Author: Giuseppe Bianchi EEL6935: Wireless Ad Hoc Networks Performance Analysis of IEEE802.11 Distributed Coordination Function (DCF) -- JSAC’2000 Author: Giuseppe Bianchi Presented by: Rongsheng Huang

Outline Overview of IEEE 802.11 DCF Mathematical model EEL6935: Wireless Ad Hoc Networks Outline Overview of IEEE 802.11 DCF Mathematical model Notations and the main idea One step transition probabilities Bi-dimensional Markov Chain Stationary distribution Maximum throughput Performance evaluation of DCF Concluding remarks

EEL6935: Wireless Ad Hoc Networks Overview of IEEE 802.11 MAC and PHY layers specifications for wireless LANs MAC Protocols Fundamental: Distributed Coordination Function (DCF) -- CSMA/CA based Binary Exponential Backoff rules Optional: Point Coordination Function (PCF)

Overview of IEEE 802.11 DCF Two access techniques EEL6935: Wireless Ad Hoc Networks Overview of IEEE 802.11 DCF Two access techniques Basic mechanism: 2 way handshaking RTS/CTS mechanism: 4 way handshaking RTS DATA CTS DESt Source Source Dest DATA ACK ACK

Example of RTS/CTS Access Scheme EEL6935: Wireless Ad Hoc Networks Example of RTS/CTS Access Scheme CSMA/CA based CSMA: listen at least DIFS before talk CA: defer transmission for random back-off time after DIFS SIFS BO=3 (set) BO=0 BO=5 (set) A BO=15 (set) DIFS RTS DATA DIFS RTS DIFS collision CTS ACK B BO=8 (set) BO=5(resume) BO=10 (set) BO=5 (suspend) BUSY DIFS NAV (RTS) DIFS RTS DIFS Others NAV(CTS)

Backoff procedure—BEB algorithm EEL6935: Wireless Ad Hoc Networks Overview of IEEE 802.11 DCF Backoff procedure—BEB algorithm CW Backoff counter: Initial: uni~[0,CW-1] Non zero: decremented for each idle slot Zero: transmit CWmax c c s s c c c CWmin c c C c s t

Analytical Model to Evaluate DCF EEL6935: Wireless Ad Hoc Networks Analytical Model to Evaluate DCF Goal: saturation throughput & Maximum achievable saturation throughput: Saturation throughput: given all the system parameters, assume each station always has packets to send, the stable throughput of the whole system Maximum throughput: if the system parameters can be modified, the theoretical maximum saturation throughput achievable. Basic assumptions Constant & independent collision probability for each transmitted packet p (suppose that the backoff mechanism can adjust the p to a constant value) Ideal channel condition (no hidden terminals and capture) Fixed number of stations n operated under overload (saturation condition)

Idea of Applying Markov Chain EEL6935: Wireless Ad Hoc Networks Idea of Applying Markov Chain Transmission probability and conditional collision probability p are needed. Normalized system throughput S with n stations can be derived. Markov chain’s property: state transition diagram -> probability of each state 2-dimensional Markovian chain

Bi-dimensional Markov Chain model EEL6935: Wireless Ad Hoc Networks Bi-dimensional Markov Chain model Behavior of a single station Notations b(t): backoff time counter at time t, stochastic process. s(t): backoff stage at time t, stochastic process. Time scale: Discrete and integer, t, beginning of a slot time, when backoff time counter decrements or regenerated [t, t+1], interval between 2 consecutive slot time, can be variable length Makovian State: B(t) ={s(t), b(t)} CW_i = 2^i*CW_min m: maximum backoff stage, CW_max = 2^m*CW_min p: prob.of each transmitted packet being collided n: number of stations

One step transition probabilities (1) EEL6935: Wireless Ad Hoc Networks One step transition probabilities (1) 1) P{i,k|i,k+1}=1, k : [0,Wi-2], i : [0,m] At beginning of t Backoff counter not reach zero, no transmission Channel sensed idle for 1 mini-slot till t+1 At beginning of t+1 Backoff counter decremented by 1 1 i , k i , k+1

One step transition probabilities (2) EEL6935: Wireless Ad Hoc Networks One step transition probabilities (2) P{0,k|i,0}=(1-p)/W_0, k : [0,W_0-1], i : [0,m] At beginning of t Backoff counter reaches zero, successful transmitted [t,t+1] At beginning of t+1 Contention window reset to CW_min (backoff stage = 0) Backoff counter chosen randomly in [0,W_0-1] P{i+1,k|i,0}= p/W_(i+1), k : [0,W_(i+1)-1], i : [1,m-1] Backoff counter reaches zero, transmit in [t,t+1], collision Contention window < CW_max contention window doubled Backoff counter chosen randomly in [0,W_(i+1)-1]

Illustration of the Backoff After Collison EEL6935: Wireless Ad Hoc Networks Illustration of the Backoff After Collison

EEL6935: Wireless Ad Hoc Networks State transits upon backoff counter reach zero (Contention Window <CWmax) Tx Success 0 , 0 0 , 1 … 0 , W0-2 0, W0-1 . (1-p)/W0 i , 0 p/Wi+1 i+1, 0 i+1 , 1 i+1,Wi+1-2 i+1,Wi+1-1 . . . collision

One step transition probabilities (3) EEL6935: Wireless Ad Hoc Networks One step transition probabilities (3) P{m,k|i,0}= p/W_m, k : [0,W_m-1], i = m At beginning of t Backoff counter reaches zero, transmit in [t,t+1], collision Contention Window = CW_max At beginning of t+1 Contention Window remains at CW_max Backoff time counter chosen randomly in [0,W_m-1]

EEL6935: Wireless Ad Hoc Networks State transits upon backoff counter reach zero (Contention Window = CWmax) 0 , 0 0 , 1 … 0 , W0-2 0, W0-1 . (1-p)/W0 m , 0 m , 1 m , Wm-2 m , Wm-1 … p/Wm

Bi-dimensional Markov Chain model EEL6935: Wireless Ad Hoc Networks Bi-dimensional Markov Chain model

Bi-dimensional Markov Chain model EEL6935: Wireless Ad Hoc Networks Bi-dimensional Markov Chain model

Result from Markovian Chain Model (1) EEL6935: Wireless Ad Hoc Networks Result from Markovian Chain Model (1)

Result from Markovian Chain Model (2) EEL6935: Wireless Ad Hoc Networks Result from Markovian Chain Model (2)

Result from Markovian Chain Model (3) EEL6935: Wireless Ad Hoc Networks Result from Markovian Chain Model (3) and

Throughput Analysis Based on Model EEL6935: Wireless Ad Hoc Networks Throughput Analysis Based on Model S := Fraction of Time channel is used to successfully transmit payload bits As an outside observer, see a random slot and observe what is happening Probability, Exactly 1 TX Occurring on the channel is successful given someone transmits Hybrid Scheme also possible. Packet Length may vary and throughput may relate itself to packet size distribution mean Ts, Tc, s,P are constant for model verification constant, and determined by standard Maximizing throughput over probabilities which are in terms of t, we get S is max when Remember Slotted Aloha Stabilization ? tao depends on m and W and can be changed adaptively But m and W fixed because of Physical Layer Standard Result – S can be significantly lower than maximum

EEL6935: Wireless Ad Hoc Networks IMPORTANT RESULTS For Sufficiently Large n, Smax is practically independent of no. of stations in wireless network Maximum throughput achievable by BAS is very close to RTS/CTS mechanism RTS/CTS scheme throughput is less insensitive to transmission probability t RTS/CTS scheme is network size independent for W <= 64 values. Basic Mechanism throughput increases but significantly decreases with network size Key to these results – RTS/CTS mechanism reduces the time spent during a collision, and it becomes more effective than Basic Access when W and n increases the collision probability RTS/CTS even more effective when packet length are longer

EEL6935: Wireless Ad Hoc Networks Model Validation

Remarks Classic model for WLAN with good approximation EEL6935: Wireless Ad Hoc Networks Remarks Classic model for WLAN with good approximation Delay performance can be derived Only one-hop is considered. Only saturation is considered. A small mistake

EEL6935: Wireless Ad Hoc Networks A Small Mistake Found H. Wu, Y. Peng, K. Long, S. Cheng, and J. Ma, Performance of reliable transport protocol over IEEE 802.11 wireless LAN: analysis and enhancement,. in Proc. IEEE INFOCOM'02, New York, June 2002 Limit of retransmission m and maximum stage of backoff counter m’: [ m {5 or 7}, m’ {5} ]

EEL6935: Wireless Ad Hoc Networks Modified Model

Validation of the Modified Model EEL6935: Wireless Ad Hoc Networks Validation of the Modified Model

EEL6935: Wireless Ad Hoc Networks Thank you