1. OFDMA performance in 11ax
March 2016
OFDMA performance in 11ax
Date:
Authors:
Name
Affiliations
Address
Phone
Suhwook Kim
LG Electronics
19, Yangjea-daero 11gil, Seocho-gu, Seoul , Korea
Hyunhee Park
Jeongki Kim
Kiseon Ryu
HanGyu Cho
Suhwook Kim, LG Electronics

2. Month Year
doc.: IEEE yy/xxxxr0
March 2016
Introduction
In January, we addressed OFDMA performance when we use simple MU-RTS/CTS and CCA after trigger frame
RTS/CTS is more powerful in many STAs case and legacy cases
Nevertheless, OFDMA shows better throughput
This presentation provides update on it
Suhwook Kim, LG Electronics
John Doe, Some Company

3. Simulation update MU-RTS/CTS Preamble detection of STA side
March 2016
Simulation update
MU-RTS/CTS
We evaluate MU-RTS/CTS scheme which are defined in our SFD
Preamble detection of STA side
A STA can detect multiple preambles on different channel
BSS channelization
Channelization of each BSS can be different
Suhwook Kim, LG Electronics

4. Simulation Setup: Scheduler
March 2016
Simulation Setup: Scheduler
Scheduling resource
Maximum allocation per one station is 1 RU
1 RU: 242-tone (total 4 RUs in 80 MHz)
Scheduling policy
Random: AP selects STAs randomly
In DL case,
primary destination is fixed (Head of AP’s queue)
frame length is determined by A-MPDU length to primary destination (legacy spec rule)
Suhwook Kim, LG Electronics

5. MU-RTS/CTS in DL-OFDMA
March 2016
MU-RTS/CTS in DL-OFDMA
We evaluate two MU-RTS/CTS schemes
Simple MU-RTS/CTS
STAs can transmit CTS frame on secondary channel only
SFD MU-RTS/CTS [1]
Primary channel should be included when STAs transmit CTS frame
Suhwook Kim, LG Electronics

6. MU-RTS/CTS in DL-OFDMA
March 2016
MU-RTS/CTS in DL-OFDMA
Simple MU-RTS/CTS
Suhwook Kim, LG Electronics

7. MU-RTS/CTS in DL-OFDMA
March 2016
MU-RTS/CTS in DL-OFDMA
SFD MU-RTS/CTS
Suhwook Kim, LG Electronics

8. Simulation Setup: Parameters
March 2016
Simulation Setup: Parameters
Simulator Type
PHY/MAC Integrated simulator
Tx Power (AP/STA)
20/15 dBm (Topology 3)
Antenna Gain (AP/STA)
0/-2 dBi
Traffic Model
DL and UL mixed (CBR)
BSS Bandwidth
80MHz
Noise Figure
Noise Floor
-101dBm per 20MHz
Rate Control Algorithm
Open Loop Link Adaption [3]
MSDU size (bytes)
1472
Feedback
GENIE
Max Retx 10
Symbol length
4 usec (legacy), 16usec (OFDMA)
Metrics
Throughput
Queue Size
AP: 1000 * # of associated STA, STA: 1000
Suhwook Kim, LG Electronics

9. Topology Topology description Simulation setting
March 2016
Topology
Topology description
20 APs (Fixed location: center of room)
4 or 20 STAs per one AP (Random location)
Simulation setting
Open Loop Link Adaption [3]
Channelization: Fixed channel
Trigger frame is generated every 10 msec period and trigger frame has priority over data frame
CH1
CH2
CH3
Suhwook Kim, LG Electronics

10. March 2016
Simulation cases
Case 1: BSS channelization is fixed (all channelization 1)
Case 1-1: 4 STAs, DL/UL 50 Mbps
Case 1-2: 20 STAs, DL/UL 4 Mbps
Case 2: BSS channelization is random (1 or 2)
2-1: 4 STAs, DL/UL 50 Mbps
2-2: 20 STAs, DL/UL 50 Mbps
Suhwook Kim, LG Electronics

11. Simulation Result – Case 1
Month Year
doc.: IEEE yy/xxxxr0
March 2016
Simulation Result – Case 1
Case 1-1: 4 STAs, DL/UL 50 Mbps
Suhwook Kim, LG Electronics
John Doe, Some Company

12. Simulation Result – Case 1
Month Year
doc.: IEEE yy/xxxxr0
March 2016
Simulation Result – Case 1
Case 1-2: 20 STAs, DL/UL 4 Mbps
Suhwook Kim, LG Electronics
John Doe, Some Company

13. Simulation Result – Case 1
Month Year
doc.: IEEE yy/xxxxr0
March 2016
Simulation Result – Case 1
Throughput comparison
PER
Tput [Mbps]
NO RTS/CTS
RTS/CTS-Simple
RTS/CTS-SFD
4 STAs DL
74.93
74.05
74.59 UL
122.76
115.12
114.96
SUM
197.69
189.18
189.55
20 STAs
58.38
63.84
63.50
77.65
76.61
75.92
136.03
140.45
139.42
PER [%]
NO RTS/CTS
RTS/CTS-Simple
RTS/CTS-SFD
4 STAs DL
13.67
8.56
9.04 UL
7.17
6.08
6.15
20 STAs
16.08
11.03
11.20
8.83
7.85
7.88
Suhwook Kim, LG Electronics
John Doe, Some Company

14. Simulation Result – Case 2
Month Year
doc.: IEEE yy/xxxxr0
March 2016
Simulation Result – Case 2
2-1: 4 STAs, DL/UL 50 Mbps
Suhwook Kim, LG Electronics
John Doe, Some Company

15. Simulation Result – Case 2
Month Year
doc.: IEEE yy/xxxxr0
March 2016
Simulation Result – Case 2
2-2: 20 STAs, DL/UL 50 Mbps
Suhwook Kim, LG Electronics
John Doe, Some Company

16. Simulation Result – Case 2
Month Year
doc.: IEEE yy/xxxxr0
March 2016
Simulation Result – Case 2
Throughput comparison
PER
　Tput [Mbps]
NO RTS/CTS
RTS/CTS-Simple
RTS/CTS-SFD
4 STAs DL
69.83
73.73
72.54 UL
121.73
114.61
114.95
SUM
191.56
188.33
187.49
20 STAs
71.02
70.99
71.64
124.55
116.28
117.18
195.57
187.28
188.82
PER [%]
NO RTS/CTS
RTS/CTS-Simple
RTS/CTS-SFD
4 STAs DL
13.70
9.68
9.01 UL
7.10
6.55
6.03
20 STAs
13.66
10.48
10.25
7.02
6.39
6.40
Suhwook Kim, LG Electronics
John Doe, Some Company

17. Discussion RTS/CTS could not show better throughput in OFDMA
March 2016
Discussion
RTS/CTS could not show better throughput in OFDMA
PER of OFDMA is around 5 ~ 15 % even through RTS/CTS is not used. The reasons are
Reduced contending
Frequency diversity
So, a room for throughput enhancement by RTS/CTS is limited
Because of the situation, RTS/CTS-Simple and RTS/CTS-SFD couldn’t have different performance
We will evaluate RTS/CTS performance in following environments to verify our discussion
OFDMA with uplink EDCA contending
Dense AP
Suhwook Kim, LG Electronics

18. Next Step Following items will be added to OFDMA simulation RTS/CTS
March 2016
Next Step
Following items will be added to OFDMA simulation
RTS/CTS
OFDMA with uplink EDCA contending
Dense AP
Feedback modeling
Sophisticated schedulers
Different scenarios/traffic model
Suhwook Kim, LG Electronics

19. March 2016
Conclusion
We addressed OFDMA performance when we use MU-RTS/CTS which are defined in SFD
We evaluated three cases
NO RTS/CTS
RTS/CTS-Simple
RTS/CTS-SFD
However, No RTS/CTS, RTS/CTS-Simple, and RTS/CTS-SFD show very similar performance
Suhwook Kim, LG Electronics

20. Reference [1] 11-15/0132r15 Spec Framework
March 2016
Reference
[1] 11-15/0132r15 Spec Framework
[2] 11-14/0980r14 Simulation Scenarios
[3] 11-14/620r0 link adaptation for PHY SLS calibration
[4] 11-14/0571r10 11ax Evaluation Methodology
Suhwook Kim, LG Electronics