Doc.: IEEE 802.11-08/1132r0 Submission September 2008 Eldad Perahia (Intel)Slide 1 Measurements of Coexistence between 802.11n 40MHz and Bluetooth SCO.

September 2008

Eldad Perahia (Intel)

Slide 1

doc.: IEEE 802.11-08/1132r0

Submission Slide 1

Measurements of Coexistence between 802.11n 40MHz and Bluetooth SCO

Date: 2008-09-10Authors:

Name Company Address Phone email Mohammad Mansour

Intel [email protected]

York Liu Intel [email protected]

Eldad Perahia Intel [email protected]

September 2008

Eldad Perahia (Intel)

Slide 2

doc.: IEEE 802.11-08/1132r0

Submission Slide 2

Abstract

Over-the-air measurements of Bluetooth SCO voice performance in the presence of 802.11n 40 MHz link. Results show no difference in impact to quality of nearby Bluetooth SCO voice link due to neighboring 802.11n 40MHz link as that of 802.11n 20 MHz link

September 2008

Eldad Perahia (Intel)

Slide 3

doc.: IEEE 802.11-08/1132r0

Submission Slide 3

Setup: Test Case I

• Wired controlled network on Gigabit switch• IxChariot 6.5 traffic is managed on controlled

network• WiFi link is 2.4GHz/40MHz capable• SCO session is created over BT link with

ability to control AFH on/off.• Windows XP SP2 with Chariot endpoint is

used for all stations• Clean RF (2.40-2.48GHz) environment

verified with spectrum analyzer

Gigabit Ethernet

Chariot Server/ConsoleChariot Endpoint

802.11n AP

11n WiFi Link

BT Control

BT

Lin

k

BT Master

WiFi STA

BT ControlBT Slave

September 2008

Eldad Perahia (Intel)

Slide 4

doc.: IEEE 802.11-08/1132r0

Submission Slide 4

Setup: Test Case II

• Wired controlled network on Gigabit switch• IxChariot 6.5 traffic is managed on controlled

network• WiFi link is 2.4GHz/40MHz capable• SCO session is created over BT1 link with

ability to control AFH on/off.• BT2 link used as an additional source of

interference.• Windows XP SP2 with Chariot endpoint is

used for all stations• Clean RF (2.40-2.48GHz) environment

verified with spectrum analyzer

Gigabit Ethernet

Chariot Server/ConsoleChariot Endpoint

802.11n AP

11n WiFi Link

WiFi STA

BT Control

BT

1 Li

nk

BT1 Master

BT ControlBT1 Slave

Gigabit Ethernet

BT2 Slave BT2 MatserBT2 Link

Chariot Management

September 2008

Eldad Perahia (Intel)

Slide 5

doc.: IEEE 802.11-08/1132r0

Submission

Deployment Layout

• Test setup is contained entirely in screen room.

• Lab room is a basic room– 15 ft x 15 ft

WiFi STASCO BT

Devices

WiFi AP

BT Link

Case II

September 2008

Eldad Perahia (Intel)

Slide 6

doc.: IEEE 802.11-08/1132r0

Submission Slide 6

Test Procedure : Case I• Controlled network:

– All nodes are on the same network– Chariot management traffic is on the controlled network

• Bluetooth baseline procedure:– All WiFi nodes are powered down– Bi-directional SCO link using HV3

• Voice traffic over SCO BT link from master to slave (~72sec standard reference wav file transmission)

• No silence cancellation, so there is still traffic from slave to master

– MOS score calculated based on comparison of reference file sent and received file using Opticom’s OPERA v3.5 voice quality analyzer (PESQ Algorithm).

– MOS score averaged over 5 iterations.

• Bluetooth with 802.11 traffic procedure:– Bluetooth procedures are the same as baseline– Add WiFi traffic for 90 Seconds– 802.11n (max and 22mbps) traffic: 4 pairs – two per direction (Tx, Rx), high

performance throughput.scr script– 802.11n (1 mbps) traffic: 1 pairs – one per direction (Tx, Rx), throughput.scr script

September 2008

Eldad Perahia (Intel)

Slide 7

doc.: IEEE 802.11-08/1132r0

Submission Slide 7

Test Procedure : Case II• Controlled network:

– All nodes are on the same network– Chariot management traffic is on the controlled network

• Bluetooth baseline procedure:– All WiFi nodes are powered down– Bi-directional SCO link using HV3

• Voice traffic over SCO BT link from master to slave (~72sec standard reference wav file transmission)

• No silence cancellation, so there is still traffic from slave to master– Bidirectional voice traffic on second BT link (1 pair Tx/Rx Chariot voice) as a

source of interference.– MOS score calculated based on comparison of reference file sent and received file

using Opticom’s OPERA v3.5 voice quality analyzer.– MOS score averaged over 5 iterations.

• Bluetooth with 802.11 traffic procedure:– Bluetooth procedures are the same as baseline– Add WiFi traffic for 90 Seconds– 802.11n (max and 22mbps) traffic: 4 pairs – two per direction (Tx, Rx), high

performance throughput.scr script– 802.11n (1 mbps) traffic: 1 pairs – one per direction (Tx, Rx), throughput.scr script

September 2008

Eldad Perahia (Intel)

Slide 8

doc.: IEEE 802.11-08/1132r0

Submission Slide 8

Test Methodology

• Measurements made in accordance with test scenarios outlined in 08/971r0

• Bluetooth devices comprised of two development platforms from leading BT vendor

• Bluetooth devices forming Bluetooth link separated by 1 m to model typical distance between headset and cell phone

• Bluetooth link separated from the 802.11 link by 0.5 m to model scenario where WLAN connectivity is being using simultaneously with Bluetooth voice call

• Test measurements conducted for 802.11n 40 MHz and 802.11n 20 MHz• Test conducted with a range of 802.11 throughput

– Max throughput for each 802.11 system– Comparable throughputs for each system to compare impact based on the same

offered load– Minimum 802.11 throughput tested is 1 Mbps to model typical internet access

September 2008

Eldad Perahia (Intel)

Slide 9

doc.: IEEE 802.11-08/1132r0

Submission Slide 9

Measurement Baseline

BT Mean Opinion Score (MOS) - No 802.11 present

802.11n Throughput – no BT present

Test Scenario 1 4.0 40 MHz: 155 Mbps

20 MHz: 71 MbpsTest Scenario 2 4.0

September 2008

Eldad Perahia (Intel)

Slide 10

doc.: IEEE 802.11-08/1132r0

Submission Slide 10

Test Scenario 1 – AFH off

• 802.11 traffic type: TCP

• BT link same as baseline

SCO, AFH off

0.0

1.0

2.0

3.0

4.0

Max TPT 22Mbps 1Mbps

11n offered load

BT

MO

S s

co

re

BT+WiFi2.4/40 11n

BT+WiFi2.4/20 11n

11n 20 MHz: 44 Mbps

11n 40 MHz: 115 Mbps

September 2008

Eldad Perahia (Intel)

Slide 11

doc.: IEEE 802.11-08/1132r0

Submission Slide 11

Test Scenario 1 – AFH on

• 802.11 traffic type: TCP

• BT link same as baseline

SCO, AFH on

0.0

1.0

2.0

3.0

4.0

Max TPT 22Mbps 1Mbps

11n offered load

BT

MO

S s

core

BT+WiFi2.4/40 11n

BT+WiFi2.4/20 11n

11n 20 MHz: 55 Mbps

11n 40 MHz: 136 Mbps

September 2008

Eldad Perahia (Intel)

Slide 12

doc.: IEEE 802.11-08/1132r0

Submission Slide 12

Test Scenario 2

• 802.11 traffic type: TCP

• BT links same as baseline

AFH off AFH on

11n 20 MHz

Max TPT

11n 40 MHz

Max TPT

11n 20 MHz

Max TPT

11n 40 MHz

Max TPT

BT MOS 2.6 2.4 3.9 3.7

802.11n TPT

42 Mbps 90 Mbps 52 Mbps 105 Mbps

• BT MOS are identical to Test Scenario 1• 802.11n throughput results degraded compared to Test

Scenario 2

September 2008

Eldad Perahia (Intel)

Slide 13

doc.: IEEE 802.11-08/1132r0

Submission Slide 13

Analysis of Results

• All BT MOS measurements with AFH-on in the presence of 802.11n 40MHz are comparable to that of the BT baseline for both Test Case 1 and 2

• BT MOS score improves with decreasing 802.11n offered load with AFH-off in Test Case 1

• No measurable impact due to offered load with AFH-on in Test Case 1

• No measurable difference between 802.11n 40 MHz and 802.11n 20 MHz with regards to BT performance

• No change in BT MOS score between Test Case 1 and 2

• 802.11n 20 MHz and 40 MHz max throughput is degraded by BT even with AFH on

September 2008

Eldad Perahia (Intel)

Slide 14

doc.: IEEE 802.11-08/1132r0

Submission Slide 14

Conclusions

• Over-the-air measurements conducted to determine coexistence between 802.11n 40 MHz and Bluetooth SCO voice link

• Measurements demonstrate that neighboring 802.11n 40 MHz link has no difference in impact to quality of nearby Bluetooth SCO voice link as that of 802.11n 20 MHz link

• All BT SCO MOS measurements with AFH-on in the presence of 802.11n 40MHz are comparable to that of the BT SCO baseline