A Multichain Backoff Mechanism for IEEE WLANs Alkesh Patel & Hemant Patel ECE 695 – Leading Discussion By : Shiang- Rung Ye and Yu-Chee Tseng

Background WLAN is emerging as a promising technology. MAC plays an important role on efficient and fair use of the wireless medium. MAC plays an important role on efficient and fair use of the wireless medium. Multiple Access Scheme (CSMA) require station to sense carries on the wireless channel before transmitting

Related Work MILD increases the contention window by 1.5 times when collision occurs & decreases the contention window by 1 when transmission succeed In high traffic load this increases collusion probability and decreases throughput DCF of IEEE is a variant of persistent CSMA If medium is busy, transmission defer until the medium becomes idle If medium is busy, transmission defer until the medium becomes idle Physical Carrier Sense  Physical layer Physical Carrier Sense  Physical layer Virtual Carriers Sense  MAC Virtual Carriers Sense  MAC

Concept/Idea MCB – Multi Chain Backoff MCB – Multi Chain Backoff Enables station to adapt to different congestion level Enables station to adapt to different congestion level No restriction on number of contending stations No restriction on number of contending stations High throughput and fair channel access High throughput and fair channel access

MCB Algorithm During back off period station shall detect any collision event by other station Collision flag f col is used to record whether frame collision occurs on the wireless channel f col set to 1 if a station itself experiences a collision or medium is busy – Collision/Transmission occurring with other station f col set to 1 if a station itself experiences a collision or medium is busy – Collision/Transmission occurring with other station f col set to 0 after successful transmission. Backoff counter reaches 0, the station transmit data. f col set to 0 after successful transmission. Backoff counter reaches 0, the station transmit data.

MCB Algorithm Figure 1: The transition diagram of MCB. The j- th backoff stage of chain i is denoted by (i, j) in the figure

MCB Algorithm w i : The minimum contention window of chain i m i : The maximum back off stage of chain i u i : The transition probability from chain i to chain i+1 v i : The transition probability from chain i to chain i -1

Performance Analysis Saturation Throughput S (i, j, k) The station is in stage j of chain i & has a backoff value k. (i, j, k) The station is in stage j of chain i & has a backoff value k. T  Probability that station will transmit in a randomly chosen backoff slot T  Probability that station will transmit in a randomly chosen backoff slot Xi  Probability that a station will detect at least one collision during backoff period Xi  Probability that a station will detect at least one collision during backoff period

Performance Analysis Expressing collision probability can be expressed by T Now the saturation throughput can be obtained by Now the saturation throughput can be obtained by P tr  Probability that transmission occurs randomly P tr  Probability that transmission occurs randomly P s  Probability that transmission succeeds in a backoff slot P s  Probability that transmission succeeds in a backoff slot T s  Time require for frame exchange T s  Time require for frame exchange T c  Length of a colliding duration T c  Length of a colliding duration

Performance Evaluation Performance opposed to MILD, DCF, GDCF, EIED Comparison of saturation throughput & fairness index FI is bounded in the interval [0,1] Algorithm is fair as its FI is close to 1 Optimal value of u and v when they are the same

Fig 3: The optimal u and v with frame size 1024

Fig 4: The ratio of optimal u and v under different n and c

Fig 5: Saturation throughput under different u and v with a frame size of 1024 bytes

Fig 6: Saturation throughput under different u and v with a frame size of 128 bytes

Fig 7: Saturation throughput versus frame sizes

Fig 8: Throughput of MCB with u and v which are chosen for n=6 and n=46

Fig 9: Fairness Index

Fig 10: Saturation throughput of MCB and GDCF with frame size 128 bytes

Fig 11: Saturation throughput of MCB and GDCF with frame size 1024 bytes

Fig 12: Fairness index of MCB and GDCF with frame size 1024 bytes

Fig 13: Saturation throughput of MCB, IEEE , and MILD

Fig 14: Saturation throughput of MCB and EIED(x, y) with fixed y=2 Fig 15: Throughput of MCB and EIED(x, y) with fixed x =2 Fig: 14 Fig: 15

Fig 15: Throughput of MCB and EIED(x, y) with fixed x=2

Fig 16: Fairness index of MCB and EIED (x, y) with fixed x = 2

Conclusion MCB algorithm explores the possibility of using multiple backoff chains Considering collision event offers To choose a proper chain for transmission To choose a proper chain for transmission Capability of switching to different backoff chain Capability of switching to different backoff chain High throughput then any existing algorithms High throughput then any existing algorithms Fair access to the wireless channel Fair access to the wireless channel How to apply Multichain concept to an error- prone wireless channel to resolve the issue would be the next step in future development