Rearchitecting Wireless Networks with PHY Layer Components

1. Rearchitecting Wireless Networkswith PHY Layer Components Romit Roy Choudhury Assistant Professor

2. A little bit about ourselves

3. Webpage http://synrg.ee.duke.edu

4. Our Research Mobile Computing (top down) Collaborative Sensing Interfaces Localization Application Location Privacy Energy Security Transport Network Mobility Home networks MAC / Link Interference Mgmt. Rate Control PHY Software Radios Smart Antennas Wireless Networking (bottom up)

5. Mobile Computing Information Telescope Mobile Phones for Collaborative Sensing Location Sensing Physical and Logical Localization MobiSys 08 MobiCom 09 Infocom 09, 10 MobiCom 10 Micro-mobility Gesture and activity recognition Smart Content Context-aware content and compression MobiHeld 09 MobiSys 10 Hotmobile 11

6. Wireless Networking MobiCom 09, 10 Hotnets 09, 10 NSDI 10 MobiCom 09 Hotnets 08 Wired + Wireless Infrastructure Assisted Wireless Cross-Layer PHY Informed Protocol Design Out of Band Sensor Assisted Wireless Networking SleepWell WiFi Energy Management LANMAN 10 In Submission

7. Today’s Talk Context Cross-Layer Systems Time to Frequency AccuRate CSMA/CN MobiCom 09, 10 Hotnets 09, 10 NSDI 10 Mobile Computing Virtual Telescope Location PhonePen Cross-Layer PHY Informed Protocol Design Closing Thoughts

8. Context

9. Wireless Everywhere

10. Skyrocketing Demands FCC looking for 500 MHz spectrum by 2020 … But also calling for much better use of available spectrum • Wireless usage increased by 25x in last 5 years • Cisco predicts 40x increase by 2013 • Network outages a reality • Major carriers forcing customers to pay-per-byte

11. Capacity vs. Goodput • Problem is not of spectrum alone • Under utilization of available spectrum a major problem • Significant leaps in achievable PHY capacity • MIMO, OFDM, Coding, Beamforming … • Yet, this PHY capacity not visible to higher layers • Inefficiencies in network design … • protocols … architecture Link Throughput PHY Bitrate

12. Layering too Restrictive? Throughput Capacity • The capacity-throughput gap is not new • Researchers recognized need to share information across layers • Cross layer approaches became popular • Cross layer optimization • Several creative ideas … many analyzed and simulated

13. Layering too Restrictive? Throughput • Lack of experimentation platform •  difficult to build practical working systems • Protocol designers untrained in communications •  cross layer ideas variants of originals •  uses some PHY layer info. Capacity • The capacity-throughput gap is not new • Researchers recognized need to share information across layers • Cross layer approaches became popular • Cross layer optimization • Several creative ideas … many analyzed and simulated • However, 2 deficiencies

14. Software Radios Full view of PHY layer enabling experimentation with holistic, unconventional ideas … We intend to contribute here • Software defined radios • Changing landscape of wireless systems • Protocol designers understanding PHY concepts, using them • PHY community receiving feedback from practical systems

15. Our Goal: Rearchitect wireless networks with full access to PHY layer capabilities

16. We instantiate our ideas through WiFi However, the core ideas not specific to WiFi … choice of WiFi mainly from platform considerations

17. WiFi Protocol Structure

18. WiFi Structure Packet for R1 Packet for R2 AP1 AP2 R2 R1

19. WiFi Structure Random Backoff = 10 Random Backoff = 18 AP1 AP2 R2 R1

20. WiFi Structure Random Backoff = 10 Random Backoff = 18 AP1 AP2 R2 R1 AP1 = 10 Time AP2 = 18

21. WiFi Structure Remaining Backoff = 0 Remaining Backoff = 8 AP1 AP2 R2 R1 AP1 = 0 Time AP2 = 8

22. WiFi Structure Transmit @ rate = r1 Channel Busy AP1 AP2 R2 R1 AP1 = 0 Data ACK AP2 Waits Time AP2 = 8

23. WiFi Structure New Backoff = 15 Remaining Backoff = 8 AP1 AP2 R2 R1 AP1 = 15 Data ACK AP2 Waits AP2 = 8

24. WiFi Structure Remaining Backoff = 7 Remaining Backoff = 0 AP1 AP2 R2 R1 AP1 = 7 Data ACK AP2 Waits AP2 = 0

25. WiFi Structure Transmit @ rate r2 Channel Busy AP1 AP2 R2 R1 AP1 = 7 Data ACK AP1 Waits Data ACK AP2 Waits AP2 = 0

26. WiFi Structure Channel Busy Transmit AP1 AP2 R2 R1 Data Data ACK AP1 Waits Data ACK AP2 Waits

27. WiFi Structure ACK Not Received Channel Busy AP1 AP2 R2 R1 Data ACK AP1 Waits ✘ AP2 Waits Data Interference

28. WiFi Structure Adjust Rate & Retransmit Channel Busy AP1 AP2 R2 R1 Data ACK AP1 Waits ✘ AP2 Waits Data Data Interference

29. WiFi Structure Adjust Rate & Retransmit Channel Busy AP1 AP2 R2 R1 Data ACK AP1 Waits ✘ AP2 Waits Data Data Interference

30. WiFi Structure Adjust Rate & Retransmit Channel Busy AP1 AP2 R2 R1 Channel Wastage Data ACK AP1 Waits ✘ AP2 Waits Data Data Interference

31. WiFi Structure Adjust Rate & Retransmit Channel Busy AP1 AP2 R2 R1 Channel Wastage Data ACK AP1 Waits ✘ AP2 Waits Data Data Interference Collision or Fading

32. WiFi Structure Adjust Rate & Retransmit Channel Busy AP1 AP2 R2 R1 Channel Wastage Heuristic Rate Selection Data ACK AP1 Waits ✘ AP2 Waits Data Data Interference Collision or Fading

33. WiFi Structure Adjust Rate & Retransmit Channel Busy AP1 AP2 R2 R1 Channel Wastage Heuristic Rate Selection Redundancy Data ACK AP1 Waits ✘ AP2 Waits Data Data Interference Collision or Fading

34. Channel Wastage Heuristic Rate Selection Redundancy

35. PHY Layer Information (OFDM, Constellation, Interference Cancellation, Correlation …) Channel Wastage Heuristic Rate Selection Redundancy

36. PHY Layer Information (OFDM, Constellation, Interference Cancellation, Correlation …) Channel Wastage Heuristic Rate Selection Redundancy Software Radios (USRP, WARP, SoRa)

37. PHY Layer Information (OFDM, Constellation, Interference Cancellation, Correlation …) Channel Wastage Heuristic Rate Selection Redundancy Software Radios (USRP, WARP, SoRa) Cross-Layered Network Systems

38. Channel Wastage due to Randomized Backing off

39. Backoff Per packet backoff  35% of channel wastage at 54 Mbps. Worse at higher data rates. Data Data ACK AP1 Waits Data ACK AP2 Waits

40. Fundamentally, backoff is not a time domain operation … its implementation has been in the time domain

41. Fundamentally, backoff is not a time domain operation … its implementation has been in the time domain We intend to break away, and implement backoff on the frequency domain

42. Frequency Domain Subcarriers: 1 2 3 4 … 48 Frequency • 802.11a/g PHY adopts OFDM • Wideband channel divided into 48 narrow sub-carriers • Copes better with fast, frequency selective fading • Purely a PHY layer motivation • MAC Opportunity • Pretend OFDM subcarriers are integers • Emulate randomized backoff

43. 18 6 0 47 0 47 T2F: Main Idea • Pick random backoff, say 6 • Submit signal on 6thsubcarrier AP2 Backoff = 18 AP1 Backoff = 6

44. 18 6 Listen Antenna Listen Antenna 6 18 6 18 0 47 0 47 T2F: Main Idea • Pick random backoff, say 6 • Submit signal on 6thsubcarrier AP2 Backoff = 18 AP1 Backoff = 6 AP2 learns some other AP is winner. AP1 learns AP1 is the winner … hence, AP1 transmits

45. 0 1 2 3 4 5 Subcarrier Second Round What if Collision? • Introduce a second round of contention • Winners of first go to second 0 1 2 3 4 5 Subcarrier First Round Winner

46. Possible to do better … Why beneficial? Avg. temporal backoff = 16 slots = 144 micro sec. Frequency backoff = 1 OFDM symbol = 4 micro sec 2 rounds of backoff = 8 micro sec.

47. 0 1 2 3 4 5 Subcarrier Second Round Creating a Queue 0 1 2 3 4 5 Subcarrier First Round Rank 2 Winner

48. 0 1 2 3 4 5 Subcarrier Second Round Creating a Queue 0 1 2 3 4 5 Subcarrier First Round 0 2 4 Rank 1 Rank 2 Rank 3 Enabling TDMA 0 2 4 0 2 4

49. Improved Channel Utilization Data Data Data Data WiFi: Contention per packet TDMA Data Data Data Data T2F: OFDM contention per TDMA schedule

50. Multiple Collision Domains What happens with real-world scatterred APs