MegaMIMO: Scaling Wireless Throughput with the Number of Users - PowerPoint PPT Presentation

MegaMIMO: Scaling Wireless Throughput with the Number of Users Hariharan Rahul , Swarun Kumar and Dina Katabi

There is a Looming Wireless Capacity Crunch Given the trends in the growth of wireless demand, and based on current technology, the FCC projects that the US will face a spectrum shortfall in 2013. The iPhone 4 demo failed at Steve Jobs’s keynote due to wireless congestion. Jobs’s reaction: “If you want to see the demos, shut off your laptops, turn off all these MiFi base stations, and put them on the floor, please.”

MegaMIMO MegaMIMO alleviates the capacity crunch by transmitting more bits per unit of spectrum.

T oday’s Wireless Networks Ethernet Access Access Access Point 1 Point 2 Point 3 Interference! User 3 User 1 User 2 Access Points Can’t Transmit T ogether in the Same Channel

MegaMIMO Ethernet Access Access Access Point 1 Point 2 Point 3 User 3 User 1 User 2 Interference from Interference from Interference from x 1 +x 2 ≈0 x 2 +x 3 ≈0 x 1 +x 3 ≈0 Data : x 3 survives Data : x 1 survives Data : x 2 survives All Access Points Can Transmit Simultaneously in the Same Channel

MegaMIMO Ethernet Access Access Access Point 1 Point 2 Point 3 User 3 User 1 User 2 Interference from Interference from Interference from x 1 +x 2 ≈0 Enables senders to transmit together x 2 +x 3 ≈0 x 1 +x 3 ≈0 Data : x 3 survives Data : x 1 survives without interference Data : x 2 survives All Access Points Can Transmit Simultaneously in the Same Channel

MegaMIMO = Distributed MIMO Distributed protocol for APs to act as a huge MIMO transmitter with sum of antennas Ethernet … AP3 AP10 AP2 AP1 … User 3 User 10 User 1 User 2 10 APs  10x higher throughput

Diving Into The Details

Transmitting Without Interference AP 1 AP 2 Wants x 2 Wants x 1 Receives y 1 Receives y 2 Cli 1 Cli 2 y 1 = d 1 x 1 + 0 . x 2 y 2 = 0 . x 1 + d 2 x 2 y 1 d 1 x 1 0 = y 2 0 d 2 x 2

Transmitting Without Interference AP 1 AP 2 Wants x 2 Wants x 1 Receives y 1 Receives y 2 Cli 1 Cli 2 y 1 = d 1 x 1 + 0 . x 2 y 2 = 0 . x 1 + d 2 x 2 y 1 x 1 D = Diagonal y 2 x 2

Transmitting Without Interference AP 1 AP 2 Wants x 2 Wants x 1 Receives y 1 Receives y 2 Cli 1 Cli 2 Diagonal Matrix  Non-Interference y 1 = d 1 x 1 + 0 . x 2 y 2 = 0 . x 1 + d 2 x 2 y 1 x 1 Goal: Make the effective channel matrix D = diagonal Diagonal y 2 x 2

On-Chip MIMO • All nodes are synchronized in time to within nanoseconds of each other. • Oscillators at all nodes have exactly the same frequency, i.e., no frequency offset.

On-Chip MIMO Sends x 1 Sends x 2 AP h 11 h 22 h 12 h 21 Cli 1 Cli 2 y 2 y 1 y 1 = h 11 x 1 + h 12 x 2 y 2 = h 21 x 1 + h 22 x 2 y 1 h 11 x 1 h 12 = Non-diagonal h 21 y 2 h 22 x 2 Matrix  Interference

On-Chip MIMO Sends x 1 Sends x 2 AP h 11 h 22 h 12 h 21 Cli 1 Cli 2 y 2 y 1 y 1 = h 11 x 1 + h 12 x 2 y 2 = h 21 x 1 + h 22 x 2 y 1 h 11 x 1 h 12 = h 21 y 2 h 22 x 2

On-Chip MIMO Sends s 1 Sends s 2 AP h 11 h 22 h 12 h 21 Cli 1 Cli 2 y 2 y 1 y 1 = h 11 s 1 + h 12 s 2 y 2 = h 21 s 1 + h 22 s 2 y 1 h 11 s 1 h 12 = h 21 y 2 h 22 s 2

On-Chip MIMO Sends s 1 Sends s 2 AP h 11 h 22 h 12 h 21 Cli 1 Cli 2 y 2 y 1 y 1 = h 11 s 1 + h 12 s 2 y 2 = h 21 s 1 + h 22 s 2 y 1 s 1 H = y 2 s 2

Making Effective Channel Matrix Diagonal Sends s 1 Sends s 2 AP h 11 h 22 h 12 h 21 Cli 1 Cli 2 y 2 y 1 y 1 = h 11 s 1 + h 12 s 2 y 2 = h 21 s 1 + h 22 s 2 y 1 s 1 H = y 2 s 2

Making Effective Channel Matrix Diagonal Sends s 1 Sends s 2 AP h 11 h 22 h 12 h 21 Cli 1 Cli 2 y 2 y 1 x 1 y 1 = h 11 s 1 + h 12 s 2 H -1 y 2 = h 21 s 1 + h 22 s 2 x 2 y 1 s 1 H = y 2 s 2

Making Effective Channel Matrix Diagonal Sends s 1 Sends s 2 AP h 11 h 22 h 12 h 21 Cli 1 Cli 2 y 2 y 1 y 1 = h 11 s 1 + h 12 s 2 y 2 = h 21 s 1 + h 22 s 2 Effective channel is y 1 x 1 H H -1 = diagonal y 2 x 2

Beamforming System Description Channel Measurement: • AP1 and AP2 measure channels to clients • Clients report measured channels back to APs Data Transmission: • Packets are forwarded to both APs • Each AP computes its beamformed signal s i using the equation s = H -1 x • Clients 1 and 2 decode x 1 and x 2 independently

Distributed Transmitters • Nodes are not synchronized in time. – We use SourceSync to synchronize senders within 10s of ns (SIGCOMM 2010) – Works for OFDM based systems like Wi-Fi, LTE etc. • Oscillators are not synchronized and have frequency offsets relative to each other.

MegaMIMO • First wireless network that can scale network throughput with the number of transmitters • Algorithm for phase synchronization across multiple independent transmitters • Demonstrated in a wireless testbed implementation

What Happens with Independent Oscillators? AP 1 AP 2 h 11 h 22 h 12 h 21 Cli 1 Cli 2 h 11 h 12 h 21 h 22

What Happens with Independent Oscillators? ω T1 AP 1 AP 2 h 11 h 22 h 12 h 21 Cli 1 Cli 2 ω R1 e j( ω - ω )t h 11 h 12 T1 R1 h 21 h 22

What Happens with Independent Oscillators? ω T1 AP 1 AP 2 h 11 h 22 h 12 h 21 Cli 1 Cli 2 ω R1 e j( ω - ω )t h 11 h 12 T1 R1 h 21 h 22

What Happens with Independent Oscillators? ω T1 AP 1 ω T2 AP 2 h 11 h 22 h 12 h 21 Cli 1 Cli 2 ω R1 e j( ω - ω )t e j( ω - ω )t h 11 h 12 T1 R1 T2 R1 h 21 h 22

What Happens with Independent Oscillators? ω T1 AP 1 ω T2 AP 2 h 11 h 22 h 12 h 21 Cli 1 Cli 2 ω R1 e j( ω - ω )t e j( ω - ω )t h 11 h 12 T1 R1 T2 R1 h 21 h 22

What Happens with Independent Oscillators? ω T1 AP 1 ω T2 AP 2 h 11 h 22 h 12 h 21 ω R2 Cli 1 Cli 2 ω R1 e j( ω - ω )t e j( ω - ω )t h 11 h 12 T1 R1 T2 R1 e j( ω - ω )t e j( ω - ω )t h 21 h 22 T1 R2 R2 T2

What Happens with Independent Oscillators? ω T1 AP 1 ω T2 AP 2 h 11 h 22 h 12 h 21 ω R2 Cli 1 Cli 2 ω R1 e j( ω - ω )t e j( ω - ω )t h 11 h 12 T1 R1 T2 R1 Time Varying e j( ω - ω )t e j( ω - ω )t h 21 h 22 T1 R2 R2 T2

Channel is Time Varying ω T1 AP 1 ω T2 AP 2 h 11 h 22 h 12 h 21 ω R2 Cli 1 Cli 2 ω R1 H(t)

Does Traditional Beamforming Still Work? ω T1 AP 1 ω T2 AP 2 h 11 h 22 h 12 h 21 ω R2 Cli 1 Cli 2 ω R1 y 1 (t) s 1 (t) H(t) y 2 (t) = s 2 (t)

Does Traditional Beamforming Still Work? ω T1 AP 1 ω T2 AP 2 h 11 h 22 h 12 h 21 ω R2 Cli 1 Cli 2 ω R1 x 1 (t) y 1 (t) Not H -1 H(t) y 2 (t) = Diagonal x 2 (t)

Does Traditional Beamforming Still Work? ω T1 AP 1 ω T2 AP 2 h 11 h 22 h 12 h 21 ω R2 Cli 1 Cli 2 ω R1 Beamforming does not work x 1 (t) y 1 (t) Not H -1 H(t) y 2 (t) = Diagonal x 2 (t)

Challenge Channel is Rapidly Time Varying Relative Channel Phases of Transmitted Signals Changes Rapidly With Time Prevents Beamforming

Distributed Phase Synchronization High Level Intuition: • Pick one AP as the lead • All other APs are slaves – Imitate the behavior of the lead AP by fixing the rotation of their oscillator relative to the lead.

Decomposing H(t) e j( ω - ω )t e j( ω - ω )t h 11 h 12 T1 R1 R1 T2 e j( ω - ω )t e j( ω - ω )t h 22 h 21 T1 R2 R2 T2 e -j ω t e j( ω )t e j( ω )t h 11 h 12 0 R1 T1 T2 e -j ω t e j( ω )t e j( ω )t h 22 0 h 21 T1 T2 R2

Decomposing H(t) e -j ω t e j( ω )t e j( ω )t h 11 h 12 0 T1 R1 T2 e j( ω )t e j( ω )t e -j ω t 0 h 21 h 22 T1 R2 T2

Decomposing H(t) e -j ω t e j ω t h 11 h 12 0 0 T1 R1 e j ω t e -j ω t 0 h 21 h 22 0 R2 T2

Decomposing H(t) e -j ω t e j ω t h 11 h 12 0 0 T1 R1 e j ω t e -j ω t 0 h 21 h 22 0 R2 T2

Decomposing H(t) e j ω t e -j ω t 0 0 T1 R1 H e j ω t e -j ω t 0 0 T2 R2 Diagonal Devices cannot track their own oscillator phases…

Decomposing H(t) e j ω t e -j ω t 0 0 T1 R1 H e j ω t e -j ω t T1 T1 e j ω t e -j ω t 0 0 T2 R2

Decomposing H(t) e j( ω - ω )t 0 0 1 T1 R1 H e j( ω - ω )t 0 0 e j( ω - ω )t T1 R2 T2 T1 R(t) T(t) Depends only on transmitters

Decomposing H(t) e j( ω - ω )t 0 0 1 T1 R1 H e j( ω - ω )t 0 0 e j( ω - ω )t T1 R2 T2 T1 R(t) T(t) H(t) = R(t).H.T(t)

Beamforming with Different Oscillators y 1 (t) s 1 (t) R(t).H.T(t) H(t) y 2 (t) = s 2 (t) s 1 (t) x 1 (t) T(t) -1 H -1 s 2 (t) = x 2 (t)

Beamforming with Different Oscillators x 1 (t) y 1 (t) T(t) -1 H -1 y 2 (t) = R(t).H.T(t) H(t) x 2 (t) Diagonal s 1 (t) x 1 (t) T(t) -1 H -1 s 2 (t) = x 2 (t)

Transmitter Compensation 0 1 T(t) = 0 e j( ω - ω )t T2 T1

Transmitter Compensation 0 1 T(t) -1 = 0 e -j( ω - ω )t T2 T1 Slave AP imitates lead by multiplying each sample by oscillator rotation relative to lead Requires only local information  Fully distributed

Measuring Phase Offset • Multiply frequency offset by elapsed time • Requires very accurate estimation of frequency offset – Error of 25 Hz (10 parts per BILLION) changes complete alignment to complete misalignment in 20 ms. Need to keep resynchronizing to avoid error accumulation

Resynchronization lead h 2 AP 1 AP 2 Cli 1 Cli 2 lead e j( ω lead (t) = h 2 - ω )t h 2 T2 T1 Directly compute phase at each slave by measuring channel from lead