Submission doc.: IEEE 11-11/1413r1 Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 1 Real Air-time Occupation by Beacon and Probe Date: Authors:
Submission doc.: IEEE 11-11/1413r1 Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 2 Abstract This document is not proposal. It is reporting the real Air-time occupation due to explosive increase of Smart- phones.
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 3 Motivation 3G mobile operators have demand to offload their data traffics to WLAN network. Especially, they have higher demands for the locations where many people meet or stay for data offloading, because high data traffics occur at those locations. It’s highly expected that FILS will realize transition from 3G to WLAN in very short time.
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 4 Real world (1) Number of Smart-phone is increasing. iPhone, Android, Windows-phone, Blackberry… Smart-phone holders always touch its screen. While its screen is activated (backlight turned on), Smart-phone starts searching surrounding WLAN-APs.
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 5 Real world (2) Many Smart-phone holders are in the crowded commuter train. Imagine what happens when the train arrives at the station. Air monitoring was executed at a train station in Tokyo. Results are explained in the following slides.
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 6 Conditions Time/Date: Around 18:00 / October 11(Tue), 2011 Location: Shinjuku station (Keio line), Tokyo Monitoring CH: 6CH(2,437MHz) Monitoring period: 300 seconds (5 minutes) Measured CH Thinkpad X200 Windows XP USB Wireless Monitor Adoptor (Air Pcap NX) Wireshark
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 7 Result Observed frames FramesBytes Count% % Beacon13, ,689, Probe Request7, , Probe Response24, ,941, Other46, ,581, Total92,159 17,201,802
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 8 Result (cont.) Time occupation is more important. Doc. IEEE /1031r0 was referred for time occupation analysis. BytesFrames
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 9 Transmission Rate Rate [Mbps] BeaconProbe RequestProbe ResponseOthers Frames% % % % 113, , , , , , , , , Total13,871 7,139 24,687 46,462
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 10 Transmission Rate (cont.) Rate [Mbps] BeaconProbe RequestProbe ResponseOthers Bytes% % % % 1 1,688, , ,849, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,285, , , ,034, Total 1,689, ,797 2,941, ,581,63 4
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 11 Occupied Time Calculation (Beacon) Occupied Time DIFSCWTX TIME aSlotTime:20us aSIFSTime:10us aPreambleLength:144us aPLCPHeaderLength:48bits aCWmin:31 aCWmax:1023 DIFS:50us CW:310us Occupied Time = ∑((DIFS + CW + aPreambleLength + aPLCPHeaderLength/DATARATE) * TotalFrames + (TotalBytes * 8/DATARATE)) = (( / 1.0) * 13,861 + (1,688,640 * 8 / 1.0) + (( / 11.0) * 2 + (80 * 8 / 11.0) + (( / 24.0) * 7 + (280 * 8 / 24.0) + (( / 54.0) * 1 + (40 * 8 / 54.0) = 21,165,613 us (7.06%) Beacon
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 12 Occupied Time Calculation (Probe Request) Occupied Time = ∑((DIFS + CW + aPreambleLength + aPLCPHeaderLength/DATARATE) * TotalFrames + (TotalBytes * 8/DATARATE)) = (( / 1.0) * 6,547 + (838,510 * 8 / 1.0) + (( / 2.0) * 21 + (10,019 * 8 / 2.0) + (( / 5.5) * 13 + (12,633 * 8 / 5.5) + (( / 6.0) * 2 + (1,590 * 8 / 6.0) + (( / 9.0) * 3 + (2,631 * 8 / 9.0) + (( / 11.0) * (76,565 * 8 / 11.0) + (( / 12.0) * 2 + (3,060 * 8 / 12.0) + (( / 18.0) * 7 + (10,198 * 8 / 18.0) + (( / 24.0) * 4 + (3,154 * 8 / 24.0) + (( / 48.0) * 10 + (13,509 * 8 / 48.0) + (( / 54.0) * 20 + (17,928 * 8 / 54.0) = 10,754,454us (3.58%) Occupied Time DIFSCWTX TIME Probe Request
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 13 Occupied Time Calculation (Probe Response) Occupied Time DIFSCWTX TIME Probe ResponseACK TX TIMESIFS aSlotTime:20us aSIFSTime:10us aPreambleLength:144us aPLCPHeaderLength:48bits aCWmin:31 aCWmax:1023 DIFS:50us CW:310us ACKRate:1Mbps ACKLength:14Bytes
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 14 Occupied Time Calculation (Probe Response) (Cont.) Occupied Time = ∑((DIFS + CW + aPreambleLength + aPLCPHeaderLength/DATARATE +aSIFSTime + aPreambleLength + aPLCPHeaderLeangth / ACKRATE + ACKLength * 8 / ACKRATE) * TotalFrames + (TotalBytes * 8/DATARATE)) = (( / /1.0+14*8/1.0) * 24,606 + (2,849,896 * 8 / 1.0) + (( / / * 8 / 1.0) * 4 + (5,257 * 8 / 2.0) + (( / / * 8 / 1.0) * 7 + (10,710 * 8 / 5.5) + (( / / * 8 / 1.0) * 1 + (1,530 * 8 / 6.0) + (( / / * 8 / 1.0) * 1 + (1,530 * 8 / 9.0) + (( / / * 8 / 1.0) * 15 + (13,848 * 8 / 11.0) + (( / / * 8 / 1.0) * 4 + (6,120 * 8 / 12.0) + (( / / * 8 / 1.0) * 7 + (8,212 * 8 / 18.0) + (( / / * 8 / 1.0) * 6 + (5,279 * 8 / 24.0) + (( / / * 8 / 1.0) * 3 + (4,590 * 8 / 36.0) + (( / / * 8 / 1.0) * 2 + (2,259 * 8 / 48.0) + (( / / * 8 / 1.0) * 31 + (32,100 * 8 / 54.0) = 44,215,439us (14.74%)
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 15 Occupied Time Calculation Result Packet type Occupancy rate (%) Occupied time (sec) Beacon Probe Req Probe Res Others Total
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 16 Conclusion Because of WLAN enabled devices increasing, especially Smart-phones, WLAN air circumstances are getting more crowded. In this packet monitoring, probe responses existed 5 times more than probe requests. To see the benefits of effective FILS, improvement of air circumstances would be needed by reducing unnecessary packet exchanges.
Submission doc.: IEEE 11-11/1413r1Nobember 2011 Katsuo Yunoki, KDDI R&D LaboratoriesSlide 17 References doc. IEEE /1031r0