17
Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia- Luna-Aceves

Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Embed Size (px)

Citation preview

Page 1: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Delay Analysis of IEEE 802.11 in Single-Hop Networks

Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Page 2: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Outline

The Distributed coordination function mechanism

Service time characterization Channel probabilities Model validation Performance evaluation of DSSS and FHSS Conclusion

Page 3: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

The Distributed coordination function(DCF) backoff

DCF describes two techniques for packet transmissiom Basic access mechanism RTS/CTS access mechanism

The backoff time counter is decremented if the channel is sensed idle

Otherwise it is frozen in its current state until the channel is sensed idle more than a DIFS

Page 4: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Notation Three possible events a node can sense during its backoff

= {successful transmission} , ={idle channel} , ={collision} , Let denote the backoff stage, Let , , and denote the number of idle slots, collision

slots and successful transmission slots respectively, and , and denote their probabilities respectively.

and is the average time the channel is sensed busy due a collision and a successful transmission respectively. is the time used when the channel is sensed idle (one backoff time)

The average backoff step size is

sEcEiE

}{ ss EPP

}{ ii EPP k

}{ cc EPP

sscci ptptt

ir

ir cr srip cp sp

ct st

Page 5: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Service time characterization At k-th backoff stage, the number of backoff steps is chosen

uniformly from . The average number of backoff steps is The average time a node spends at k-th stage is Let the probability of collision. The probability of successful

transmission at k-th stage is The cumulative delay of a node at k-th backoff stage is

)

2()(2

2)(

)2

(22

)1(22

)(

)1(2

)1()1()(

min2

1min

1

min

1

1 1

wtkt

wkD

tktw

tkkwkD

tkw

tkDkD

cck

c

k

ic

ic

k

ik

c

k

i

k

i

kck

2

1 k

k

wD

1)1( kk ppp

p

2

1kw kw,0

Page 6: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Service time characterization The average single-hop delay considering the frame

retry limit is given by the following expression

Where

c

k

kc

k

kc

k

k

k

k

i ik

tp

p

p

wD

pw

tpkptppwD

ppkDpkDD

)1(

2

)1(

12

12

12

1

min

1

1

min

1

1

1

1min

1 1

1

121

12

mm

p

pq

Page 7: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Channel Probabilities This model is applicable whenever , , and are known These probabilities are computed for saturated, single hop ad

hoc network under ideal channel condition (i.e, no hidden terminals and capture), with fixed number of nodes

This analysis is based on Bianchi framework, the probability that a node transmits in a randomly chosen slot time is

Where p is the probability of a collision experienced by a transmitted packet on the channel

mppwwp

p

21121

212

minmin

111 np

ip cp sp

Page 8: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Channel Probabilities Let then

Using the Taylor series expansion of at p=0 , the first order approximation of is

is approximated the following

1 m

m

ppwwp

ppwwp

21121

21121

minmin

minmin

p p

pw

w

w

wp 2

min

min

min

min

1

2

1

1

p

121

12

min2

min

min

nww

nwp

Page 9: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Channel Probabilities

Page 10: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Channel Probabilities

Page 11: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Channel Probabilities

Probability that there is one transmission in the considered time slot is

Let is the probability that a transmission occurring on the channel is successful

The probability of collision is

The probability of successful transmission is

The probability of idle slot is

111 nsucp

1

211

11

1

n

nn

sucp

sucn

s pp 111

sucn

c pp 111 1

1111 nip

Page 12: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Model Validation Ns-2 simulator Network size from 8 to 56

nodes Packet size 1500 bytes Nodes randomly distributed Network area 20*20 m No mobility Performance metrics:

service time and jitter

Page 13: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Model Validation

Page 14: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Performance Evaluation

Page 15: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Performance evaluation

Page 16: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Performance Evaluation

Page 17: Delay Analysis of IEEE 802.11 in Single-Hop Networks Marcel M. Carvalho, J.J.Garcia-Luna-Aceves

Conclusion DSSS performs better than FHSS in term of delay and

jitter The higher the initial contention window size, the

smaller the average service time and jitter are, especially for large networks

The binary exponential backoff algorithm has negative impact if both the maximum backoff stage and the number of nodes in the network are large