1. Qualcomm MAC Supplementary Presentation
Month 2002
doc.: IEEE /xxxr0
November 2004
Qualcomm MAC Supplementary Presentation
Sanjiv Nanda, John Ketchum
QUALCOMM, Inc.
Sanjiv Nanda, Qualcomm, Inc.
John Doe, His Company

2. List of Topics Backward Compatibility and Legacy Sharing Scalability
November 2004
List of Topics
Backward Compatibility and Legacy Sharing
Scalability
Future-proof Highest Performance
Permit low complexity implementations for low capability devices
Highlights of Throughput and Performance Results
Essential MAC features and quantification of benefits
Block Ack Mechanism
20/40 MHz Operation
Sanjiv Nanda, Qualcomm, Inc.

3. Backward Compatibility
November 2004
Backward Compatibility
802.11a/g Preamble and SIGNAL field is unchanged.
RATE field in a/g SIGNAL is set to a value “undefined” in a/g
Legacy STAs will abandon further decoding of the PPDU and go to CCA when they detect an undefined value of RATE field.
This is different from “spoofing” specified in other proposals. Spoofing provides protection but at a cost:
Legacy SIGNAL field is consumed. Benefit has not been quantified.
Power drain at legacy STAs.
Sanjiv Nanda, Qualcomm, Inc.

4. November 2004
Legacy Sharing: CC15
CC15 Results are provided for the case of a single n STA-AP link sharing the medium with:
802.11g STA-AP link in the 2.4 GHz band
802.11a STA-AP link in the 5 GHz band
We show that the n MAC allows the legacy STA to “fairly” share the bandwidth.
Legacy STA only: 24.4 Mbps
Legacy STA in shared network: 11.9 Mbps
The MAC permits other possible sharing policies
Sanjiv Nanda, Qualcomm, Inc.

5. November 2004
Legacy Sharing: CC15
Sanjiv Nanda, Qualcomm, Inc.

6. Scalability Scalable, future-proof, highest performance
November 2004
Scalability
Scalable, future-proof, highest performance
High data rates with up to 4 streams
Highest throughputs with Eigenvector Steering
Lowest latency MAC operation with ACF
However, scalable design also permits low complexity implementations for low capability STAs
Low complexity PHY receiver designs to accommodate
Decoding delays
MIMO processing delays
Receiver limit on maximum aggregate frame size
Scheduled operation for lowest power consumption
Sanjiv Nanda, Qualcomm, Inc.

7. Throughput and Performance Highlights
November 2004
Throughput and Performance Highlights
We have shown in Berlin:
Significantly higher throughput compared to other proposals.
100 Mbps BSS throughput in realistic scenarios with 20 MHz BW
Significantly higher range compared to other proposals.
Updated results show further improvement.
Sanjiv Nanda, Qualcomm, Inc.

8. Summary of System Simulation Results
November 2004
Summary of System Simulation Results
Metric
Scenario
ACF 2x2
ACF 4x4
CC3 Aggregate goodput (Metric 2) [Mbps]
Scenario 1 HT
105.9
193.5
Scenario 4
119.4
237.3
Scenario 6 EXT
69.4
137.0
CC18 Aggregate non-QoS throughput [Mbps]
53.6
141.2
108.1
224.2
23.7
86.5
CC19 Number of QoS flows supported
16 / 17
17 / 17
18 / 18
37 / 39
39 / 39
CC58 HT spectral efficiency [bps/Hz]
5.9
Parameters
5.25 GHz; BW = 20 MHz
SGI-52 OFDM symbols, except for Scenario 6 (Hot Spot)
Significantly higher throughput compared to other proposals.
Sanjiv Nanda, Qualcomm, Inc.

9. November 2004
Extended Scenarios
Mandatory Scenarios have been extended to demonstrate the capabilities of our design.
Scenario 1 HT is an extension of Scenario 1:
Additional FTP flow of up to 130 Mbps at 15.6 m from the AP.
Scenario 1 EXT is an extension of Scenario 1:
Maximum delay requirement for all video/audio streaming flows is decreased from 100/200 ms to 50 ms. More realistic.
Two HDTV flows are moved from 5 m from the AP, to 25 m from the AP. Representative of real deployments.
Scenario 6 EXT is an extension of Scenario 6:
One FTP flow of 2 Mbps at 31.1 m from the AP is increased up to 80 Mbps for 4x4.
Sanjiv Nanda, Qualcomm, Inc.

10. Throughput versus Range for Channel Model B
November 2004
Throughput versus Range for Channel Model B
Throughput above the MAC of 100 Mbps is achieved at:
29 m (35 m) for 2x2, 5.25 GHz for Channel Model B (D).
40 m (54 m) for 2x2, 2.4 GHz
47 m (65 m) for 4x4, 5.25 GHz
75 m (102 m) for 4x4, 2.4 GHz
Significantly higher range compared to other proposals.
Sanjiv Nanda, Qualcomm, Inc.

11. Throughput versus Range for Channel Model D
November 2004
Throughput versus Range for Channel Model D
Throughput above the MAC of 100 Mbps is achieved at:
29 m (35 m) for 2x2, 5.25 GHz for Channel Model B (D).
40 m (54 m) for 2x2, 2.4 GHz
47 m (65 m) for 4x4, 5.25 GHz
75 m (102 m) for 4x4, 2.4 GHz
Significantly higher range compared to other proposals.
Sanjiv Nanda, Qualcomm, Inc.

12. Adaptive Coordination Function
November 2004
Adaptive Coordination Function
ACF Features
No Immediate ACK
For BlockAckRequest and BlockAck
SCHED
PPDU Aggregation with Reduced and Zero IFS
Multi-Poll
Low latency closed loop operation
Achieve higher PHY rates de to smaller backoff and MIMO Mode Control
Sanjiv Nanda, Qualcomm, Inc.

13. Benefits of ACF ACF versus HCF with Frame Aggregation
November 2004
Benefits of ACF
ACF versus HCF with Frame Aggregation
Throughput Gain in Scenario 1: 70%
Sanjiv Nanda, Qualcomm, Inc.

14. Benefits of ACF ACF Features No Immediate ACK SCHED
November 2004
Benefits of ACF
ACF Features
No Immediate ACK
For BlockAckRequest and BlockAck
MAC Efficiency Gain over HCF with Frame Aggregation: ~18%
SCHED
PPDU Aggregation with Reduced and Zero IFS
Multi-Poll
MAC Efficiency Gain over HCF with Frame Aggregation : ~7%
Low latency closed loop operation
Achieve higher PHY rates due to smaller backoff and MIMO Mode Control
Mean PHY Rate Gain over HCF: ~35%
Sanjiv Nanda, Qualcomm, Inc.

15. Benefit of Data Rate Feedback
November 2004
Benefit of Data Rate Feedback
16-bit field in PLCP header extension specifies up to four preferred rates
Tx PHY rate is maximized after single ACK received
Accurate PHY rate tracking for time varying channels
Substantial throughput gains:
Scenario 1: 41%
82.6 Mbps versus 58.7 Mbps
Scenario 6: 46%
81.5 Mbps versus 55.9 Mbps
Sanjiv Nanda, Qualcomm, Inc.

16. Benefit of Eigensteering
Month 2002
doc.: IEEE /xxxr0
November 2004
Benefit of Eigensteering
Comparison of Spatial Spreading only with closed loop MIMO Mode Selection
Sample Gains:
30% throughput gain in Scenario 1.
40% in Scenario 4.
Sanjiv Nanda, Qualcomm, Inc.
John Doe, His Company

17. Reduced Complexity PHY Implementations
Month 2002
doc.: IEEE /xxxr0
November 2004
Reduced Complexity PHY Implementations
MAC design choices must not impose excessive complexity on PHY implementation
Allow low complexity PHY implementations for low capability devices, e.g., VoIP phone, PDA.
Fewer antennas.
MIMO processing delay.
Delayed decoding.
Permit STA designs with reduced PHY complexity
Limit on reception of Aggregate frames,
Limit on reception of Aggregate PPDUs,
Turn-around time for Block Ack.
Turn-around time for estimation of steering vectors.
Sanjiv Nanda, Qualcomm, Inc.
John Doe, His Company

18. Frame Aggregation On-the-fly frame aggregation at the transmitter
Month 2002
doc.: IEEE /xxxr0
November 2004
Frame Aggregation
On-the-fly frame aggregation at the transmitter
Flexible, without additional complexity.
Maximum aggregate size in Block Ack negotiation.
To permit low receiver complexity.
No frame aggregation across multiple RAs.
Requires reception of the largest Multiple-RA aggregate. Unnecessary burden on receiver complexity.
Use PPDU Aggregation instead.
Sanjiv Nanda, Qualcomm, Inc.
John Doe, His Company

19. PPDU Aggregation Improved throughput efficiency SCHED Frame:
Month 2002
doc.: IEEE /xxxr0
November 2004
PPDU Aggregation
Improved throughput efficiency
Reduced or Zero IFS
No preamble if Tx power is unchanged
SCHED Frame:
Message indicates TA and RA, start offset and duration for scheduled TXOPs.
Advantages of PPDU Aggregation with SCHED:
Inclusion of RA and start offset in SCHED means STA needs to decode only its own PPDU rather than the entire aggregate.
Permits reduced receiver complexity.
Permits optimum sleep mode.
Start offset is with respect to the SCHED frame transmission. No global synchronization issue.
May be used by AP or STA
Sanjiv Nanda, Qualcomm, Inc.
John Doe, His Company

20. November 2004
PPDU Aggregation
SCAP (Scheduled Access Period) initiated by SCHED message
Acts as consolidated multi-STA poll
Indicate TA, RA, start offset and duration of TXOP.
Permits effective PPDU Aggregation
Eliminate Immediate ACK for Block Ack frames
MIMO training in SCHED message functions as broadcast sounding waveform for channel estimation and SVD calculation
Sanjiv Nanda, Qualcomm, Inc.

21. Compression MAC Header Compression Compressed Block Ack
Month 2002
doc.: IEEE /xxxr0
November 2004
Compression
MAC Header Compression
Compressed Header Formats: Eliminate, TA, RA, Duration/ID fields
Gain: unavailable.
Compressed Block Ack
Transmitter option. Receiver mandatory.
Gain: 2% throughput gain. Results for Scenario 1 and Scenario 6.
Sanjiv Nanda, Qualcomm, Inc.
John Doe, His Company

22. Robust Block Ack Operation
Month 2002
doc.: IEEE /xxxr0
November 2004
Robust Block Ack Operation
The e Block Ack mechanism offers a flexible Window-Based Selective Reject ARQ engine.
Does not require Immediate Block Ack to maximize throughput.
Keep the “pipe” full; avoid window “stalling.”
Simple corrections/clarifications of e Block Ack mechanism.
“Synchronized” operation between frame transmissions and Block Acks is not required.
Block Acks need not be acknowledged.
Operates seamlessly if Block Acks are lost.
Operates seamlessly if Block Acks are delayed.
Block Ack Requests can be implicit.
Sanjiv Nanda, Qualcomm, Inc.
John Doe, His Company

23. Operation with Lost BlockAck Frame
November 2004
Operation with Lost BlockAck Frame
Frame Transmissions can continue even if Block Acks are lost.
Frames EFGH can be transmitted even though BlockAck is lost.
Up to size of window.
Immediate ACK of BlockAck or BlockAckRequest Frame is not required
Sanjiv Nanda, Qualcomm, Inc.

24. Operation with Decoding Delay
November 2004
Operation with Decoding Delay
Frame Transmissions can continue even with Decoding Delay
Frames EFGH can be transmitted even if Frames ABCD are still being decoded at the receiver
BAR Request Count is “echoed” back to the transmitter
Indicates last received BAR Request Count
Frames EFGH should not be retransmitted.
Sanjiv Nanda, Qualcomm, Inc.

25. Month 2002
doc.: IEEE /xxxr0
November 2004
Robust 40/20 MHz Operation
40 MHz channels are defined as (2n, 2n+1) carrier pairs. (2n+1, 2n+2) pair is not allowed.
Primary and Secondary Carriers
Ensures that overlapping 40 MHz BSS always have the same primary carrier.
Medium Access (CSMA/CA) is managed on the Primary carrier.
CCA must be done on Secondary carrier also.
Secondary Carrier Interference Events (SCIE) are tracked by AP and STAs.
Excessive SCIE Count implies overlap with a 20 MHz BSS on secondary.
Procedures are defined to eliminate overlapping 20 MHz BSS on Secondary.
Sanjiv Nanda, Qualcomm, Inc.
John Doe, His Company

26. November 2004
40 MHz Channel Pairs
40 MHz channels are defined as (2n, 2n+1) carrier pairs. (2n+1, 2n+2) pair is not allowed.
Carrier 2n: Primary carrier
Carrier 2n+1: Secondary carrier
Ensures that overlapping 40 MHz BSS always have the same primary carrier.
Sanjiv Nanda, Qualcomm, Inc.

27. November 2004
Overlapping 40 MHz BSS
Medium Access (CSMA/CA) is managed on the Primary carrier.
Overlapping 40 MHz BSS have the same primary carrier.
Sanjiv Nanda, Qualcomm, Inc.

28. November 2004
Overlapping 40/20 MHz BSS
Medium Access (CSMA/CA) is managed on the Primary carrier.
Overlapping 20 MHz BSS on the primary carrier is permitted.
Procedures are defined to eliminate overlapping 20 MHz BSS on secondary.
Sanjiv Nanda, Qualcomm, Inc.

29. Overlapping 20 MHz BSS on Secondary
November 2004
Overlapping 20 MHz BSS on Secondary
Clear Channel Assessment must be done on Secondary carrier also.
During CCA, if there is a transmission on the secondary. Transmit only on primary.
Secondary Carrier Interference Events (SCIE) are tracked by AP and STAs.
Sanjiv Nanda, Qualcomm, Inc.

30. Secondary Carrier Interference Events
November 2004
Secondary Carrier Interference Events
Secondary Carrier Interference Events (SCIE) are tracked by AP and STAs.
During reception of a 20 MHz transmission on the primary if there is energy on the secondary.
During CCA, if there is a transmission on the secondary.
During reception of a 40 MHz transmission if there is lower SNR on the secondary.
Sanjiv Nanda, Qualcomm, Inc.

31. November 2004
Fall-Back to 20 MHz
Excessive SCIE Counts imply overlap with a 20 MHz BSS on secondary.
Mandatory fall-back to 20 MHz operation.
Can move to another 40 MHz FA.
Sanjiv Nanda, Qualcomm, Inc.

32. Robust 40/20 MHz Operation: Summary
Month 2002
doc.: IEEE /xxxr0
November 2004
Robust 40/20 MHz Operation: Summary
40 MHz channels are defined as (2n, 2n+1) carrier pairs. (2n+1, 2n+2) pair is not allowed.
Carrier 2n: Primary carrier
Carrier 2n+1: Secondary carrier
Ensures that overlapping 40 MHz BSS always have the same primary carrier.
Medium Access (CSMA/CA) is managed on the Primary carrier.
Overlapping 40 MHz BSS have the same primary carrier.
Overlapping 20 MHz BSS on the primary carrier is permitted.
Procedures are defined to eliminate overlapping 20 MHz BSS on secondary.
CCA must be done on Secondary carrier also.
During CCA, if there is a transmission on the secondary. Transmit only on primary.
Secondary Carrier Interference Events (SCIE) are tracked by AP and STAs.
During CCA, if there is a transmission on the secondary.
During reception of a 40 MHz transmission if there is lower SNR on the secondary.
During reception of a 20 MHz transmission on the primary if there is energy on the secondary.
Excessive SCIE Count implies overlap with a 20 MHz BSS on secondary.
Mandatory fall-back to 20 MHz operation.
Can move to another 40 MHz FA.
Sanjiv Nanda, Qualcomm, Inc.
John Doe, His Company