Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [QoS Support in Wireless BANs] Date Submitted: [November, 2008] Source: [Feng Shu and Guido Dolmans] Company: [Holst Centre / IMEC-NL] Address [High Tech Campus 31, Eindhoven, the Netherlands] Voice:[+31 40 2774382], FAX: [+44 40 2746400], E-Mail:[{feng.shu, guido.dolmans}@imec-nl.nl] Abstract: [Presentation for QoS support in wireless BANs.] Purpose: [To discuss QoS support issues for BAN] Notice: This document has been prepared to assist the IEEE P802.15. It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein. Release: The contributor acknowledges and accepts that this contribution becomes the property of IEEE and may be made publicly available by P802.15.

QoS Support in Wireless BANs Feng Shu and Guido Dolmans IMEC-NL November, 2008, Dallas

QoS support in wireless BANs Outline QoS support in wireless BANs QoS differentiation – gives priority to critical nodes (two examples based on CSMA-CA protocol) latency differentiation scheme (LDS) reliability differentiation scheme (RDS) Simulation and evaluation

QoS support in wireless BANs As technical requirement document (TRD) specifies, major QoS for wireless BANs include Latency Reliability When critical & non-critical app’s co-exist, give priorities to critical ones QoS differentiation mechanisms needed Reliability def.: BER, packet loss, packet success rate Critical Non-critical Reliability Latency Two examples of QoS differentiation based on CSMA-CA

Traffic Class Definition Type  : traffic has stricter QoS of latency and/or reliability usually less amount of traffic treated with higher priority Type  : traffic has less strict QoS of latency and/or reliability usually more amount of traffic treated with lower priority

Energy Conserving CSMA-CA Protocol W0 = 8 W1 = 16 4 3 2 1 11 10 9 8 7 6 5 4 3 2 1 packet transmission Type  /  nodes: W0 / W0 : initial backoff windows M / M : maximum # of attempts in a single backoff process R / R : maximum # of collisions before a packet is aborted

Latency Differentiation Scheme (LDS) LDS – prioritize by adjusting backoff window sizes Type  nodes start with smaller initial backoff window W0 ≤ W0 In case of collision, type  nodes restart with maximum backoff window Wmax to reduce contention level

Reliability Differentiation Scheme (RDS) RDS – prioritize by giving more transmission attempts Type  nodes are given more backoff attempts when channel is sensed busy M > M In case of collisions, type  nodes are given more opportunities of retransmissions R > R

Performance Evaluation Implemented a C++ based simulator PHY layer: IEEE 802.15.4, 2.4 GHz Node generate packets based on Poisson arrivals, let  indicate the average packet arrival rate packet length L = 60 bytes, ie, 6 backoff slots # of type  nodes: three # of type  nodes: variable Backoff parameters are taken from default values of 802.15.4 standard Simulations run 100,000 time slots, and each point is averaged over 10 independent simulation runs

Latency – LDS When initial backoff window is smaller, latency is lower  = 28.5 Critical nodes have much lower average latency When initial backoff window is smaller, latency is lower

Packet Success Rate - LDS W0 = 16  = 28.5 LDS does not give priority to critical nodes for transmission reliability

Packet Success Rate – RDS & LDS M = 8, M = 4 R = 6, R = 3 W0=4, W0 =16  = 12.5 When LDS & RDS implemented, reliability of critical nodes is increased

Latency – RDS & LDS M = 8, M = 4 R = 6, R = 3 W0=4, W0 =16  = 12.5 Critical nodes can have larger latency when traffic load gets heavy This reveals an intrinsic trade-off between latency & reliability

Conclusions Major QoS of wireless BANs are latency & reliability QoS differentiation needed when critical & non-critical app’s competing for resources Latency differentiation scheme (LDS) Reliability differentiation scheme (RDS) It reveals that an intrinsic trade-off exists between requirements such as latency & reliability