Wireless Networking
Fundamentals and Applications
Kate C.-J. Lin (林靖茹)
Network & Mobile System Lab Research Center for IT Innovation
Academia Sinica
Agenda
• Auto Rate Adaptation
• Orthogonal Frequency Division Modulation (OFDM)
• Multi-Input Multi-Output Systems (MIMO)
2
Auto Rate Adaptation
• Modulations and bit-rates
• SNR and bit-error rate
• Bit-rate selection algorithms
Modulations
I Q
BPSK
1 à 1+0i 0 à -‐1+0i
Constella)on Points
Modulate digital bits
to a complex number (sample)
Modulations
64QAM
I Q
I Q
BPSK
1 0
QPSK
I Q
01 11
10 00
16QAM
I Q
1000 1100
1001 1101
1011 1111
1010 1110
0100 0000
0101 0001
0111 0011
0110 0010
Demodulation
Map the received complex number back to digital bits
I Q
BPSK Received sample
Closet constella?on point
If Tx is actually sending ‘0’, bit error occurs
Bit-Rates in 802.11
Bit- 802.11 DSSS Modulation Bits Coding Mega-
rate Stan- or per Rate Symbols
dards OFDM Symbol per
second
1 b DSSS BPSK 1 1/11 11
2 b DSSS QPSK 2 1/11 11
5.5 b DSSS CCK 1 4/8 11
11 b DSSS CCK 2 4/8 11
6 a/g OFDM BPSK 1 1/2 12
9 a/g OFDM BPSK 1 3/4 12
12 a/g OFDM QPSK 2 1/2 12
18 a/g OFDM QPSK 2 3/4 12
24 a/g OFDM QAM-16 4 1/2 12
36 a/g OFDM QAM-16 4 3/4 12
48 a/g OFDM QAM-64 6 2/3 12
54 a/g OFDM QAM-64 6 3/4 12
Figure 2-1: A summary of the 802.11 bit-rates. Each bit-rate uses a specific combination of modulation and channel coding. OFDM bit-rates send 48 symbols in parallel.
a channel. In the presence of fading, multi-path interference, or other interference that is not additive white Gaussian noise, predicting the combinations of modulation and channel coding that will be most effective at masking bit errors is difficult.
All 802.11 packets contain a small preamble before the data payload which is sent at a low bit-rate. The preamble contains the length of the packet, the bit-rate for the data payload, and some parity information calculated over the contents of the preamble. The preamble is sent at 1 megabit in 802.11b and 6 megabits in 802.11g and 802.11a. This results in the unicast packet overhead being different for 802.11b and 802.11g bit-rates; a perfect link can send approximately 710 1500-byte unicast packets per second at 12 megabits (an 802.11g bit-rate) and 535 packets per second at 1 megabit (an 802.11b bit-rate). This means that 12 megabits can sustain nearly 20% loss before a lossless 11 megabits provides better throughput, even though the bit-rate is less than 10% different.
2.2 Medium-Access Control (MAC) Layer
For the purposes of this thesis, the most important properties of the 802.11 MAC layer are the medium access mechanisms and the unicast retry policy.
To prevent nodes from sending at the same time, 802.11 uses carrier sense multiple access
14
Coding Rate
• Avoid random errors
1/2: Add 1x redundant bits
3/4: Add 1/3x redundant bits
• Haven’t solved the problem yet
Data input: 1, 1, 0, 1, 0, 1, 1, 0, …
After encoding:
1, 1, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 0, 0, ….
Still one bit error à Suffer from bursty errors
Interleave and De-interleave
So ur ce c od in g Inter leav e Mo dul atio n D/A
channel
noise
+
1, 1, 0, 1, 0 1, 1, 1, 1, 0,
0, 1, 1, 0, 0 1, 0, 0, 1, 0,
1, 1, 0, 1, 1 1, -‐1, -‐1, 1, -‐1, 1, 1, -‐1, 1, 1
D ec od in g De-inter leav e De-mo dulatio n A/ D
1, 0, 1, 1, 0,
0, 1, 1, 1, 0 1, 0, 1, 0, 0,
1, 1, 0, 1, 1 1, -‐1, 1, -‐1, -‐1, 1, 1, -‐1, 1, 1 1, 1, 0, 1, 0
Transmitter
Receiver Create a more uniform
distribu?on of errors
Channel Quality vs. Bit-Rate
• When channels are very good
Encode more bits as a sample
• When channels are noisy
Encode fewer bits as a sample
Why is it affected by the channel quality?
Error Probability vs. Modulations
I Q
BPSK
‖ noise‖
SNR = 10log10 (‖signal‖/‖noise‖)
‖ signal‖
Decode correctly
QPSK
I Q
01 11
10 00
‖ noise‖
Decode incorrectly
Given the same SNR
Given the same SNR, decodable for BPSK,
but un-decodable for QPSK
SNR vs. BER (Bit Error Rate)
1e-05 0.0001 0.001 0.01 0.1 1
0 5 10 15 20 25 30 35
Bit Error Rate
S/N (dB)
BPSK (1 megabit/s) QPSK (2 megabits/s) QAM-16 (4 megabits/s) QAM-64 (6 megabits/s)
Figure 1-2: Theoretical bit error rate (BER) versus signal-to-noise ratio for several modu- lation schemes assuming AGWN. The y axis is a log scale. Higher bit-rates require larger S/N to achieve the same bit-rate as lower bit-rates.
9
802.11 operating region 5dB
SNR vs. PER (Packet Error Rate)
• In 802.11, a packet is received correctly if it passes the CRC check ( all bits are correct )
Receive all or none
0 1 2 3 4 5 6
5 10 15 20 25 30
Throughput (Megabits per Second)
S/N (dB) BPSK (1 megabit/s) QPSK (2 megabit/s) QAM-16 (4 megabits/s) QAM-64 (6 megabits/s)
Figure 1-4: Theoretical throughput in megabits per second using packets versus signal-to- noise ratio for several modulations, assuming AGWN and a symbol rate of 1 mega-symbol per second.
packets can be estimated using the following equation:
throughput = (1− BER)n∗ bitrate
This equation assumes the transmitter sends packets back-to-back, the receiver knows the location of each packet boundary, the receiver can determine the integrity of the data with no overhead, there is no error correction, and the symbol rate is 1 mega-symbol per second.
Packets change the throughput versus S/N graph dramatically; Figure 1-4 shows through- put in megabits per second versus S/N for 1500-byte packets after accounting for packet losses caused by bit-errors. The range where each modulation delivers non-zero throughput but suffers from loss is much smaller in Figure 1-4 than in Figure 1-3. For most S/N values in the range from 5 to 30 dB, the best bit-rate delivers packets without loss.
Bit-rate selection is easier for links that behave as in Figure 1-4 than as in Figure 1-3;
the sender can start on the highest bit-rate and switch to another bit-rate whenever the
11
PER = 1-(1-BER)
nThroughput
= (1-BER)
n* bit-rate
Throughput degrades
quickly even with a
little BER
Bit-Rate Selection
• Given the SNR, select the bit-rate that can achieve the highest throughput
0 1 2 3 4 5 6
5 10 15 20 25 30
Throughput (Megabits per Second)
S/N (dB) BPSK (1 megabit/s) QPSK (2 megabit/s) QAM-16 (4 megabits/s) QAM-64 (6 megabits/s)
Figure 1-4: Theoretical throughput in megabits per second using packets versus signal-to- noise ratio for several modulations, assuming AGWN and a symbol rate of 1 mega-symbol per second.
packets can be estimated using the following equation:
throughput = (1− BER)n∗ bitrate
This equation assumes the transmitter sends packets back-to-back, the receiver knows the location of each packet boundary, the receiver can determine the integrity of the data with no overhead, there is no error correction, and the symbol rate is 1 mega-symbol per second.
Packets change the throughput versus S/N graph dramatically; Figure 1-4 shows through- put in megabits per second versus S/N for 1500-byte packets after accounting for packet losses caused by bit-errors. The range where each modulation delivers non-zero throughput but suffers from loss is much smaller in Figure 1-4 than in Figure 1-3. For most S/N values in the range from 5 to 30 dB, the best bit-rate delivers packets without loss.
Bit-rate selection is easier for links that behave as in Figure 1-4 than as in Figure 1-3;
the sender can start on the highest bit-rate and switch to another bit-rate whenever the
11
QPSK
64QAM
Difficulties with Rate Adaptation
• Channel quality changes very quickly
Especially when the device is moving
• Can’t tell the difference between
poor channel quality due to noise/interference/
collision (high ‖ noise‖)
poor channel quality due to distance (low
‖ signal‖)
Ideally, we want to decrease the rate due to low
signal strength, but not interference/collision
Types of Auto-Rate Adaptation
Transmitter-based Receiver-Based
SNR-based RBAR, OAR
ACK-based ARF, AARF Throughput-based SampleRate
(default in Linux)
RRAA Selected by Tx
(Less accurate)
Selected by Rx
(Higher overhead)
Sync. ACK vs. Async ACK
• Sync. ACK
Cost the minimum overhead
Only know whether the packet is transmitted correctly
Don’t know whether the packet error is due to incorrect rate selection or collision
• Async. ACK
Cost extra overhead
Can include more detailed information Tx
Rx
backoff Data
ACK SIFS
backoff A-‐ACK DIFS
Robust Rate Adaption Algorithm (RRAA)
• Dynamically enable RTS/CTS before data transmission
• Detect that the low throughput is due to the incorrect bit-rate selection or
collision (hidden terminals)
• Estimate the correct number of transmissions to keep RTS/CTS
• Disable RTS/CTS if it does not help
S. Wong, H. Yang, S. Lu, V. Bharghavan, “Robust Rate Adapta?on for
802.11 Wireless Networks,” ACM MOBICOM, 2006
SampleRate
• Periodically send packets at bit-rates other than the current bit-rate
• Calculate the transmission time of each packet
packet length, bit-rate, number of retries, backoff time
pkt1 pkt1’ pkt1’’ pkt2 … pkt10 r*
retry 1
pkt r’
pkt
retry 2 retry 1
• Look up the destination and add the transmission time to the total transmission times for the bit-rate.
• If the packet succeeded, increment the number of successful packets sent at that bit- rate.
• If the packet failed, increment the number of successive failures for the bit-rate. Oth- erwise reset it.
• Re-calculate the average transmission time for the bit-rate based on the sum of trans- mission times and the number of successful packets sent at that bit-rate.
• Set the current-bit rate for the destination to the one with the minimum average transmission time.
• Append the current time, packet status, transmission time, and bit-rate to the list of transmission results.
SampleRate’s remove stale results() function removes results from the transmission results queue that were obtained longer than ten seconds ago. For each stale transmission result, it does the following:
• Remove the transmission time from the total transmission times at that bit-rate to that destination.
• If the packet succeeded, decrement the number of successful packets at that bit-rate to that destination.
After remove stale results() performs these operations for each stale sample, it re-
calculates the minimum average transmission times for each bit-rate and destination. remove stale results() then sets the current bit-rate for each destination to the one with the smallest average trans-
mission time.
To calculate the transmission time of a n-byte unicast packet given the bit-rate b and number of retries r, SampleRate uses the following equation based on the 802.11 unicast retransmission mechanism detailed in Section 2.2:
tx time(b, r, n) = dif s + backof f (r) + (r + 1) ∗ (sifs + ack + header + (n ∗ 8/b) (5.1)
37
Sample Rates
• Select the rate that has the smallest predicted average packet transmission time
• Do not sample the rates that
have failed four successive times
are unlikely to be better than the current one
• Is thought of the most efficient scheme for static environments
J. Bicket, “Bit-‐rate Selec?on in Wireless Networks,” Ph.D Thesis, MIT, 2005
Rate Adaptation for Multicast?
• Can only assign a single rate to each packet
• Possible Solutions
For reliable transmission: select the rate based on the worst node
For non-reliable transmission: provide clients
heterogeneous throughput
Recent Proposals
• ZipTx
K. Lin, N. Kushman and D. Katabi, “Harnessing Partial Packets in 802.11 Networks,” ACM MOBICOM, 2008
Exploit partial packets with consideration of bit-rate adaptation
• SoftRate
M. Vutukuru, H. Balakrishnan and K. Jamieson, “Cross-Layer Wireless Bit Rate Adaptation,” ACM SIGCOMM, 2009
Exploit soft information to improve selection accuracy
• FARA
H. Rahul, F. Edalat, D. Katabi and C. Sodini, “Frequency-Aware Rate Adaptation and MAC Protocols,” ACM MOBICOM, 2009
Adapt the bit-rate for every OFDM subcarrier
• ESNR
D. Halperin, W. Hu, A. Sheth and D. Wetherall, “Predictable 802.11 Packet Delivery from Wireless Channel Measurements”, ACM SIGCOMM, 2010
Consider frequency selective fading
Frequency-Aware Rate Adaptation (FARA)
H. Rahul, F. Edalat, D. Katabi, C. Sodini
MOBICOM 2009
Frequency Diversity
• Frequency diverse across 100MHz of 802.11a spectrum
• The SNRs of different frequencies can be as much as 20dB on a single link
• Different receivers could prefer different frequencies
Frequency-Aware Rate Adaptation and MAC Protocols
Hariharan Rahul
†, Farinaz Edalat
ℵ, Dina Katabi
†, and Charles Sodini
††
Massachusetts Institute of Technology
ℵRKF Engineering Solutions, LLC
ABSTRACT
There has been burgeoning interest in wireless technologies that can use wider frequency spectrum. Technology advances, such as 802.11n and ultra-wideband (UWB), are pushing toward wider fre- quency bands. The analog-to-digital TV transition has made 100- 250 MHz of digital whitespace bandwidth available for unlicensed access. Also, recent work on WiFi networks has advocated discard- ing the notion of channelization and allowing all nodes to access the wide 802.11 spectrum in order to improve load balancing. This shift towards wider bands presents an opportunity to exploit frequency diversity. Specifically, frequencies that are far from each other in the spectrum have significantly different SNRs, and good frequencies differ across sender-receiver pairs.
This paper presents FARA, a combined frequency-aware rate adaptation and MAC protocol. FARA makes three departures from conventional wireless network design: First, it presents a scheme to robustly compute per-frequency SNRs using normal data trans- missions. Second, instead of using one bit rate per link, it en- ables a sender to adapt the bitrate independently across frequencies based on these per-frequency SNRs. Third, in contrast to traditional frequency-oblivious MAC protocols, it introduces a MAC protocol that allocates to a sender-receiver pair the frequencies that work best for that pair. We have implemented FARA in FPGA on a wide- band 802.11-compatible radio platform. Our experiments reveal that FARA provides a 3.1× throughput improvement in comparison to frequency-oblivious systems that occupy the same spectrum.
Categories and Subject Descriptors
C.2.2 [Computer Sys- tems Organization]: Computer-Communications NetworksGeneral Terms
Algorithms, Design, PerformanceKeywords
Wireless, Cognitive Radios, Wideband, Rate Adapta- tion, Cross-layer1 I
NTRODUCTIONWireless technologies are pushing toward wider frequency bands than the 20 MHz channels employed by existing 802.11 networks.
802.11n already includes a 40 MHz mode that bonds together two 20 MHz bands [23]. Emerging ultra-wideband (UWB) technolo- gies employ hundreds of MHz to support multimedia homes and offices [24, 50, 9, 40]. The FCC has recently permitted unlicensed
Permission to make digital or hard copies of all or part of this work for per- sonal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to re- publish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
MobiCom’09, September 20–25, 2009, Beijing, China.
Copyright 2009 ACM 978-1-60558-702-8/09/09 . . . $10.00.
0 5 10 15 20 25 30
-40 -20 0 20 40
SNR (dB)
Freq (Mhz)
Figure 1: Frequency diversity across 100 MHz of 802.11a spec- trum as observed by two receivers for transmissions from the same sender. The figure shows that the SNRs of different frequen- cies can differ by as much as 20 dB on a single link. Further, different receivers prefer different frequencies.
use of digital TV whitespaces that occupy 100-250 MHz of spectrum vacated by television bands in the analog-to-digital transition [12].
Furthermore, recent empirical studies show that the 802.11 chan- nelization model which limits each node to a single 20 MHz chan- nel can lead to severe load imbalance [19, 28, 37]. They advocate discarding channelization and allowing all nodes to access the en- tire 802.11 spectrum based on demand [19, 37]. This push towards wider bands is further enabled by the constantly lowering prices of high-speed ADC and DAC hardware [38, 31].1 In particular, today, wireless cards that span over 100 MHz of spectrum can be built us- ing off-the-shelf hardware components [35].
As wireless networks push towards wider bands, we can no longer afford to ignore frequency diversity. Specifically, multipath effects cause frequencies that are far away from each other in the spectrum to experience independent fading. Thus, different frequencies can exhibit very different SNRs for a single sender-receiver pair. Further, the frequencies that show good performance for one sender-receiver pair may be very different than the frequencies that show good per- formance for another pair. Fig. 1 shows empirical measurements of the SNRs across 100 MHz of the 802.11a spectrum, as observed by 2 clients for transmissions from the same AP (see §9 for exper- imental setup). The figure reveals that different frequencies show a difference in SNR of over 20 dB both for a single link and across links. Existing bitrate adaptation and MAC protocols however are frequency-oblivious. They assign the same bitrate to all frequencies and allocate the medium in a time-based manner, ignoring the fact that different frequencies work better for different sender-receiver pairs. Thus, current rate adaptation and MAC protocols can neither deal with the challenge nor exploit the opportunities introduced by the frequency diversity of wide bands or unchannelized 802.11.
1The wider the band, the faster the ADC and DAC have to sample the signal.
FARA
• Instead of assigning the same rate to the entire frequency band, it allows each
OFDM sub-carrier to pick a modulation and a code rate that match its SNR
Frequency-Aware Rate Adaptation and MAC Protocols
Hariharan Rahul
†, Farinaz Edalat
ℵ, Dina Katabi
†, and Charles Sodini
††
Massachusetts Institute of Technology
ℵRKF Engineering Solutions, LLC
ABSTRACT
There has been burgeoning interest in wireless technologies that can use wider frequency spectrum. Technology advances, such as 802.11n and ultra-wideband (UWB), are pushing toward wider fre- quency bands. The analog-to-digital TV transition has made 100- 250 MHz of digital whitespace bandwidth available for unlicensed access. Also, recent work on WiFi networks has advocated discard- ing the notion of channelization and allowing all nodes to access the wide 802.11 spectrum in order to improve load balancing. This shift towards wider bands presents an opportunity to exploit frequency diversity. Specifically, frequencies that are far from each other in the spectrum have significantly different SNRs, and good frequencies differ across sender-receiver pairs.
This paper presents FARA, a combined frequency-aware rate adaptation and MAC protocol. FARA makes three departures from conventional wireless network design: First, it presents a scheme to robustly compute per-frequency SNRs using normal data trans- missions. Second, instead of using one bit rate per link, it en- ables a sender to adapt the bitrate independently across frequencies based on these per-frequency SNRs. Third, in contrast to traditional frequency-oblivious MAC protocols, it introduces a MAC protocol that allocates to a sender-receiver pair the frequencies that work best for that pair. We have implemented FARA in FPGA on a wide- band 802.11-compatible radio platform. Our experiments reveal that FARA provides a 3.1× throughput improvement in comparison to frequency-oblivious systems that occupy the same spectrum.
Categories and Subject Descriptors
C.2.2 [Computer Sys- tems Organization]: Computer-Communications NetworksGeneral Terms
Algorithms, Design, PerformanceKeywords
Wireless, Cognitive Radios, Wideband, Rate Adapta- tion, Cross-layer1 I
NTRODUCTIONWireless technologies are pushing toward wider frequency bands than the 20 MHz channels employed by existing 802.11 networks.
802.11n already includes a 40 MHz mode that bonds together two 20 MHz bands [23]. Emerging ultra-wideband (UWB) technolo- gies employ hundreds of MHz to support multimedia homes and offices [24, 50, 9, 40]. The FCC has recently permitted unlicensed
Permission to make digital or hard copies of all or part of this work for per- sonal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to re- publish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
MobiCom’09, September 20–25, 2009, Beijing, China.
Copyright 2009 ACM 978-1-60558-702-8/09/09 . . . $10.00.
0 5 10 15 20 25 30
-40 -20 0 20 40
SNR (dB)
Freq (Mhz)
Figure 1: Frequency diversity across 100 MHz of 802.11a spec- trum as observed by two receivers for transmissions from the same sender. The figure shows that the SNRs of different frequen- cies can differ by as much as 20 dB on a single link. Further, different receivers prefer different frequencies.
use of digital TV whitespaces that occupy 100-250 MHz of spectrum vacated by television bands in the analog-to-digital transition [12].
Furthermore, recent empirical studies show that the 802.11 chan- nelization model which limits each node to a single 20 MHz chan- nel can lead to severe load imbalance [19, 28, 37]. They advocate discarding channelization and allowing all nodes to access the en- tire 802.11 spectrum based on demand [19, 37]. This push towards wider bands is further enabled by the constantly lowering prices of high-speed ADC and DAC hardware [38, 31].1 In particular, today, wireless cards that span over 100 MHz of spectrum can be built us- ing off-the-shelf hardware components [35].
As wireless networks push towards wider bands, we can no longer afford to ignore frequency diversity. Specifically, multipath effects cause frequencies that are far away from each other in the spectrum to experience independent fading. Thus, different frequencies can exhibit very different SNRs for a single sender-receiver pair. Further, the frequencies that show good performance for one sender-receiver pair may be very different than the frequencies that show good per- formance for another pair. Fig. 1 shows empirical measurements of the SNRs across 100 MHz of the 802.11a spectrum, as observed by 2 clients for transmissions from the same AP (see §9 for exper- imental setup). The figure reveals that different frequencies show a difference in SNR of over 20 dB both for a single link and across links. Existing bitrate adaptation and MAC protocols however are frequency-oblivious. They assign the same bitrate to all frequencies and allocate the medium in a time-based manner, ignoring the fact that different frequencies work better for different sender-receiver pairs. Thus, current rate adaptation and MAC protocols can neither deal with the challenge nor exploit the opportunities introduced by the frequency diversity of wide bands or unchannelized 802.11.
1The wider the band, the faster the ADC and DAC have to sample the signal.
54Mb/s
6Mb/s
FARA
• Receiver driver protocol
Initially, the sender transmit few symbols using the lowest bit-rate for all sub-carriers
The receiver selects the bit-rate based on an SNR-Rate mapping table
where H
iis the channel, x
i[k ] is the k
thtransmitted signal sam- ple in subband i , and n
i[k ] is the corresponding noise sample. The receiver knows H
ifor all subbands because it is estimated using known OFDM symbols in the preamble [20]. In the case of a pi- lot subband, x
i[k ] is also known at the receiver since pilot subbands contain a known data sequence. As a result, the receiver can estimate the noise samples, n
i[k ], and the noise power, N
0, as:
n
i[k ] = y
i[k ] − H
ix
i[k ] (4) N
0= E
i ,k(n
i[k ]
2) (5) where the function E (.) is the mean computed using all pilot bits across all symbols in the data packet.
Thus, every received packet allows the receiver to obtain a new SNR measurement for each OFDM subband. The receiver maintains a time weighted moving average of the SNR in each subband, which it updates on the reception of a data packet.
A few points are worth noting:
(a) What happens when the data packet is corrupted (i.e. does not pass the checksum test)? Even when the packet is corrupted, the receiver can still compute an accurate estimate of the per-subband SNRs. This is because the receiver can compute the average received power, regardless of whether the packet is corrupted or not. Further- more, the receiver can still obtain an accurate estimate of the noise power since this only requires the pilots which are known, and sent at BPSK, which is the most robust modulation rate and hence al- low synchronization and packet recovery even at low SNRs. Thus, FARA can get accurate estimates of the per-subband SNRs from ev- ery captured packet, including corrupted packets.
(b) How accurate are FARA’s SNR estimates? We note that since FARA has access to the PHY layer, it can collect accurate SNR estimates. In particular, traditional estimates of the SNR use RSSI readings, which measure the received power of a few samples at the beginning of the packet (i.e., the AGC gain) [6], or infer the SNR using just the correlation of header symbols in the preamble of the packet [49]. In contrast, FARA exploits the known pilot bits to ac- curately estimate the noise power and utilize it in its SNR compu- tation. Furthermore, FARA computes its signal and noise estimates over the whole packet and not just a few samples at the beginning of the packet, which allows it to obtain more stable estimates.
(c) Do different choices of bitrate affect the accuracy of FARA’s SNR estimation? OFDM data subbands use a different modulation scheme depending on the choice of bitrate. The modulation scheme in a subband, however, does not affect our per-subband SNR esti- mate. The estimation of SNR involves only the measured power in each subband and hence can be performed on any packet indepen- dent of the modulation and coding schemes used by the transmitter.
6 F REQUENCY -A WARE R ATE A DAPTATION
The goal of rate adaptation is to determine the highest bitrate that a channel can sustain at any point in time. Traditional 802.11 rate adaptation schemes are frequency-oblivious, and use the same modulation scheme and coding rate across all frequencies. Thus, they cannot exploit the frequency diversity present across the 802.11 spectrum. In contrast, FARA exploits this frequency diversity via a frequency-aware rate adaptation scheme that picks different bitrates for different frequencies depending on their SNRs.
6.1 PHY Architecture
In 802.11, a particular bit rate implies a single modulation scheme and code rate over all OFDM subbands in the entire packet. For
!"#$$#$#"!
!"#$$#$#"!
%&''()* ) +,*&-
.(/0 !""#
$"%&'#()*+,-../('0-1/
2%3
%&''()* $
""#
$"%&'4/5/-6/('0-1/
1,*0"2"
3)/04.0(50
607,*&- .(/0
608,*0"2"
609)/04.0(50
%&''()* )
%&''()* $
%23
(a) Schematic of 802.11 PHY
!"#$$#$#"!
!"#$$#$#"!
!""#
$"%&'#()*+,-../('0-1/
2%3 ""#
$"%&'4/5/-6/('0-1/
&7189)./ %23
371/':'!*./(9/)6/ %/571/':'%/-*./(9/)6/
%/,7189).-/
(b) Schematic of FARA-enabled 802.11 PHY
Figure 3: OFDM PHY semantics with and without FARA. In FARA-enabled devices, the choice of modulation and FEC code rate is done independently for each OFDM subband.
Minimum Required SNR Modulation Coding
<3.5 dB Suppress subband
3.5 dB BPSK 1/2
5.0 dB BPSK 3/4
5.5 dB 4-QAM 1/2
8.5 dB 4-QAM 3/4
12.0 dB 16-QAM 1/2
15.5 dB 16-QAM 3/4
20.0 dB 64-QAM 2/3
21.0 dB 64-QAM 3/4
Table 1: Minimum required SNR for a particular modulation and code rate (i.e., bitrate). Table is generated offline using the WiGLAN radio platform by running all possible bit rates for the whole operational SNR range. The SNR field refers to the minimum SNR required to maintain the packet loss rate below 1% (see §9 for experimental setup).
example, a bitrate of 24 Mbps corresponds to 16-QAM modula- tion scheme and a half-rate code. 802.11 has 4 possible modulation schemes (BPSK, 4-QAM, 16-QAM, and 64-QAM), and 3 possible code rates (1/2, 2/3, and 3/4). In current 802.11, a transmitter imple- ments a particular bitrate by first taking the input bit stream, passing it to the convolutional coder, and puncturing to achieve the desired coding rate. The bits are then interleaved, modulated and striped over the OFDM subbands, as shown in Fig. 3(a). The process is reversed on the receiver as shown in the figure.
FARA makes a few modifications to the existing 802.11 PHY layer, as shown in Fig. 3(b). Specifically, FARA employs the same set of modulation schemes and code rates supported by the existing 802.11. However, it allows each OFDM subband to pick a modu- lation scheme and a code rate that match its SNR, independently from the other subbands. Note that this design does not require addi- tional modulation/demodulation or coding/decoding modules in the PHY layer. In particular, since we use standard 802.11 modulation and coding options, we only need to buffer the samples and process them through the same pipeline.
6.2 Mapping Subband SNRs to Optimal Bitrates
The receiver needs to map the average SNR in each subband to
the optimal bitrate for that band. To do so, the receiver uses an SNR
characterization table like the one in Table 1 that lists the minimum
SNR required for a particular combination of modulation and cod-
Predictable 802.11 Packet Delivery from Wireless Channel Measurements
D. Halperin, W. Hu, A. Sheth and D. Wetherall
ACM SIGCOMM, 2010
SNR-based Rate Adaptation
• SNR-based rate adaptation is usually inaccurate because we
Assume frequency-flat fading
Select the bit-rate based on “average SNR”
across bins
• However, this will over-estimate the channel quality because