IEEE Journal on Selected Areas in Communications Performance Analysis of the IEEE 802.11 Distributed Coordination Function Giuseppe Bianchi IEEE Journal on Selected Areas in Communications March 2000
Outline Introduction 802.11 Distributed Coordination Function Maximum & Saturation Throughput Performance Throughput Analysis Model Validation Maximum Saturation Throughput Performance Evaluation Conclusion
Introduction 802.11 Distributed Coordination Function The fundamental mechanism to access the medium Based on CSMA/CA Two techniques Basic Access Mechanism RTS/CTS Mechanism
802.11 DCF Two access techniques Basic mechanism: 2 way handshaking RTS/CTS mechanism: 4 way handshaking RTS DATA CTS DESt Source Dest Source DATA ACK ACK
802.11 DCF
802.11 DCF
802.11 DCF
Maximum and Saturation Throughput Performance Maximum throughput performance Saturation throughput performance Maximum load in stable condition
Throughput Analysis Assumption Fixed # of stations Always having a packet available for transmission Transmission queues are always nonempty Two parts of analysis Study the behavior of single station with a Markov model Study the events that occur within a generic slot time & expressed throughput for both Basic & RTS/CTS access method Obtain the stationary probability
Throughput Analysis n stations b(t) W = 𝐶𝑊 𝑚𝑖𝑛 ; 𝐶𝑊 𝑚𝑎𝑥 = 2 𝑚 W s(t) Each station always has a packet available for transmission b(t) Stochastic Process representing backoff time counter W = 𝐶𝑊 𝑚𝑖𝑛 ; 𝐶𝑊 𝑚𝑎𝑥 = 2 𝑚 W s(t) Stochastic Process representing backoff stage (0,m)
Throughput Analysis Each packet collide with constant and independent probability p Model bi-dimensional process {s(t) , b(t)} with discrete-time Markov chain
Markov Chain model
Markov Chain model Stationary distribution of the chain 𝑏 𝑖,𝑘 = lim 𝑡→∞ 𝑃{𝑠 𝑡 =𝑖, 𝑏 𝑡 =𝑘} i ϵ ( 0, m ) , k ϵ ( 0, 𝑊 𝑖 -1 )
Markov Chain model 𝑏 𝑖,𝑘 = 𝑊 𝑖 − 𝑘 𝑊 𝑖 𝑏 𝑖,0 1= 𝑖=0 𝑚 𝑘=0 𝑊 𝑖 −1 𝑏 𝑖,𝑘 𝑏 𝑖,𝑘 = 𝑊 𝑖 − 𝑘 𝑊 𝑖 𝑏 𝑖,0 1= 𝑖=0 𝑚 𝑘=0 𝑊 𝑖 −1 𝑏 𝑖,𝑘 𝑏 0,0 = 2(1−2𝑝)(1−𝑝) 1−2𝑝 𝑊+1 +𝑝𝑊(1− (2𝑝) 𝑚 ) Probability τ a station transmits in randomly chosen slot time
Markov Chain model Some note In general, τ depends on p If m = 0 , 𝜏= 2 𝑊+1 Independent of p In general, τ depends on p 𝑝=1− (1−𝜏) 𝑛−1
Throughput Normalized system throughput S Probability of transmission 𝑃 𝑡𝑟 At least one transmission in the slot time 𝑃 𝑡𝑟 =1− 1−𝜏 𝑛 Probability of successful transmission 𝑃 𝑠 Transmit successfully 𝑃 𝑠 = 𝑛𝜏 (1−𝜏) 𝑛−1 𝑃 𝑡𝑟 = 𝑛𝜏 (1−𝜏) 𝑛−1 1− (1−𝜏) 𝑛
Throughput 𝑆= 𝐸[𝑝𝑎𝑦𝑙𝑜𝑎𝑑 𝑖𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 𝑡𝑟𝑎𝑛𝑠𝑚𝑖𝑡𝑡𝑒𝑑 𝑖𝑛 𝑎 𝑠𝑙𝑜𝑡 𝑡𝑖𝑚𝑒] 𝐸[𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑎 𝑠𝑙𝑜𝑡 𝑡𝑖𝑚𝑒] E[P]: average packet payload size 𝑇 𝑠 : average time the channel is sensed busy because of a successful transmission 𝑇 𝑐 : average time the channel is sensed busy by each stationi during a collusion
Throughput
Throughput
Maximum Saturation Throughput Optimal 𝜏≈ 1 𝑛 𝑇 𝑐 /2𝜎 𝑆 𝑚𝑎𝑥 = 𝐸[𝑃] 𝑇 𝑠 +𝜎𝐾+ 𝑇 𝑐 (𝐾 𝑒 1 𝐾 −1 −1) K= 𝑇 𝑐 /2𝜎
Model Validation
Performance Evaluation Basic RTS/CTS
Performance Evaluation Basic RTS/CTS
Performance Evaluation
Performance Evaluation
Performance Evaluation
Conclusion Evaluated the 802.11 DCF throughput performance Model suited for both Basic Access and RTS/CTS Access mechanisms The model is extremely accurate in predicting the system throughput Basic Access strongly depends on n and w RTS/CTS is better in large network scenarios