Presentation is loading. Please wait.

Presentation is loading. Please wait.

Improving Loss Resilience with Multi-Radio Diversity in Wireless Networks Allen Miu, Hari Balakrishnan MIT Computer Science and Artificial Intelligence.

Similar presentations


Presentation on theme: "Improving Loss Resilience with Multi-Radio Diversity in Wireless Networks Allen Miu, Hari Balakrishnan MIT Computer Science and Artificial Intelligence."— Presentation transcript:

1 Improving Loss Resilience with Multi-Radio Diversity in Wireless Networks Allen Miu, Hari Balakrishnan MIT Computer Science and Artificial Intelligence Laboratory C. Emre Koksal Computer and Communication Sciences, EPFL

2 2 The problem New wireless applications demand high performance But: wireless channels are loss-prone  Interference  Noise  Attenuation  Multi-path  Mobility, etc…  Inconsistent and poor performance

3 3 Current solutions are inefficient for recovering losses Coding (e.g., FEC)  Hard on highly variable channels Retransmission  Wasteful: outage durations can be long (tens to hundreds of milliseconds) Bit-rate adaptation  Hard on variable channels  Slows down other clients

4 4 May use only one path Uses only one communication path Today’s wireless LAN (e.g., 802.11) AP 1 Interne t

5 5 May use only one path Allow multiple APs to simultaneously receive transmissions from a single transmitter Today’s wireless LAN (e.g., 802.11) AP 1 AP 2 Multi-Radio Diversity (MRD) – Uplink Interne t MRDC 10% 20% Loss independence  simultaneous loss = 2%

6 6 Multi-Radio Diversity (MRD) – Downlink Interne t AP 1 AP 2 MRDC Allow multiple client radios to simultaneously receive transmissions from a single transmitter

7 7 Are losses independent among receivers? Broadcast 802.11 experiment at fixed bit-rate: 6 simultaneous receivers and 1 transmitter Compute loss rates for the 15 receiver-pair (R1, R2) combinations  Frame loss rate FLR(R1), FLR(R2) vs. simultaneous frame loss rate FLR(R1 ∩ R2)

8 8 Individual FLR > Simultaneous FLR FLR(R1 ∩ R2) FLR y = x R1 R2 R1*R2

9 9 Challenges in developing MRD How to correct simultaneous frame errors?  Frame combining How to handle retransmissions in MRD?  Request-for-acknowledgment protocol How to adapt bit rates in MRD?  MRD-aware rate adaptation

10 10 2. Select bit combination at unmatched bit locations, check CRC PatternsCRC Ok Bit-by-bit frame combining 1100 0000 R1 1101 1010 R2 0001 1010 1. Locate bits with unmatched value 1100 0000-- 1100 0010X 1100 1000X 1100 1010O Corrected frame TX: 1100 1010 Problem: Exponential # of CRC checks in # of unmatched bits. Combine failure 1 1 0

11 11 Block-based frame combining Observation: bit errors occur in bursts Divide frame into N B blocks (e.g., N B = 6) Attempt recombination with all possible block patterns until CRC passes # of checks upper bounded by 2 N B  Failure rate increases with N B

12 12 1.0 Failure decreases with N B and burst size Burst error length parameter Probability of failure N B = 2 N B = 4 N B = 6 N B = 16 … 0.8 0.6 0.4 0.2 0 0 1020304050 Frame size = 1500B

13 13 Flawed retransmission schemes Conventional link-layer ACKs do not work  Final status known only to MRDC Two levels of ACKs are redundant  Cannot disable link-layer ACKs

14 14 Request-for-acknowledgment (RFA) for efficient feedback link IP link MRDC link MRD DATA ACK DATA RFA MRD-ACK DATA IP

15 15 MRD-aware rate adaptation Standard rate adaptation does not work  Reacts only to link-layer losses from 1 receiver  Uses sub-optimal bit-rates MRD-aware rate adaptation  Reacts to losses at the MRD-layer Implication: First use multiple paths, then adapt bit rates.

16 16 Experimental setup ~20 m R1 R2 L 802.11a/b/g implementation in Linux (MADWiFi) L transmits 100,000 1,500B UDP packets w/ 7 retries 802.11a @ auto bit rate (6, 9, 12, 18, 24, 36, 48, 54) L is in motion at walking speed, > 1 minute per trial Variants: R1, R2, MRD (5 trials each)

17 17 MRD improves throughput 2.3x Improvement R1 R2MRD Throughput (Mbps) 8.25 Mbps 18.7 Mbps Each color shows a different trial

18 18 MRD maintains high bit-rate Frame recovery data (% of total losses at R1) via R2 42.3% frame combining 7.3% Total 49.6% Selected bit rate (Mbps) Fraction of transmitted frames 6 0 9121824364856

19 19 Delay Analysis One way delay (10 x s) Fraction of delivered packets 10 -3 10 -4 0 10 -2 10 -1 1 User space implementation caused high delay

20 20 Related work Physical layer spatial diversity techniques  Antenna diversity  802.16 MIMO/802.11n Macro-diversity in CDMA networks Retransmission with memory [Sindhu ’77] Opportunistic forwarding [Biswas ‘05][Jain ’05] Bit rate selection (AARF, RBAR, MiSer, OAR, Sample Rate)

21 21 Summary Design of Multi-Radio Diversity WLAN  Block-based frame combining  Request-for-acknowledgment protocol  MRD-aware rate adaptation Analysis of block-based frame combining Experimental evaluation  MRD reduces losses by 50% and improves throughput by up to 2.3x


Download ppt "Improving Loss Resilience with Multi-Radio Diversity in Wireless Networks Allen Miu, Hari Balakrishnan MIT Computer Science and Artificial Intelligence."

Similar presentations


Ads by Google