Allocating Priority On The Mac of Ieee 802.11 DCF in Dissemination of Beacon Messages

International Journal of Mechanical Engineering and Technology (IJMET)
(IJM
Volume 9, Issue 12, December 2018,
201 pp. 20–28, Article ID: IJMET_09_12_003
Available online at http://www.iaeme.com/ijmet/issues.asp?JType=IJMET&VType=9&IType=12
ISSN Print: 0976-6340
6340 and ISSN Online: 0976-6359
0976
© IAEME Publication Scopus Indexed
ALLOCATING PRIORITY ON THE MAC OF

IEEE 802.11 DCF IN DISSEMINATION
DISSEMINATION OF
BEACON MESSAGES
C. Bazil Wilfred, M. Selva Rathi and P. Xavier
Department of Mathematics,
Karunya Institute of Technology and Sciences, Coimbatore, India
ABSTRACT
The MAC performance of IEEE 802.11 DCF and the IEEE 802.11e EDCA is
analyzed in non-saturated
saturated condition which means sometimes
sometimes the device does not have
a packet for transmission. In this paper,
paper, it is assumed that there is no creation of
message when the previous message is still in progress. The number of messages in
flow is distributed geometrically. In the non-saturatedd condition, the packet which
comes first at the idle station is passed on without entering into backoff procedure.
This paper attempts at modelling the stochastic performance of one station as a
discrete time Markov Chain.
Keywords: DCF, Discrete Time Markov Chain, MAC, Non-Saturation
Saturation Condition.
Condition
Cite this Article: C. Bazil Wilfred, M. Selva Rathi and P. Xavier, Allocating Priority
on The Mac of IEEE 802.11 DCF In Dissemination of Beacon Messages,
Messages International
Journal of Mechanical Engineering and Technology, 9(12), 2018, pp. 20–28.
20
http://www.iaeme.com/ijmet/issues.asp?JType=IJMET&VType=9&IType=12
1. INTRODUCTION
With the increase of global temperature, especially in the tropical regions and the increase of
pollution, noise etc., no vehicle runs without air-conditioner,
air conditioner, thereby ‘no hand signal’
indication is stuck on almost every vehicle. Indicators are the only helpinghe hands which
bridge the communication between the vehicles and the drivers. The increase in the road
accidents is directly proportional to the increase in the traffic on the roads. It is imperative at
this juncture to impose a priority on the beacon messages
messages transmitted in the MAC layer.
In order to facilitate and assist road safety applications electronically, a fool proof and an
effective inter-vehicular
vehicular communication is critically required. Thus, in the procedure of
grouping vehicles, which is termed
termed as Vehicle Platooning, a high speed inter-vehicular
inter
communication is inevitable, lest the delayed communication causes havoc to the entire
process. Hence, a highly effective Medium Access Control (MAC) procedure becomes
significant. The paper is aimed at the making of a MAC method to suit a vehicular cluster,
which will be efficient to transmit beacon message very swiftly but with a better sustainability
http://www.iaeme.com/IJMET/index.
IJMET/index.asp 20 editor@iaeme.com
Allocating Priority on The Mac of IEEE 802.11 DCF In Dissemination of Beacon Messages
than IEEE 802.11p, simultaneously transmitting event-driven messages promptly. The

protocol allots the turn within the cluster based on backoff procedure.
The medium access control (MAC) of this wireless communication system employs a
mandatory contention-based channel access function called the Distributed Coordination
Function (DCF). Due to the robust increase of real time usage of data for varied internet
applications, quality of service (QoS) of these applications such as the guaranteed packet
delay and packet loss probability is essential. However, 802.11 DCF protocol does not
provide QoS support. A new Hybrid coordination function (HCF) in IEEE 802.11e is to
ensure the QoS support.
Generally, there are two classifications such as Saturation Condition and Non-saturation
condition. The assessment of the performance and the receiving of output are very complex in
non-saturated condition when compared to saturated condition. Despite its complexities, it is
enormously deployed in numerous applications in real time WLAN networks. Thus, the non-
saturated condition is analysed.
Based on the onset of the flow patterns of streams of messages at stations the non-
saturated mode is classified into three types:
Messages arrive and line up in a buffer at a station to form a queue. There can arise a
situation when even during the process time of a message, a new message arrives.
A stream consisting of many messages are arrayed according to a Poisson process and
every stream in itself produces a new active station.
A new message is not generated until the previous message completes its transmission.
For instance, when a user transmits a message using his mobile phone or laptop in the WLAN
area, a request to send (RTS) protocol is generated and sent. This message may contain
several packets. These packets get transmitted using 802.11 DCF protocol.
Normally, the user does not create a new request when the previous message is still in
service. When the message gets transmitted, the user gets prepared to send the next message
or to get a response to his previous message. The time taken between these two actions is the
idle time for the user, during which the station remains idle with no messages to transmit.
Based on the third condition mentioned in the non-saturated condition, we proceed to
assess the performance of IEEE 802.11 DCF. We model the stochastic behavior of one
station as a discrete time Markov chain and obtain the channel throughput.
2. PERFORMANCE ANALYSIS OF IEEE 8201.11 DCF

1.1. Basic assumptions
Let n be the number of stations. Considering the third case of the non-saturation condition the
station does not generate flows while the station has a flow in service. After completion of
transmission of a flow the station goes to idle state and it takes exponential duration with rate
λ for a station to generate a new flow. That is the inter arrival time of flow is exponentially
distributed with rate λ . We assume each flow consists of geometrically distributed number of
1
packets with mean that is the distribution of the number L of packets in a flow is given
1− φ
by P( L = k ) = φ k −1 (1 − φ ) . The station generates a new packet with probability φ immediately
after a previously packet has been transmitted and goes to idle state with probability (1 − φ ) .
To make our model to be a Markov chain we adopt the latter idea. All packets are assumed to
have the same payload length.
http://www.iaeme.com/IJMET/index.asp 21 editor@iaeme.com
1.2. Mathematical modeling

The backoff counter decrement is freezed when the channel is sensed busy. Embedded points
of the Markov chain are epochs where the backoff counter of the tagged station decrements
and so a slot is classified as an idle slot, a successful transmission slot, and a collision slot.
Therefore the time interval between two consecutive slot may be much longer than the idle
slot time size σ , as it could be the duration of a packet transmission. For convenience we
denote W = CWmin and Wi = min {2i W , 2 N W } where N is the maximum backoff stage [2].
The state space of our Markov chain are;
• Idle: the state in which the station has no packet to transmit
• (−1, d ),1 ≤ d ≤ 3 : The states in which the station monitors the channel activities during
DIFS when the first packet of a flow arrives. Since DIFS = SIFS + 2 ⋅ σ ,
(−1,1) ,is the state of sensing channel during SIFS,
(−1, 2) , and (−1,3) denote the states of two slot next SIFS period.
• (i, k ) : the states in which the station is in backoff procedure. i denotes the backoff stage with
0 ≤ i ≤ M where M is retransmission limit excluding initial attempts. So a packet can
experience M + 1 transmission attempts. If a packet is not successfully transmitted at the
( M + 1) th attempt, the packet is discarded. The backoff stage is reset to 0 and the contention
window is reset to CWmin after every successful packet transmission or packet discard. k
denotes the backoff counter with 0 ≤ k ≤ Wi .
The three cases that the first packet of the flow of our tagged station can experience are as
below;
• The first packet arrived at the idle station is immediately transmitted without backoff
procedure after it senses the channel being idle during DIFS period. Figure a.
The channel is occupied by other stations during which the tagged station services channel
condition during DIFS period after packet arrival. In this case, the tagged station starts a back
off procedure after DIFS period following the other station’s transmission. Figure b.
The first packet of the tagged station arrives during busy slot. The tagged station
postpones a backoff procedure until the channel is idle during DIFS period. On the other
hand, the ordinary packets are always transmitted through backoff procedure. Figure c.
1.3. Markov Chain modelling

The following figure d is the one-step transition probability of the Markov chain model
describing backoff procedures for the tagged station in non-saturation condition. In Fig. d,
p is the conditional collision probability which is assumed as constant regardless of the
backoff stage. Let p0 , p1 and p∗ be the conditional probabilities that a randomly chosen slot
is an idle slot, a successful transmission slot and a collision slot, given that the tagged station
has no packets to transmit, respectively. Then,
p0 = (1 − τ ) n −1 (1)
p1 = (n − 1)τ (1 − τ ) n − 2 (2)
p∗ = 1 − p0 − p1 (3)
where is the probability that the tagged station transmits in a randomly chosen slot.
Let pa and pb be the probabilities of packet arrival in an idle slot and a busy slot of
channel when the tagged station has no packet to transmit respectively. Since inter-arrival
time of flows is exponentially distributed with rate λ as mentioned pa and pb are calculated
as,
pa = p0 ⋅ (1 − e− λ ⋅σ ) (4)
pb = p1 ⋅ (1 − e − λ ⋅TS ) + p ∗ ⋅ (1 − e− λ ⋅TC ) (5)

TS and TC the durations that the channel is sensed busy during a successful transmission
and a collision respectively, is calculated by
TS = T0 + Tp + SIFS + Ta + DIFS (6)
TC = T0 + Tp∗ + DIFS (7)
where T0 and Ta denote the durations to transmit overhead (PHY overhead + MAC
overhead) and ACK packet, respectively. Tp is the average transmission time of data payload
and Tp∗ is the average transmission time of the longest data payload involved in a collision.
Since we assume that all packets have the same payload size, Tp∗ = Tp .
Fig. d.
1.4. Computation of stationary probabilities

Computation of the stationary probabilities namely bidle , b−1,d and bi ,k , of the Markov chain.
For the state bidle
b00 (1 − p )(1 − φ ) + b1,0 (1 − p )(1 − φ ) + ... + bM −1,0 (1 − p )(1 − φ ) + bM ,0 (1 − p )(1 − φ ) = bidle ( pa + pb )
[b00 + b1,0 + ... + bM −1,0 ](1 − p )(1 − φ ) + +bM ,0 (1 − φ ) = bidle ( pa + pb )
[b00 + pb0,0 + p 2 b0,0 + ... + p M −1b0,0 ](1 − p )(1 − φ ) + p M b00 (1 − φ ) = bidle ( pa + pb )
b00 [1 + p + p 2 + ... + p M −1 ](1 − p )(1 − φ ) + p M b00 (1 − φ ) = bidle ( pa + pb )
 1− pM  M
b00   (1 − p)(1 − φ ) + p b00 (1 − φ ) = bidle ( pa + pb )
 1− p 
b00 (1 − φ ) − p M b00 (1 − φ ) + p M b00 (1 − φ ) = bidle ( pa + pb )
b00 (1 − φ )
∴ bidle = (8)
( pa + pb )
For the state b−1,D
bidle pa = b−1, D (1 − p0 ) + p0 
b−1, D = bidle pa
For the state b−1,1
b−1,1 = b−1,2 p0
b−1,2 = b−1,3 p0
b−1,1 = b−1,3 p0 p0
= bidle p0 2 ⋅ pa
b−1,2 = bidle pa ⋅ p0
= bidle p0 ⋅ pa
b−1,3 = bidle p0 0 ⋅ pa
b−1,d = bidle p03− d ⋅ pa if d ∈ [1, 3] (9)
For the state bi ,0 , bi ,0 : 1≤ i ≤ M

p
bi ,1 + bi −1,0 = bi ,01
cwi
p
bi ,2 + bi −1,0 = bi ,1
cwi
p
bi ,3 + bi −1,0 = bi ,2
cwi
p
bi ,cw −1 + bi −1,0 = bi ,cw − 2
i
cwi i
p
bi −1,0 = bi ,cw −1
cwi i
Adding
cwi −1 cwi −1
p
∑b
j =1
i, j + bi −1,0
cwi
= ∑b
j =0
i, j
bi −1,0 ⋅ p = bi ,0
bi ,0 = bi −1,0 ⋅ p, i ≥1
bi −1,0 = p ⋅ bi − 2,0
b1,0 = p ⋅ b0,0
i
Therefore bi ,0 = p ⋅ b0,0 (10)
For the state bi ,k

p
bi ,k = bi .k −1 − bi −1,0
cwi
cwi −1 cwi −1
p
∑b
k =1
i ,k = ∑b
k =0
i , k −1 − (cwi − 1)bi −1,0
cwi
, for, 1 ≤ k ≤ cwi − 1
cwi − 2
p
= ∑b
k =0
i ,k − (cwi − 1)bi −1,0
cwi
cwi −1 cwi − 2
p
∑b
k =1
i ,k = bi ,0 + ∑b
k =1
i ,k − (cwi − 1)bi −1,0
cwi
bi −1,0 p
bi ,cw −1 = bi ,0 − (cwi − 1)
i
p cwi
cw − 1
= bi ,0 − ( i )bi ,0
cwi
cw − 1
= bi ,0 − ( ii )bi ,0
2 w0
cwi = 2i w0
bi ,0 p
bi ,k = bi , k −1 −
p cwi
bi ,0
bi , k = bi ,k −1 −
cwi
bi ,0
bi ,k −1 = bi , k − 2 −
cwi
bi ,0
bi , k − 2 = bi ,k −3 −
cwi
bi ,0
bi ,1 = bi ,0 −
cwi
Adding
k
bi ,k = bi ,0 − bi ,0
cwi
k
bi ,k = bi ,0 − i bi ,0 , for N + 1 ≤ i ≤ M
2N
 k 
bi , k = bi ,0 1 − i  , for i ∈ [1, N ] (11)
 2 N
Similarly we get
 k 
bi ,k = bi ,0 1 − i  , for i ∈ [ N + 1, M ] (12)
 2 N
To find b0,k
D D
pb 1 1 1
b0,k = bi ,k +1 + bidle + ∑ b−1, j (1 − p0 ) + ∑ bi ,0 (1 − p)φ + bM ,0 ⋅ φ
CW0 j =1 CW0 i = 0 CW0 CW0
M −1
p 3
(1 − p0 ) 1 1
b0,k +1 = bi , k − bidle b − ∑ b−1, j − ∑ bi ,0 (1 − p )φ − bM ,0 ⋅ φ
CW0 j =1 CW0 i =0 CW0 CW0
M −1
p 3
(1 − p0 ) φ φ
b0,k = bi , k −1 − bidle b − ∑ b−1, j − ∑ bi ,0 (1 − p) − bM ,0 ⋅
M −1
p 3
(1 − p0 ) φ φ
b0,1 = b0,0 − bidle b − ∑ b−1, j − ∑ bi ,0 (1 − p) − bM ,0 ⋅
k k  3 M −1

b0,k = b0,0 − pb ⋅ bidle −  (1 − p 0 ∑ −1, j
) b + ∑ bi ,0 (1 − p)φ + bM ,0 ⋅ φ 
w0 CW0  j =1 i =0 
k k  3 M −1

b0,k = b0,0 − pb ⋅ bidle −  (1 − p0 )∑ b−1,d + ∑ bi ,0 (1 − p )φ + bM ,0 ⋅ φ 
w0 w0  d =1 i =0 
k  3
 k  1 − p M  
0 ∑ b−1, d  −
M
= b0,0 − p ⋅ b
 b idle + (1 − p )  (1 − p )φ   b0,0 + φ p b0,0 
w0  d =1  w0   1− p  

k 3
 k
= b0,0 − 

w0
p b ⋅ bidle + (1 − p 0 ) ∑
d =1
b−1,d  −
 w0
φ b0,0 
k  3

b0, k = b0,0 − 
w0 
pb ⋅ bidle + (1 − p 0 ) ∑
d =1
b−1,d + φ b0,0 

(13)
Equations (8),(9),(10),(11) and (12) give the stationary probabilities of the Markov Chain.
Thus we have expressed all the probabilities in terms of b0,0 . b0,0 is determined by the
normalization condition. As any transmission occurs when the backoff time counter is equal
M
to zero, the probability τ is given by τ = ∑ bi ,0 and τ is a function of p . The conditional
i=0
collision probability p is the same as probability that, in a time slot, at least one of the n − 1
remaining stations transmit. Therefore, the probability p can be written as p = 1 − (1 − τ )n −1 .
Thus the two equations represent a nonlinear system of two unknown variables, from which
we obtain and p by using numerical techniques.
3. CONCLUSION
In this paper we have developed an analytical model to evaluate the performance of the IEEE
802.11 DCF in non-saturation condition. We have constructed a discrete time Markov chain
to describe the stochastic behavior of one station and the stationary probabilities are
evaluated.
REFERENCES
[1] G. Bianchi, “Performance Analysis of the IEEE 802.11 Distributed Coordination
Function,” IEEE Journal on Selected Areas in Communications, Vol. 18, No. 3, pp. 535–
547, Mar. 2000.
[2] Tae Ok KIM, Kyung Jae KIM, Bong Dae CHOI, “Performance analysis of IEEE 802.11
DCF and IEEE 802.11e EDCA in non-saturation condition”, IEICE Trans.
Communication., VOL. E91-B, No.4 April 2008.
[3] IEEE, “International Standard Information Technology-Telecommunications and
information exchange between systems-Local and metropolitan area networks-Specific
requirements Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer
(PHY) Specifications,” IEEE 802.11-1999, 1999.
[4] IEEE, “IEEE Standard for Information Technology - Telecommunications and
information exchange between systems - Local and metropolitan area networks - Specific
requirements - Part 11: Wireless Medium Access Control (MAC) and Physical Layer
(PHY) Specifications: Amendment 8 Medium Access Control (MAC) quality of service
(QoS) enhancements,” IEEE P802.11e-2005, 2005.

Allocating Priority On The Mac of Ieee 802.11 DCF in Dissemination of Beacon Messages

Enviado por

Dados do documento

Título original

Direitos autorais

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Allocating Priority On The Mac of Ieee 802.11 DCF in Dissemination of Beacon Messages

Enviado por

Direitos autorais:

International Journal of Mechanical Engineering and Technology (IJMET)

© IAEME Publication Scopus Indexed

ALLOCATING PRIORITY ON THE MAC OF

than IEEE 802.11p, simultaneously transmitting event-driven messages promptly. The

2. PERFORMANCE ANALYSIS OF IEEE 8201.11 DCF

1.2. Mathematical modeling

1.3. Markov Chain modelling

pb = p1 ⋅ (1 − e − λ ⋅TS ) + p ∗ ⋅ (1 − e− λ ⋅TC ) (5)

TS = T0 + Tp + SIFS + Ta + DIFS (6)

TC = T0 + Tp∗ + DIFS (7)

1.4. Computation of stationary probabilities

For the state bidle

b00 (1 − p )(1 − φ ) + b1,0 (1 − p )(1 − φ ) + ... + bM −1,0 (1 − p )(1 − φ ) + bM ,0 (1 − p )(1 − φ ) = bidle ( pa + pb )

[b00 + b1,0 + ... + bM −1,0 ](1 − p )(1 − φ ) + +bM ,0 (1 − φ ) = bidle ( pa + pb )

[b00 + pb0,0 + p 2 b0,0 + ... + p M −1b0,0 ](1 − p )(1 − φ ) + p M b00 (1 − φ ) = bidle ( pa + pb )

b00 [1 + p + p 2 + ... + p M −1 ](1 − p )(1 − φ ) + p M b00 (1 − φ ) = bidle ( pa + pb )

b00 (1 − φ ) − p M b00 (1 − φ ) + p M b00 (1 − φ ) = bidle ( pa + pb )

For the state b−1,D

bidle pa = b−1, D (1 − p0 ) + p0 

For the state b−1,1

b−1,d = bidle p03− d ⋅ pa if d ∈ [1, 3] (9)

For the state bi ,0 , bi ,0 : 1≤ i ≤ M

For the state bi ,k

Você também pode gostar