doc.: IEEE /0882r1 Submission DSC leveraging uplink RTS/CTS control Authors: July 2015

doc.: IEEE /0882r1 Submission Outline July 2015 M. Shahwaiz Afaqui (UPC)Slide 2 1.Context 2.Simulation Environment: NS-3 3.Simulation scenario and assumptions 4.RTS/CTS threshold setting 5.Impact of frame size on DSC+RTS 6.Conclusions 7.References Appendix

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide 3 IEEE ax enabled devices are expected to maintain and reduce energy consumed per successful information bit. However, different amendments that are proposed to increase the efficient operation of PHY and MAC layer would work against the aforementioned requirement. Since interference management is of paramount importance in dense deployments, IEEE ax aspires to intelligently utilize RTS/CTS method based on observed channel conditions on per node basis. In [1], the authors highlight the possible mechanism through which an AP can control the RTS/CTS policy for the associated stations. It has been shown in our previous presentations [2-3] that the use of DSC can increase per-user throughput in dense scenarios over the cost of increase in Frame Error Rate (FER) due to increased number of hidden nodes. In this submission we, – investigate the performance of DSC, By utilizing intelligent RTS/CTS control method as means to mitigate the drawbacks associated with DSC and to improve performance gains. – recommend method to select threshold to enable RTS/CTS on each node. – study the impact of frame size on RTS enabled DSC stations. 1. Context

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide 4 NS-3 –Allows the study of protocols and network performance of large-scale systems in a controlled and scalable environment. Main characteristics, –Discrete event simulator –Packet level simulator (layer 2 and above) –Layered architecture –Free and open source –Frequent updates ( latest version ns release date ) Large number of protocol implementations and models available, –TCP, UDP –IPV4, IPV6, static routing –IEEE and variants, WiMAX, LTE –IEEE 802 physical layer –Mobility models and routing protocols –Ability to design indoor, outdoor or hybrid networks –etc. 2. Simulation Environment: NS-3

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide 5 2. Simulation Environment: NS-3 Limitations of NS3 –Simplified abstraction of PHY layer –MPDU aggregation is not yet mature and thus not used within these simulations. –IEEE ac model has not yet been developed and current results focus on IEEE n.

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide 6 3. Simulation scenarios and assumptions Topology –multi-floor residential building, 5 stories 2×10 apartments per story. Apartment size: 10m×10m×3m. –1 AP placed randomly in each apartment at 1.5m height. –channel selected randomly for each cell. Three channel scheme (1, 6, 11)  1/3 of the cells share the same channel –5 STAs placed randomly around their respective AP.

doc.: IEEE /0882r1 Submission July2015 M. Shahwaiz Afaqui (UPC)Slide 7 Frequency band: focused on 2.4GHz, Intended to investigate the impact of DSC in a band that is more restricted in dense environments. Traffic: UDP CBR uplink transmission in saturation conditions is considered, Worst case in terms of contention. Pathloss model: Hybrid Building Propagation loss model (adapted to fit TGax channel model [3]) We simulated specific scenarios (with same STA and AP positions) with and without DSC and RTS/CTS control method. Additional simulation details are provided in the appendix. 3. Simulation scenarios and assumptions

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide 8 4. RTS/CTS threshold setting FER of nodes used to enable and disable RTS/CTS control from the AP. –i.e. if FER of a node is greater than RTSThrehold → enable RTS/CTS for that node. –Formula: RTSThreshold = N x AvgFER Improvement over DSC-enabled network for longest frame (MCS0 with maximum MSDU size of 2302bytes) : ~14% improvement in throughput over DSC when RTSThrehold = 0.6 x AvgFER –N = 0.6 used in all the following experiments.

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide 9 5. Impact of frame size on DSC+RTS (1/2) From previous slide, intelligent use uplink RTS/CTS control can have multifold benefits. Tested with long frames (i.e. 1000, 1600 and 2302Bytes at MCS0) for comparative evaluation with RTS enabled DSC nodes (RTSDSC) with, a.IEEE802.11n network with RTS disabled DSC nodes (NORTSDSC) b.IEEE802.11n network with RTS disabled non-DSC nodes (NORTSNODSC) Up to 55% improvement in throughput (longest frame duration). RTS control adds to benefits of DSC. Maximum FER improvements are achieved for small frame size.

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide Impact of frame size on DSC+RTS (2/2) Tested with shorter frames (2302Bytes at MCS7) for comparative evaluation with RTS enabled DSC nodes (RTSDSC) with, a.IEEE802.11n network with RTS disabled DSC nodes (NORTSDSC) b.IEEE802.11n network with RTS disabled non-DSC nodes (NORTSNODSC) For short frames, RTS/CTS not paying off (in terms of throughput). Better off DSC on its own.

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide Conclusions More efficient DSC is achieved when combined with uplink RTS/CTS control, as proposed in [1]. –One of the major drawbacks of DSC scheme was the increase in FER due to hidden nodes  RTS enabled DSC network can mitigate the aforementioned problem. We propose to intelligently enable RTS/CTS, –The AP knows uplink FER of all associated STAs –AP decides which STA uses RTS/CTS based on: Threshold FER derived from average FER (RTSThreshold = 1.4 x AvgFER for long frames) Frame size (excessive overhead for small packets)

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide References [1] Sigurd Schelstraete, IEEE , Uplink RTS/CTS Control. [2] M. Shahwaiz Afaqui, IEEE , Simulation-based evaluation of DSC in residential scenario. [3] M. Shahwaiz Afaqui, IEEE , Proposal and simulation based evaluation of DSC-AP Algorithm [4] Jianhan Liu et al., IEEE r4, TGax Channel Model Document

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide 13 Simulation assumptions PHY parameters Parameters4ValuesParametersValues Wireless StandardIEEE nMSDU size Bytes Frequency band2.4 GHzSTA TX power16dBm Physical transmission rate for IEEE n i.7.2 Mbps ii.72.2 Mbps Transmission gain1 dB Channel width20 MHzReception gain1 dB Propagation delay model Constant speed propagation delay Noise figure7 dB Propagation loss modelHybrid buildings propagation loss Rx Wall penetration loss12 dBInitial CCA threshold-80dBm Floor penetration loss17 dBGuard intervalShort Data preambleShort

doc.: IEEE /0882r1 Submission July 2015 M. Shahwaiz Afaqui (UPC)Slide 14 MAC parameters ParametersValuesParametersValues Access protocolEDCARetransmission attempts16 RTS/CTSDisabled/uplink controlMaximum missed beacons for re-association Association100% STAs associated to AP in an Apartment Active probingDisabled QOSEnabledTraffic modelBest effort AggregationDisabled Simulation parameters ParametersValuesParametersValues Simulation time25 secondsNumber of simulations per case 25 Confidence interval95%