bringing life to dead spots

30
Bringing Life to Dead Spots Grace Woo Pouya Kheradpour, Dawei Shen, and Dina Katabi

Upload: paniz

Post on 12-Jan-2016

35 views

Category:

Documents


0 download

DESCRIPTION

Bringing Life to Dead Spots. Grace Woo Pouya Kheradpour, Dawei Shen, and Dina Katabi. Many APs But Still Poor Coverage. mit1. mit4. mit5. Problem increases with mobility and low power devices. Poor Coverage Is Not No Coverage!. X. X. 010101011111. 011101011011. Loss. Loss. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Bringing Life to Dead Spots

Bringing Life to Dead SpotsGrace Woo

Pouya Kheradpour, Dawei Shen, and Dina Katabi

Page 2: Bringing Life to Dead Spots

Many APs But Still Poor Coverage

Problem increases with mobility and low power devices

Problem increases with mobility and low power devices

mit1

mit4

mit5

Page 3: Bringing Life to Dead Spots

Poor Coverage Is Not No Coverage!

010101011111X 011101011011X

Can recover a correct packet if we combine the correct bits from these receptions

Spatial Diversity APs are unlikely to have same bit error

LossLoss

Currently High Bit Error Persistent Loss Dead Spot

Page 4: Bringing Life to Dead Spots

But Which AP Got the Right Bit?

• Clearly can’t have per bit checksum

• Prior work (MRD) tries all block combinations to satisfy checksum

• Exponential Complexity• Works for a few bit errors But not dead spots

First bit is “0”First bit is “1”

Page 5: Bringing Life to Dead Spots

SOFT• Recovers a correct packet from its

faulty receptions at APs

• Leverages physical layer hints to identify correct bits

• SOFT’s delivery rate is up to10x higher than current WLANs and MRD

Page 6: Bringing Life to Dead Spots

SOFT Architecture

APs leverage high-speed Ethernet to combine their receptions

Internet

AP1 AP2

SOFT Combiner

APn

Wired Ethernet

But which bits are correct?But which bits are correct?

010001001111X010101001011X

010001001011

Page 7: Bringing Life to Dead Spots

Physical Layer Knows More!

PHY already estimates a confidence in its 0-1 decision Soft Value

PHY 01011001

0.0

0.2

0.4

0.6

0.8

1.0

-1.0 -0.5 0.0 0.5 1.0

Measured Soft Values

PDF of per bit soft values

Larger absolute soft values More confidence in bit

Larger absolute soft values More confidence in bit

Soft Value < 0 “0”

Soft Value > 0 “1”

Page 8: Bringing Life to Dead Spots

We Use the Soft Values• SOFT changes the PHY interface to expose the soft values to higher layers• SOFT combines the soft values of a bit to decode it correctly

• The combiner forwards the decoded packet if it satisfies the 802.11 checksum

AP1AP2

SOFT Combiner

Soft packet Soft packet

010110111

Page 9: Bringing Life to Dead Spots

How Do We Combine Soft Values?

0.4 - 0.1 - 0.2

How do we decode the bit?• Maximum soft value Bit is “1”• Majority vote Bit is “0”• Average Bit is “1”

Say for a particular bit, we got

Different Combining Methods Different Answers!

Page 10: Bringing Life to Dead Spots

SOFT Combining Algorithm

"0" else "1" is Bit 02

jS

ii

ij

2iLet be the noise variance on the channel to

APi

For AWGN and dead spots rule is proven optimal.

Let Sij be the soft value of bit j reported by APi

SOFT decision rule:

Intuitively, we want to favor less noisy channels

Page 11: Bringing Life to Dead Spots

But, How Does SOFT Get the Noise Variance?

Randomness in soft values is caused by channel noise

Estimate from the PDF of the soft values in packet2i

0.0

0.2

0.4

0.6

0.8

1.0

-1.0 -0.5 0.0 0.5 1.0

Measured Soft Values

PDF of per bit soft values

Page 12: Bringing Life to Dead Spots

How About Overhead?

• PHY soft values can be 32-bit float Excessive Ethernet traffic

Solution• Invoke SOFT only when associated

AP can’t decode• Quantize soft values (we used 3 bits)

Page 13: Bringing Life to Dead Spots

What About the Downlink?

Use Time Diversity Combine a packet with its retransmission

Use Time Diversity Combine a packet with its retransmission

010001001011X

010101000011X

Page 14: Bringing Life to Dead Spots

Performance

Page 15: Bringing Life to Dead Spots

SOFT Implementation• Software – GNURadio

codebase• Hardware – USRP frontend• GMSK and DBPSK

modulations• Soft values are inputs to

the slicer• Poor Coverage:

• SNR 5 – 12 dB

• BER about 10-3

Page 16: Bringing Life to Dead Spots

Experimental Setup• 13 GNURadio nodes• Compared

– Current 802.11 WLAN (user associates with best AP)

– MRD– SOFT

• Each Experiment– 3 random APs – Random source – Transmit 500 packets

Page 17: Bringing Life to Dead Spots

Does SOFT Help?CDF of 100 experiments

Packet Delivery Rate

Page 18: Bringing Life to Dead Spots

Does SOFT Help?CDF of 100 experiments

Packet Delivery Rate

Page 19: Bringing Life to Dead Spots

Does SOFT Help?CDF of 100 experiments

Packet Delivery Rate

Page 20: Bringing Life to Dead Spots

Does SOFT Help?CDF of 100 experiments

Packet Delivery RateSOFT’s delivery rate can be 10x

higher SOFT’s delivery rate can be 10x

higher

10x

Page 21: Bringing Life to Dead Spots

Packet Delivery Rate

Bit Error Rate

Performance with Increasingly Poor Coverage

Page 22: Bringing Life to Dead Spots

Packet Delivery Rate

Bit Error Rate

Performance with Increasingly Poor Coverage

Current Approach

Page 23: Bringing Life to Dead Spots

Packet Delivery Rate

Bit Error Rate

Current ApproachMRD

Performance with Increasingly Poor Coverage

Page 24: Bringing Life to Dead Spots

Packet Delivery Rate

Bit Error Rate

Current Approach

SOFTMRD

SOFT Addresses Dead SpotsSOFT Addresses Dead Spots

Performance with Increasingly Poor Coverage

Page 25: Bringing Life to Dead Spots

0.00

0.20

0.40

0.60

0.80

1.00

0 5 10 15 20 25 30

Current Approach

CDF over 50,000 packets

Number of Retransmissions Until Correct Packet

SOFT on Downlink

Page 26: Bringing Life to Dead Spots

0.00

0.20

0.40

0.60

0.80

1.00

0 5 10 15 20 25 30

Current ApproachSOFT

CDF over 50,000 packets

SOFT on Downlink

17 ReTx

Number of Retransmissions Until Correct PacketMuch Higher Throughput!Much Higher Throughput!

Page 27: Bringing Life to Dead Spots

Combining Method Is Important

CDF of 100 experiments

Packet Delivery Rate

MAX

SOFTMajority

SOFT Outperforms MAX and MAJORITY SOFT Outperforms MAX and MAJORITY

Page 28: Bringing Life to Dead Spots

Effect of Quantization SOFT Average Delivery Rate

3 Bits32 Bits

2 Bits

All presented results are for 3-bit quantization!Overhead on Wired Ethernet is AcceptableOverhead on Wired Ethernet is Acceptable

Page 29: Bringing Life to Dead Spots

Related Work

• Soft and softer handoff in cellular networks

• Theoretical Maximum Ratio Combining (MRC) [Brennan55,Yang99]

• H-ARQ & Chase Combining [ASX03]

• Partial Packet Recovery [Jam07]

Page 30: Bringing Life to Dead Spots

Conclusion• WLAN can have better coverage if the

interface to the PHY exposes soft values

• Delivery rate can be up to 10x higher

• Ethernet overhead is acceptable

• The new architecture, SOFT, can co-exist with unmodified 802.11 cards and APs