May 2008 doc IEEE 802 11 080535 r

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 PHY and MAC Throughput Analysis with 80 MHz for VHT below 6 GHz Date: 2008/05/09 Authors: Submission 1 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 Abstract This presentation shows the achievable PHY data rate and MAC throughput for VHT below 6 GHz with 80 MHz channel bandwidth with PHY and MAC optimizations Submission 2 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 Overview • • Previous MAC throughput results 80 MHz PHY configuration 802. 11 n MAC A-MPDU limitations New MAC throughput results Submission 3 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 Previous MAC Throughput Analysis • 11 -07 -2431 -00 -0 vht-analysis-on-ieee-802 -11 n-mac-efficiency. ppt • Using 800 ns GI: 540 Mbps x 2 = 1080 Mbps • PER ~ 0. 11 • No limit on the maximum size of A-MPDU • Number of MPDUs in one A-MPDU up to 64 Submission 4 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 80 MHz PHY Simulation • 80 MHz PHY simulation was developed based on 802. 11 n 40 MHz • To avoid partial overlap with 40 MHz TGn, we chose to use 80 MHz • New allocation of data and pilot tones to support 80 MHz – 20 MHz 802. 11 n: 52 data tones and 4 pilot tones – 40 MHz 802. 11 n: 108 data tones and 6 pilot tones – New 80 MHz: 224 data tones and 10 pilot tones • New MCSs for higher PHY data rates – Code rate 7/8, 64 QAM for 4 streams : 1. 307 Gbps – Code rate 3/4, 256 QAM for 4 streams : 1. 493 Gbps Submission 5 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 PHY Simulation Overview • For comparison purposes, the same TGn channel model was used in the 80 MHz simulator • Simulations were run with all the impairments listed in CC 67 (comparison criteria) with the exception of the PA – To avoid lengthy simulation time required by TGn PA model over sampling requirement, PA was not included in the simulations – Based on previous work in early days in TGn it is expected that additional 2 d. B backoff would be required for 256 QAM • Packet size was 3000 bytes • SNR of 40 d. B was selected for the MAC results to provide an optimistic operating point • 800 ns GI was used for the simulations, however 400 ns GI was used to compute the data rate to maximize throughput Submission 6 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 PHY Data Rates and PER at 40 d. B SNR (1) - Square configurations (2 x 2, 3 x 3, and 4 x 4) MCS Configuration Code Rate Modulation Data Rate (400 ns GI) 80 MHz PER at 40 d. B SNR – TGn channel model D PER at 40 d. B SNR - TGn channel model B 14 2 x 2 2 Stream 3/4 64 QAM 560 Mbps 0. 0 %* 1. 5 % 15 2 x 2 2 Stream 5/6 64 QAM 622 Mbps 0. 3 % 4. 6 % new 2 x 2 2 Stream 7/8 64 QAM 653 Mbps 1. 7 % 11. 9 % new 2 x 2 2 Stream 3/4 256 QAM 747 Mbps 3. 7 % 18. 4 % 22 3 x 3 3 Stream 3/4 64 QAM 840 Mbps 0. 0 % 13. 9 % 23 3 x 3 3 Stream 5/6 64 QAM 933 Mbps 1. 0 % 31. 6 % new 3 x 3 3 Stream 7/8 64 QAM 980 Mbps 4. 6 % 43. 6 % new 3 x 3 3 Stream 3/4 256 QAM 1. 120 Gbps 10. 5 % 69. 1 % 30 4 x 4 4 Stream 3/4 64 QAM 1. 120 Gbps 0. 3 % 48. 9 % 31 4 x 4 4 Stream 5/6 64 QAM 1. 244 Gbps 5. 5 % 76. 6 % new 4 x 4 4 Stream 7/8 64 QAM 1. 307 Gbps 17. 6 % 88. 5 % new 4 x 4 4 Stream 3/4 256 QAM 1. 493 Gbps 29. 5 % 97. 7 % *) Note: 0% PER means less then 15 errors in 15, 000 packets Submission 7 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 PHY Data Rates and PER at 40 d. B SNR (2) - Rx diversity (2 x 3, 3 x 4, and 4 x 5) MCS Configuration Code Rate Modulation Data Rate (400 ns GI) 80 MHz PER at 40 d. B SNR – TGn channel model D PER at 40 d. B SNR – TGn channel model B 14 2 x 3 2 Stream 3/4 64 QAM 560 Mbps 0. 0 % * 0. 0 % 15 2 x 3 2 Stream 5/6 64 QAM 622 Mbps 0. 0 % new 2 x 3 2 Stream 7/8 64 QAM 653 Mbps 0. 0 % new 2 x 3 2 Stream 3/4 256 QAM 747 Mbps 0. 0 % 22 3 x 4 3 Stream 3/4 64 QAM 840 Mbps 0. 0 % 23 3 x 4 3 Stream 5/6 64 QAM 933 Mbps 0. 0 % 0. 3 % new 3 x 4 3 Stream 7/8 64 QAM 980 Mbps 0. 0 % 0. 8 % new 3 x 4 3 Stream 3/4 256 QAM 1. 120 Gbps 0. 0 % 6. 4 % 30 4 x 5 4 Stream 3/4 64 QAM 1. 120 Gbps 0. 0 % 0. 7 % 31 4 x 5 4 Stream 5/6 64 QAM 1. 244 Gbps 0. 0 % 6. 2 % new 4 x 5 4 Stream 7/8 64 QAM 1. 307 Gbps 0. 0 % 11. 1 % new 4 x 5 4 Stream 3/4 256 QAM 1. 493 Gbps 0. 0 % 45. 3 % *) Note: 0% PER means less then 40 errors in 15, 000 packets Submission 8 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 MAC Throughput Analysis with PHY Results • Assume A-MPDU aggregation with RTS/CTS exchange • MAC throughput = (1 -PER) MAC efficiency = • Total MAC payload (bits) Time to transmit total MAC payload (sec) MAC throughput PHY rate A-MPDU limitations in 802. 11 n – Maximum size of A-MPDU = 64 KB – Maximum number of MPDUs in one A-MPDU =64 MPDUs Submission 9 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 MAC Throughput Results - Channel model D with the limitations in A-MPDU • With the limitations in A-MPDU Square configurations 3 x 3 Rx diversity configurations 4 x 5 4 x 4 3 x 4 2 x 2 • 2 x 3 No PHY configurations meet 1 Gbps MAC throughput Submission 10 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 MAC Throughput Results - Channel model D without the limitations in A-MPDU • Without the limitations in A-MPDU – Maximum size of A-MPDU > 64 KB – Maximum number of MPDUs in one A-MPDU > 64 MPDUs Square configurations 3 x 3 Rx diversity configurations 4 x 5 3 x 4 4 x 4 2 x 2 • • • 2 x 3 4 x 4 configuration just barely meets 1 Gbps MAC throughput For reliable 1 Gbps MAC throughput requires 5 antennas To achieve 1. 3 Gbps MAC throughput required A-MPDU of length 510 Kbytes (168 MPDUs aggregation) Submission 11 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 MAC Throughput Results - Channel model B without the limitations in A-MPDU • Without the limitations in A-MPDU Square configurations Rx diversity configurations 3 x 4 4 x 5 2 x 3 2 x 2 3 x 3 4 x 4 • • Even with no limitations in A-MPDU, 4 antennas doesn’t come close to achieving 1 Gbps MAC throughput 5 antennas just barely meets 1 Gbps MAC throughput Submission 12 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 Conclusions • Even with the PHY and MAC optimizations, 4 x 4 will not reliably achieve 1 Gbps MAC throughput • To achieve 1 Gbps MAC throughput required – 40 d. B SNR – 5 antennas – 510 Kbytes aggregate • The study group has two choices – Lower the throughput number in the below 6 GHz scope – Clarify in the below 6 GHz scope that 1 Gbps MAC throughput is optional in a device Submission 13 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 Backup Submission 14 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 PHY Simulation Results - Square configurations over channel model D • • Configurations: 2 x 2, 3 x 3, and 4 x 4 TGn channel model D Channel bandwidth: 80 MHz Packet size = 3000 bytes Submission 15 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 PHY Simulation Results - Square configurations over channel model B • • Configurations: 2 x 2, 3 x 3, and 4 x 4 TGn channel model B Channel bandwidth: 80 MHz Packet size = 3000 bytes Submission 16 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 PHY Simulation Results - Rx diversity configurations over channel model D • • Configurations: 2 x 3, 3 x 4, and 4 x 5 TGn channel model D Channel bandwidth: 80 MHz Packet size = 3000 bytes Submission 17 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 PHY Simulation Results - Rx diversity configurations over channel model B • • Configurations: 2 x 3, 3 x 4, and 4 x 5 TGn channel model B Channel bandwidth: 80 MHz Packet size = 3000 bytes Submission 18 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.

May 2008 doc. : IEEE 802. 11 -08/0535 r 0 MAC Throughput Results - Channel model B with the limitations in A-MPDU • Channel model B • With the limitations in A-MPDU Square configurations Submission Rx diversity configurations 19 Thomas Kenney, Minyoung Park, Eldad Perahia, Intel Corp.