Measurement-based, Practical Techniques to Improve 802.11ac - PowerPoint PPT Presentation

Measurement-based, Practical Techniques to Improve 802.11ac Performance Apurv Bhartia, Bo Chen , Feng Wang, Derrick Pallas, Raluca Musaloiu-E, Ted Tsung-Te Lai, Hao Ma ACM Internet Measurement Conference 2017 Nov 2017

Why is this talk important? • Wireless networks have been undergoing changing o more capacity, higher speed, denser networks o new 802.11 standards have been introduced • New techniques are required o take advantage of the wireless evoluCon What can we do about it? 1

Agenda • Large scale wireless measurement study o shows changing trend in today’s wireless networks • TurboCA - A new channel planning algorithm o reduces wireless TCP latency by up to 40% • FastACK – A TCP over wireless enhancement o increases wireless throughput by up to 38% 2

Agenda • Large scale wireless measurement study o shows changing trend in today’s wireless networks • TurboCA - A new channel planning algorithm o reduces wireless TCP latency by up to 40% • FastACK – A TCP over wireless enhancement o increases wireless throughput by up to 38% 2

What is 802.11ac? • A wireless standard published in Dec 2013 o wider channel width (up to 160 MHz) 100 2015 2017 % of client devices o more antenna usage (4x4 MIMO) o higher data rate (up to 3.39 Gbps) 50 • Rapid adopJon of 802.11ac o default in new wireless devices o more than 50% of Meraki APs 0 802.11 abg only 802.11n 802.11ac o 45.7% of the clients devices 27.7% more clients support 802.11ac 3

Measurement Framework

Coud Management More than one Million Networks 4


 
 100K active APs 
 1.7M associated clients 
 50G packets

Client Device Trends (2015 – 2017) 2017 2015* Band Channel Width % of client devices % of client devices 100 100 50 50 0 0 2.4GHz only 2.4GHz + 5GHz 20MHz only 40MHz 80MHz 45.7% supporCng 80MHz now Significant usage of 2.4GHz devices MulJple Antennas (MIMO) 100 % of client devices 50 0 No MIMO 2x2 MIMO 3x3 MIMO 4x4 MIMO MulC-antenna devices gaining tracCon * Large-scale Measurements of Wireless Network Behavior, SIGCOMM 2015 5

Spectrum Trend 5GHz 2.4GHz 5 GHz 2.4 GHz 1 1 0.8 0.8 0.6 0.6 CDF CDF 1.7% 15% 5 7 0.4 0.4 0.2 0.2 0 0 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 Channel uJlizaJon Number of interferers Median # of interferers is similar for 2.4GHz is much more congested than 5GHz both bands 6

Traffic Analysis 0.4 DistribuJon of Packets 0.3 0.2 0.1 0 0 2 8 32 128 256 512 1024 0 2 8 32 128 256 512 1024 Bit rate (Mbps) DistribuCon matches types of clients seen in the field 7


 More and more 802.11ac devices 
 Wireless spectrum getting more crowded 


Agenda • Large scale wireless measurement study o shows changing trend in today’s wireless networks • TurboCA - A new channel planning algorithm o reduces wireless TCP latency by up to 40% • FastACK – A TCP over wireless enhancement o increases wireless throughput by up to 38% 8

TurboCA – Motivation 802.11ac introduces extended channel width • o channel width is manually configured lower with 35% of the Meraki 802.11ac APs. 802.11ac makes RTS/CTS enforced • o RTS/CTS changes the behavior of neighbor APs on the same channel 9

TurboCA – Motivation One day Snapshot of an AP at Meraki HQ 20 100 8 UJlizaJon (%) Usage (Mbps) AcJve Clients 10 50 4 0 0 0 0:00 8:00 16:00 24:00 0:00 8:00 16:00 24:00 0 480 960 1440 0 480 960 1440 0:00 8:00 16:00 24:00 0 480 960 1440 • Traffic trend indicates frequent channel switch • Channel switch disrupts traffic flow 10

TurboCA – System • Carrier sensing based channel model • User-experience based opJmizaJon AP load o AP load o channel quality channel switch Channel quality o channel switch Metric of network Metric of single AP • Aggressive channel re-evaluaJon to catch up with environment change • Details in the paper 11

TurboCA – Evaluation • Test networks o museum (171 of 201 APs supporCng 802.11ac) o university (653 of 735 APs supporCng 802.11ac) • Comparison o baseline (default before Nov. 2016) o TurboCA (default ader Nov. 2016) • DuraJon o 03/25/17 -- 05/07/17 Real-world deployment 12

TurboCA – Evaluation • Usage (throughput) Baseline TurboCA University Usage Museum Usage 600 80 Usage(GB) Usage(GB) 60 400 40 200 20 0 0 Per Hour Usage Peak Hour Usage Per Hour Usage Peak Hour Usage Uplink saturated for the university from ISP constraint TurboCA improves peak hour usage by 27% in the museum 13

TurboCA – Evaluation • TCP latency 1 Baseline TurboCA 0.8 0.6 CDF 40% lower 0.4 0.2 31.1ms 18.6ms 0 1 10 100 1000 Wireless TCP Latency (ms) TurboCA reduces the TCP latency by 40% 14


 TurboCA 
 Better channel assignment

Agenda • Large scale wireless measurement study o shows changing trend in today’s wireless networks • TurboCA - A new channel planning algorithm o reduces wireless TCP latency by up to 40% • FastACK – A TCP over wireless enhancement o increases wireless throughput by up to 38% 15

FastACK– Motivation • TCP is most popular protocol for transmission • But TCP proposed for wired networks not aware of the variable latency introduced in wireless o not aware of the aggregaCon used in 802.11ac o TCP sender TCP receiver AP TCP packet 802.11 ACK TCP ACK TCP ACK TCP ACK Key Insight 802.11 latency << TCP latency strong hint for correct recepCon 16

` FastACK – System • Exploit 802.11 ACK hints for larger aggregates • TCP end-to-end semanJcs maintained • No client modificaJon TCP sender TCP receiver AP TCP packet TCP ACKs 802.11 ACK TCP ACK TCP ACK TCP ACK • Details in the paper 17

FastACK – Evaluation • Testbed – 40, 3x3 clients • Ixchariot is used to measure the throughput Switch 1 TCP sender 1 TCP sender 2 Switch 2 30 feet AP2 AP1 40 wireless clients 85 feet 18

FastACK – Evaluation • Single-AP mulJple clients test 600 Aggregate Throughput (Mbps) Baseline FastACK 450 300 150 5 10 20 30 40 Number of Clients Up to 38% aggregate throughput improvements 19

FastACK – Evaluation • MulJ-AP mulJple clients test AP1 AP2 Aggregate Throughput (Mbps) 395 Mbps 400 325 Mbps 300 251 Mbps 200 100 0 Baseline Baseline FastACK FastACK 1,2 2 1,2 1 Number of Clients FastACK does not suffer in isolaCon Combined performance is 60% beker 20


 FastACK 
 Improved TCP over wireless

s Related Work • Wireless network measurement Vivek, S [NSDI 11], Ratul M [SIGCOMM 06], Yu-Chung, C [SIGCOMM 06], Sanjit, B [SIGCOMM 15] The largest scale of study for wireless and first evaluaJon on 802.11ac. • Channel assignment Apurv. B[MOBIHOC 16], Paramvir, B[SIGCOMM 09], Shravan R [MOBICOM 09] TurboCA considers real enterprise challenges. • TCP enhancement A. Bakre [ICDCS 95], Ajav V [MLICS 95], Hari B [MOBICOM 95], Stefan S [SIGCOMM 99] FastACK looks at the impact of 802.11ac aggregaJon for TCP. 21

Conclusion • We observe interesJng trends in today’s wireless networks o rapid adopCon of 802.11ac-enabled devices o network densificaCon conCnues to increase o wireless spectrum geqng more crowded • Based on observed trends, we present two fundamental techniques o TurboCA, a new channel assignment algorithm o FastACK, a TCP enhancing technique over 802.11ac • New techniques show significant potenJal in enterprise networks 22

Q & A