementation

15 node distributed randomly in the environment One Sender-Receiver pair is chosen randomly each node have two cart WiFi, with different frequency

channels f1 and f2 If one node receive the packets

It can be a relay with probability p The relay nodes broadcast information in the

environment There is not any routing protocol

It is done in NS

Topology

0 100 200 300 400 500 6000

100

200

300

400

500

600

Sender

Receiver

Throughput and relay load

10-3

10-2

10-1

100

102

103

104

105

106

10-3

10-2

10-1

100

0

5

10

15

20

25

30

35

Toward Software radio

• Common technology for multiple radio platforms

NetworkI/F

Dup

Channel-izer

D/AUpcon-verter

A/DLNA RF/IF Rx Chan

Tx ChanMCPA

Cellsite controller middleware

CommonDSP

platform

Antenna

ATMWideband transceiver

Inte

rfac

e

Inte

rfac

e

Inte

rfac

e

Conclusions

Crosslayer design needed to meet requirements and constraints of future wireless networks

Key synergies in crosslayer design must be identified The design must be tailored to the application Crosslayer design should include adaptivity, scheduling and

diversity across protocol layers Energy can be a precious resource that must be shared by different

protocol layers

Coming Challenges MIMO: how to take advantage of Multiple Antenna Software Radio: How to enable adaptation of physical layer

from upper layer

Interesting Question

MIMO or Ad Hoc, that’s the question? Routing can be seen as a diversity

Not shortest path !