January 2012 doc IEEE 802 11 120034 r

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Performance Eveluation Date: 2012 -01 -12 Authors: Name Affiliations Address Phone email Hitoshi MORIOKA Allied Telesis R&D Center 2 -14 -38 Tenjin, Chuo-ku, Fukuoka 810 -0001 JAPAN +81 -92 -771 -7630 hmorioka@root-hq. com Submission Slide 1 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Abstract This document describes an example of measurement and performance evaluation of existing protocol. Submission Slide 2 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Example of Existing Protocols • • Authentication: EAP-TTLS/MS-CHAPv 2 Key Exchange: EAPOL Key IP address assignment: DHCPv 4 Address Resolution: ARP Submission Slide 3 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Sequence of Existing Protocols STA Auth Assoc AP AS EAP-TTLS /MS-CHAPv 2 EAPOL Key DHCP Server DHCP Gateway ARP Submission Slide 4 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Performance Definition • Link Setup Latency – from non-AP STA transmits Authentication – to complete resolution of the default gateway MAC address – This shows how fast a non-AP STA to setup the link. • Occupied Airtime – Total airtime occupied by each frame for one non-AP STA to complete link setup. – This includes transmission time, IFSs, CW and ACK transmission time (unicast). – This shows how many non-AP STAs can be accomodated. Link Setup Latency ACK ARP Reply Auth SIFS DIFS CW Occupied Airtime (Auth) Submission Occupied Airtime (ARP Reply) Slide 5 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Link Setup Latency Measurement • Most of Link Setup Latency is caused by the latency of processing on STA, AP and servers. • So I measured the latency of existing protocols. • Latency is strongly depends on the environment. • It’s just an example. WLAN I/F (to capture WLAN frames) STA i. Phone 4 AS DHCP Server Gateway AP Internet Capture both WLAN frames and Ethernet frames. (for timestamp syncronization) Pentium. M 1. 7 GHz, Free. BSD Submission Slide 6 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Measured Latency Fragmented Submission Frame Auth Req Auth Rep Assoc Req Assoc Resp EAP Req Id EAP Resp Id EAP Req TTLS EAP Resp Cl Hello EAP Req Sv Hello, Cert EAP Resp EAP Req Sv Hello, Cert TLS Cl key exch. TLS Cipher Spec… MSCHAP ID MSCHAP Challenge EAP Success EAPOL Key 1 EAPOL Key 2 EAPOL Key 3 EAPOL Key 4 STA AP 1 20 150 33 29 26 19 20 3 4 > < < > < > < > Slide 7 Server Subtotal 2 2 90 2 32 2 11 1 16 1 15 1 8 1 3 1 83 2 5 > < > < 1 0 0 0 25 0 0 590 5 14 [ms] Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Measured Latency (Cont’d) Frame DHCPDISCOVER DHCPOFFER Wait for other offer DHCPREQUEST DHCPACK Duplicate Address ARP Req. Check & Gratuitous ARP Rep. ARP Total STA AP 34 1035 1629 > < > < Server 1 62 1 3 1 1 347 3003 > < > < Subtotal Duplicate Address Check 193 1 1330 0 220 1631 3570 [ms] Optimized (Optimistic & Aggressive) DHCP Implementation Assuming same as below From receipt of DHCPACK to transmission of ARP for duplication check Frame DHCPDISCOVER DHCPOFFER DHCPREQUEST DHCPACK ARP Req. ARP Rep. Total STA AP 34 5 5 > < > < 349 Server 1 3 1 1 288 Maybe environmental issue (congestion) Submission Slide 8 > < > < Subtotal 0 1 0 27 From receipt of DHCPDISCOVER to 48 transmission of ARP 7 664 [ms] more than 80% reduced Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Occupied Airtime Calculation (DS 1) • – Parameters – – – – • TXRate: (DS 1) a. Slot. Time: a. SIFSTime: a. Preamble. Length: a. PLCPHeader. Length: a. CWmin: a. CWmax: 1 Mbps – 20 us 144 us 48 us 31 1023 – – DIFS = a. SIFSTime+2*a. Slot. Time = 50 us CWave = a. CWmin*a. Slot. Time/2 = 310 us (No contention assumed) ACKLength: 18 octets Frame. Length (inluding MAC Header): n octets Rough Occupied Airtime by n octets frame (including MAC header and FCS) • Broadcast from AP (no ACK) Tbroadcast(n) = a. Preamble. Length+a. PLCPHeader. Length+n/TXRate+DIFS+CWave = 144+48+n/1+50+310 [us] = n+552 [us] • Other Tunicast(n) = Tbroadcast(n)+a. Preamble. Length+a. PLCPHeader. Length+ACKLength/TXRate+a. SIFSTime = n+552+144+48+18/1+10 [us] = n+772 [us] Submission Slide 9 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Occupied Airtime Calculation (OFDM 6) • – Parameters – – – – • TXRate: (OFDM 6) a. Slot. Time: a. SIFSTime: a. Preamble. Length: a. PLCPHeader. Length: a. CWmin: a. CWmax: 6 Mbps – 9 us 16 us 4 us 15 1023 – – DIFS = a. SIFSTime+2*a. Slot. Time = 34 us CWave = a. CWmin*a. Slot. Time/2 = 67. 5 us (No contention assumed) ACKLength: 18 octets Frame. Length (inluding MAC Header): n octets Rough Occupied Airtime by n octets frame (including MAC header and FCS) • Broadcast from AP (no ACK) Tbroadcast(n) = a. Preamble. Length+a. PLCPHeader. Length+n/TXRate+DIFS+CWave = 16+4+n/6+34+67. 5 [us] = n/6+121. 5 [us] • Other Tunicast(n) = Tbroadcast(n)+a. Preamble. Length+a. PLCPHeader. Length+ACKLength/TXRate+a. SIFSTime = n/6+121. 5+16+4+18/6+16 [us] = n/6+160. 5 [us] Submission Slide 10 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Occupied Airtime Frame Size Airtime Auth Req 45 Auth Rep 34 Assoc Req 90 Assoc Resp 50 EAP Req Id 50 EAP Resp Id 49 EAP Req TTLS 46 EAP Resp Cl Hello 192 EAP Req Sv Hello, Cert 1064 EAP Resp 46 EAP Req Sv Hello, Cert 232 TLS Cl key exch. 376 TLS Cipher Spec… 109 MSCHAP ID 183 MSCHAP Challenge 135 MSCHAP Challenge 46 EAP Success 44 EAPOL Key 1 135 EAPOL Key 2 157 EAPOL Key 3 191 EAPOL Key 4 135 Submission (DS 1) Subtotal (payload) Airtime (OFDM 6) Subtotal (payload 817 168 806 166 862 176 822 3307 (219) 169 679 (37) 822 169 821 169 818 168 964 193 1836 338 818 168 1004 199 1148 223 881 179 955 191 907 183 818 168 816 15262 (3682) 168 3021 (614) 907 183 929 187 963 192 907 3706 (618) 183 745 (103) Slide 11 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Occupied Airtime (Cont’d) Frame DHCPDISCOVER DHCPOFFER DHCPREQUEST DHCPACK ARP Req ARP Rep Total Submission Size Airtime 380 380 80 98 (DS 1) Subtotal (payload) Airtime (OFDM 6) Subtotal (payload 1152 224 1152 4608 (1520) 224 895 (253) 852 174 870 1722 (178) 177 351 (30) 28605 (6217) 5691 (1036) Slide 12 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Conclusion • • • In practical environment, major link setup latency is brought by DHCP. Optimistic and Aggressive DHCP implementation may reduce most of DHCP latency. But it’s unrecommended procedure according to the protocol specification. Another configuration procedure which is optimized for wireless network should be considered. • • • Major airtime occupancy is brought by authentication phase. Most of airtime is consumed by overheads (IFSs, CW, preamble…). Reducing number of frames is effective. • DS consumes much more airtime than OFDM. Especially this is caused by long overhead. To quit using DS is effective. • Submission Slide 13 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Future Work • Evaluate the performance of the proposed protocols to help the TG decision. Latency proposal C proposal B proposal A Airtime Submission Slide 14 Hitoshi Morioka, Allied Telesis R&D Center

January 2012 doc. : IEEE 802. 11 -12/0034 r 0 Questions & Comments Submission Slide 15 Hitoshi Morioka, Allied Telesis R&D Center