Crowdsourcing Access Network Spectrum Allocation Using Smartphones - PowerPoint PPT Presentation

pocketsniffer@blue.cse.buffalo.edu Crowdsourcing Access Network Spectrum Allocation Using Smartphones Jinghao Shi † , Zhangyu Guan †§ , Chunming Qiao † Tommaso Melodia § , Dimitrios Koutsonikolas † and Geoffrey Challen † † University at Buffalo § Northeastern University HotNets'14, October 28, 2014

Wireless Spectrum Management… . …is hard. Crowded clients. User mobility. Hidden terminals. … 1 / 14

Client-Side Measurements… . …are valuable. 123 Hidden Interference Interference Yet collecting them is disruptive . 1 Arunesh Mishra, Vladimir Brik, et al. ``A Client-Driven Approach for Channel Management in Wireless LANs''. In: Proc. of the 25th IEEE International Conference on Computer Communications (INFOCOM 2006). Apr. 2006 2 Arunesh Mishra, Suman Banerjee, and William Arbaugh. ``Weighted coloring based channel assignment for WLANs''. In: ACM SIGMOBILE Mobile Computing and Communications Review 9.3 (July 2005), pp. 19–31 3 Shravan Rayanchu et al. ``FLUID: improving throughputs in enterprise wireless lans through flexible channelization''. In: Proc. of the 17th annual international conference on Mobile computing and networking (MobiCom 2011). Sept. 2011 2 / 14

An Example . Wi-Fi is slow. Can you do something? . 3 / 14

An Example . Sure, please scan channel X, Y, Z. . 4 / 14

An Example . “Thank you” for the broken link. . A slow link is still better that a broken link. 5 / 14

What if… . I can help. . 6 / 14

Smartphone to the Rescue . I can help. . Carried with you. Always on. Mostly idle. 7 / 14

PocketSniffer: A Framework For… . …client-side measurements collection. PocketSniffer Server PocketSniffer AP Query Active session Health Spectrum monitoring measurement Proximity detection Idle PocketSniffer Client Active Client Spectrum utilization. Wi-Fi scan results. 8 / 14

Q1: Measurement Approximation . Sure, tell me your condition of channel X, Y, Z. What we want What we get Physical proximity. Wi-Fi signature similarity. 9 / 14

Q2: Incentivize Measurement . Incentives: Bob's phone help Bob's laptop. Trade measurements for better QoS. 10 / 14

Q3: Measurement Efficiency . ...which one to use? Battery Level Proximity Network performance improvement per Joule? Measurement? … 11 / 14

Q4: Measurement Validation . Umm... are you serious? Laziness. Leverage controlled APs. Selfishness. Cross validation. 12 / 14

Big Picture & Open Questions . . Global information? Cooperative networks? Game theory? Wifi-cellular interaction? 13 / 14

Summary . Use smartphones to help PocketSniffer Server PocketSniffer AP other devices. Approximation. Query Active Incentive. session Health Spectrum monitoring measurement Efficiency. Proximity detection Validation Idle PocketSniffer Client Coordination. Active Client 14 / 14

References I . Arunesh Mishra, Suman Banerjee, and William Arbaugh. ``Weighted coloring based channel assignment for WLANs''. In: ACM SIGMOBILE Mobile Computing and Communications Review 9.3 (July 2005), pp. 19–31. Arunesh Mishra, Vladimir Brik, et al. ``A Client-Driven Approach for Channel Management in Wireless LANs''. In: Proc. of the 25th IEEE International Conference on Computer Communications (INFOCOM 2006). Apr. 2006. Shravan Rayanchu et al. ``FLUID: improving throughputs in enterprise wireless lans through flexible channelization''. In: Proc. of the 17th annual international conference on Mobile computing and networking (MobiCom 2011). Sept. 2011. 1 / 1