doc IEEE 802 11 -13 0872 r 0

doc. : IEEE 802. 11 -13/0872 r 0 June 2013 Clarification on outdoor deployments Authors: Submission Date: 2013 -07 -16 Slide 1 Laurent Cariou (Orange)

doc. : IEEE 802. 11 -13/0872 r 0 Cellular network outdoor deployment strategies (1/2) • 2 G-3 G cellular network deployments had the objective of a full blanket coverage – operator have deployed base station in what is called “macro” sites. The cells are called macrocells. – Such sites are usually huge tours/masts in rural areas and at rooftop or 15 m above rooftop in cities. – The backhaul is available and the site (expensive) is acquired by operators for cellular networks • First deployments of 4 G have the same blanket coverage objective and macro sites are deployed in priority. – Transmit power is between 20 and 60 W – Inter-site distance (ISD) is 1950 m in rural areas and 500 m in dense urban – usually 3 sectors per site Submission

doc. : IEEE 802. 11 -13/0872 r 0 Cellular network outdoor deployment strategies (2/2) • 3 G-4 G small cells are used to cover localized zones of high density (which require increased capacity) or badly covered by macrocells. The objective is no longer for blanket coverage – small cells incorporate microcells (10 W) and picocells (1 -2 W) – they are deployed under rooftop (between 3 to 10 m): lamp poles, hanged on cables, stuck to walls… – site acquisition is less expensive than macro sites but is still with the backhaul the most important Capex/Opex. • 3 GPP defines multiple scenarios for pico deployments – rural (10 picos per macro sector) – dense urban (4 picos per macro sector) • • Submission with 500 m macro ISD and 3 sectors per macro: density of 48 picos/Km²

doc. : IEEE 802. 11 -13/0872 r 0 Wi-Fi outdoor deployment strategies • • With its limited transmit power, Wi-Fi is by nature a small cell technology. Wi-Fi will therefore be deployed in priority in zones which require high capacity (i. e. with high density of STAs) and will follow the same deployment strategy as 3 G-4 G small cells – • • Early operator deployments should be Wi-Fi-only but it is likely that deployments will evolve toward Wi-Fi co-located with 3 G-4 G small cells. deployed under rooftop (between 3 to 10 m): lamp poles, hanged on cables, stuck to walls… side coverage (omni or directional) and overhead coverage (downtilt omni) Submission

doc. : IEEE 802. 11 -13/0872 r 0 Wi-Fi outdoor deployment strategies • Micro/macro deployments: – For some reasons (site fees, low heights obstruction, power and network availability), APs can be placed at rooftop or even above • it can therefore be co-localized with macro sites but it should be less frequent than pico deployments as macro sites are not specificaly placed in zones of high density and as the range difference between LTE and Wi-Fi would be too high Submission

doc. : IEEE 802. 11 -13/0872 r 0 Outdoor AP densities and channel model • For these reasons, – The lowest density of outdoor APs should be around 50 APs/Km² for pico deployments • maximum seperation between APs should be 160 -200 m AP AP STA – higher densities will be common of course but as the backhaul and site rental are expensive for such deployments, there should be as low sites as possible. Consequence: • neighboring cells should overlap close to minimum sensitivity (high proportion of BSS-edge users with low SNR/SINR, especially in uplink) • APs should use maximum Tx power • question regarding the frequency reuse for one operator deployment: should frequency reuse 1 be favorable? • If there were one channel model to select for outdoor deployments, we should clearly favor ITU UMi over ITU UMa. Submission

doc. : IEEE 802. 11 -13/0872 r 0 Picocell interference characteristics • It is very likely that in outdoor dense environments, multiple operators will make their own network deployment – even more true in case of co-location with 3 G-4 G small cells • There will also be a high proportion of interfering private APs in the coverage of outdoor hotspots – indoor home or shop private APs leaking outdoors (usually in hidden node situation) – some of these APs can have community Wi-Fi (one private SSID and one public SSID) Submission

June 2013 doc. : IEEE 802. 11 -13/0872 r 0 Traffic type on hotspots: focus on video • In hotspot scenario, highly compressed video will be the most common • Compressed video throughput evolution takes into account: – New resolution, new coding/compression schemes, throughput increase in the backhaul or ADSL/VDSL/fiber network, availabilities of content in such format… • Here are some prediction numbers that we got from our video experts: – Currently Orange HD requires 6. 5 Mbps for IPTV (15 Mbps for DVB broadcast, 7 Mbps for Netflix) – Evolution with HEVC: • With new versions of H 264, this IPTV will move to 5 or 5. 5 Mbps. • With HEVC, we will gain between 40 to 50%, so a HD flow will be around 2. 75 Mbps. – Some 4 k will be 3 -4 times this, so there could be VHD flows as low as a bit more than 10 Mbps. • this is probably the lower bound and the average should be around 20 Mbps. This will depend on the average available bandwidth in different networks. – predictions that 4 k will start market around 2018 and will reach mass market around 2023, including the availability of content (predicted to be slightly slower than HD penetration). Submission Slide 8

doc. : IEEE 802. 11 -13/0872 r 0 Conclusion • Most outdoor deployments should be based on pico-sites – below rooftop (3 to 10 m high) • The lowest density of outdoor APs should be around 50 APs/Km² – maximum ISD between APs should be 160 -200 m – can be smaller than that obviously in denser networks • We believe ITU UMi is more fitted than ITU UMa for outdoor deployments Submission