Throughput and delay analysis of IEEE 802.11 protocol

Throughput and delay analysis of IEEE 802.11 protocol
P. Chatzimisios, V. Vitsas and A. C. Boucouvalas
Multimedia Communications Research Group,
School of Design, Engineering and Computing,
Bournemouth University, Fern Barrow, Poole, Dorset, BH12 5BB, UK
{pchatzimisios, vvitsas, tboucouv}@bournemouth.ac.uk
Abstract - Wireless technologies in the LAN environment are
becoming increasingly important. The IEEE 802.11 standard is
the most mature technology for Wireless Local Area Networks
(WLANs). The performance of the Medium Access Control
(MAC) layer, which consists of Distributed Coordination
Function (DCF) and Point Coordination Function (PCF), has
been examined over the past years. In this paper, we present an
analytical model to compute the saturated throughput of 802.11
protocol in the absence of hidden stations and transmission
errors. A throughput analysis is carried out in order to study the
performance of 802.11 DCF. Using the analytical model, we
develop a frame delay analysis under traffic conditions that
correspond to the maximum load that the network can support
in stable conditions. The behaviour of the exponential backoff
algorithm used in 802.11 is also examined.
I. INTRODUCTION
Recent advances in wireless technology have equipped
portable devices with wireless capabilities that allow
networked communication even while a user is mobile. These
devices include palmtop computers, personal digital assistants
(PDAs), portable computers, digital cameras and printers.
To deal with this wireless connectivity need, various
wireless communication standards have been developed [1].
Two major projects have been involved in standardizing the
physical and the medium access control (MAC) layers for
wireless LANs, namely IEEE 802.11 [2] and ETSI HiperLAN
[3]. This paper focuses on the analysis of the MAC protocol
of the IEEE 802.11 protocol which is the most widely used
WLAN protocol today.
The IEEE 802.11 standard for wireless networks
incorporates two medium access methods. The mandatory
Distributed Coordination Function (DCF) method and the
optional Point Coordination Function (PCF) which provides
Time Bounded Services (TBS). DCF is an asynchronous data
transmission function, which best suits delay insensitive data
(e.g. email, ftp). It is available in ad-hoc or infrastructure
network configurations and can be either used exclusively or
combined with PCF in an infrastructure network. PCF, on the
other hand, best suits delay sensitive data transmissions (e.g.
real-time audio or video) and is only available in
infrastructure environments.
A common problem in wireless LAN systems is hidden
stations. The presence of hidden stations may result in
significant network performance degradation and causes
unfairness in accessing the medium because a station’s
location may result in a larger transmission privilege. The
hidden station problem occurs when a station is causing
interference due to not been able to detect the existence of a
transmission from another station and thus assumes that the
medium is free and available to transmit. As an example, lets
assume that stations A and B are within communication range
of each other and station C is within communication range of
station B, but not of A. Therefore, it is possible that both
stations A and C could try to transmit to station B at the same
time causing a collision. The influence of hidden stations [4]
on the performance of an IEEE 802.11 network has been
studied in [5].
The performance of CSMA protocols over radio channels
was investigated in [6]. The MACAW protocol [7] was
designed to improve wireless communication performance
based on collision avoidance technique. MACAW is based on
the Multiple Access Collision Avoidance (MACA) protocol
introduced in [8] and enhances MACA’s performance.
Several other papers [9][10][11][12] have studied the
efficiency of the IEEE 802.11 protocol by investigating the
maximum throughput that can be achieved under various
network configurations. Reference [13] analyses the backoff
mechanism and proposes alternatives. Also, in [14] the
throughput of a CMSA/CA protocol is calculated using a
simple model that is space dependent and the fairness
problem is considered. Reference [15] investigates the IEEE
802.11 MAC protocol capacity by deriving an accurate
analytical estimate of it. Moreover, an extension of the
protocol backoff algorithm is proposed.
Bianchi [16] presents a simple analytical model to compute
the saturation throughput performance assuming a finite
number of stations and ideal channel conditions. In contrast,
[17] is based on the same model and assumptions and takes
into account the frame retry limit. As a result, the throughput
of 802.11 can be predicted more accurately.
In our work a delay analysis of both access mechanisms,
basic and RTS/CTS, is developed. This delay analysis is
applied to the model initially presented in [17]. Since, the
frame retry limit is included, the model is considered to
predict the 802.11 frame delay in an accurate way.
The paper is organized as follows: Section II presents in
detail the IEEE 802.11 MAC protocol and includes both basic
access and RTS/CTS mechanism. Section III reviews the
protocol model presented in [17]. Section IV presents the
performance evaluation of both DCF access mechanisms. In
Section V, a frame delay analysis is developed based on the
previous mathematical model. Finally, section VI concludes
the paper and presents future work.
II. DISTRIBUTED COORDINATION FUNCTION (DCF)
The basic service set (BSS) is the basic building block of
IEEE 802.11 WLANs. The coverage area of a BSS is referred
to as the basic service area (BSA). A station that is a member
of the BSS within the BSA may continue communicating
with other members of the BSS. The IEEE 802.11 defines
two types of network architecture, the ad hoc network and the
infrastructure network. An ad hoc network deliberates on the
grouping of stations into a BSS without the need for any
infrastructure implementation. This type of IEEE 802.11
WLAN is often formed for only as long as the WLAN is
needed. The infrastructure networks, in contrast to the ad hoc
networks, create a range extension and obtain some specific
services from other wired or wireless LANs via infrastructure
implementations.
The DCF is based on the Carrier Sense Multiple Access
with Collision Avoidance (CSMA/CA) protocol. Under DCF,
data frames are transferred via two methods. The essential
method used in DCF is called Basic Access method and it is
shown is Fig. 1. The 802.11 standard also provides an
alternative way of transmitting data frames, namely the
RTS/CTS method, illustrated in Fig. 2.
Carrier sensing can be performed on both the physical and
MAC layers. On the physical layer, physical carrier sensing is
done by detecting any channel activity by other stations. In
addition to the physical channel sensing, virtual carrier
sensing is achieved by using time fields in the frames, which
indicate to other stations the duration of the current
transmission. All stations that hear the data or the RTS frame,
update their Network Allocation Vector (NAV) field based on
the value of the duration field in the received frame which
includes the SIFS and the ACK frame transmission time
following the data frame, before sensing the medium again.
A. The basic access method
Priority access to the wireless medium is controlled by the
use of the interframe space (IFS) time period between the
transmission of frames. The IFS defines the minimal time that
a station has to let pass after the end of a frame, before it may
start transmitting a certain type of frame. In 802.11 three
different IFS intervals have been specified to provide various
priority levels for access to the wireless medium: Short IFS
(SIFS), Point Coordination Function IFS (PIFS) and DCF-IFS
(DIFS). The SIFS is the smallest followed by PIFS and DIFS.
After a SIFS (the shortest interframe space) only
acknowledgements, CTS and data frames may be sent. The
use of the PIFS and the DIFS is used to separate the PCF and
DCF modes, giving a higher priority to the former.
In order to minimize the probability of collisions, a random
backoff mechanism is used to randomize moments at which
stations are trying to access the wireless medium. This
contention resolution technique is called binary exponential
backoff (BEB). In particular, the time following an idle DIFS
is slotted and a station is allowed to transmit only at the
beginning of each slot. A slot time is equal to the time needed
by any station to detect the transmission of a frame from any
other station. The backoff counter is decremented when the
medium is idle and is frozen when the medium is sensed
busy. After a busy period the backoff resumes only after the
medium has been idle longer than DIFS. A station initiates a
frame transmission when the backoff counter reaches zero.
Contention Window
DIFS
DIFS
SIFS
Backoff Window
Busy medium
Slot time
Select slot and decrement backoff
as long as medium is idle
Defer access
Fig. 1 Basic Access mechanism
DIFS
RTS
Source
(TX)
DATA
SIFS
SIFS
SIFS
CTS
ACK
Destination
(TX)
Other
DIFS
NAV (RTS)
NAV (CTS)
NAV (DATA)
Defer access
Backoff
Fig. 2 RTS/CTS mechanism
Every station maintains a station short retry count (SSRC)
as well as a station long retry count (SLRC), both of which
have an initial value of zero. The short retry count indicates
the number of retransmissions of RTS frames or of the data
frames when RTS/CTS is not used. The long retry count
indicates the number of retransmissions of data frames when
RTS/CTS is used.
The contention window (CW) is chosen in the interval
(0,CW-1). The value of CW depends on the number of failed
transmissions of a frame. At the first transmission attempt,
CW is set equal to CWmin which is called minimum contention
window. A collision occurs when two or more stations start
transmission simultaneously in the same slot. After each
retransmission due a collision, CW is doubled up to a
′
maximum value, CW max = 2 m ⋅ CW min where m' is the number
of different contention window sizes . Once the CW reaches
CWmax, it will remain at the value of CWmax until it is reset.
The CW is reset to CWmin in the following cases: (a) after
every successful transmission of a data frame, (b) when SSRC
reaches the ShortRetryLimit and (c) when SLRC reaches the
LongRetryLimit. When either of these limits is reached, retry
attempts shall cease and the frame shall be discarded. The
SSRC is reset to 0 whenever a CTS is received in response to a
RTS or whenever an ACK is received in response to a data
frame. The SLRC is also reset to 0 whenever a ACK is
received in response to a data frame when the RTS/CTS is
used.
After a successful frame transmission, if the station still has
frames buffered for transmission, it must execute a new backoff
process. The set of CW values are sequentially ascending
integer powers of 2 minus 1 and Fig. 3 is illustrates this
exponential increase of CW.
1023 1023
CWmax
511
255
31
63
127
31
CWmin
First transmission attempt
6th retransmission
5th retransmission
4th retransmission
3rd retransmission
2nd retransmission
1st retransmission
First transmission attempt
Fig. 3 The exponential increase of CW
After receiving correctly a frame in the destination station, an
immediate positive acknowledgment (ACK) is sent to confirm
the successful reception of the frame transmission after a time
interval equal to SIFS. Since the SIFS interval is shorter than
the DIFS interval, the station sending an ACK attempts
transmission before stations attempting to send data and hence
takes priority. If the source station does not receive an ACK,
the data frame is assumed to have been lost and a
retransmission is scheduled.
B. The RTS/CTS access method
In 802.11, DCF also defines an optional way of transmitting
data frames with transmitting short RTS and CTS frames
before the transmission of the actual data frame. The RTS/CTS
mechanism is mainly used to minimize the amount of time
wasted when a collision occurs and to address the hidden
station problem. When the destination receives a RTS frame,
it transmits a CTS frame immediately after a SIFS interval.
The source station is allowed to transmit its data frame only if
it receives the CTS correctly. If the CTS is not received by
the source station, it is assumed that a collision occurred and a
RTS retransmission is scheduled. After the data frame is
received by the destination station, an acknowledgement
frame is sent back to the source, verifying successful data
reception. The RTS/CTS exchange is shown in Fig.2.
Since all stations are adjusting their NAV based on the
duration field value of the RTS from the source station or the
CTS from the destination station, the use of RTS/CTS frames
helps to minimize the duration of collisions and the collisions
caused by hidden stations. More specifically, if a collision
occurs with two or more small RTS frames, the time loss is
smaller compared to the collision of long data frames.
Furthermore, the successful exchange of small messages,
RTS and CTS, reserves the area within the range of the
receiver and the sender for the intended transmission period
guaranteeing undisturbed transmission for the longer data
frame. On the other hand, RTS/CTS decreases efficiency
since it transmits two additional frames without any payload.
For that reason, there is a manageable object RTS_Threshold
that indicates the data length under which the data frames
should be sent without RTS/CTS. The data frame size is the
only parameter that is used to decide whether the mechanism
is applied. The RTS_Threshold parameter is not fixed in the
802.11 standard and has to be set separately by each station.
III. ANALYTICAL MODEL
In this paper, the assumptions necessary for the presented
analytical framework are as follows:
1. We ignore the effect of frame errors due to bit errors
introduced by channel noise. Therefore frames are received in
error only when they encounter collisions due to other
simultaneous transmissions.
2. No hidden stations and propagation delays are considered.
3. We assume that the network consists of a finite number
of contending stations n and that every station always has a
packet available for transmission (saturated conditions).
4. The main approximation is that the collision probability
of a transmitted frame is constant and independent of the
number of retransmissions that this frame has experienced in
the past.
Let us consider the following sequence of events in a
successful frame transmission using the RTS/CTS access
method:
1. t0 : Station A begins transmitting RTS.
2. t1 = t0+RTS : RTS finished.
3. t2 = t0+RTS+SIFS : Station B begins transmitting CTS.
4. t3 = t0+RTS+SIFS+CTS: CTS finished.
5. t4 = t0+RTS+SIFS+CTS+SIFS: Station A begins
transmitting DATA frame.
6. t5 = t0+RTS+SIFS+CTS+SIFS+DATA: DATA finished.
7. t6 = t0+RTS+SIFS+CTS+SIFS+DATA+SIFS: Station B
begins transmitting ACK frame.
8. t7 = t0+RTS+SIFS+CTS+SIFS+DATA+SIFS+ACK:
ACK finished.
9. t8 =t0+RTS+SIFS+CTS+SIFS+DATA+SIFS+ACK+DIFS:
Next contention period starts.
From the above sequence of events, the time required for a
successful transmission, i.e., the time interval between the start
of a non-colliding transmission and the reception of the ACK
frame, can be easily calculated.
A. Transmission Probability
i − 1,0
bi , 0 = p i ⋅ b0,0 ,
0≤i≤m
W i = 2 i ⋅ W

m′
W i = 2 ⋅ W
i ≤ m′
(2)
(3)
i > m′
where m represents the station short retry count and is equal to
7 according to the standard [2]. Here m is also the maximum
backoff stage and can have a value larger or smaller than m'.
For the DSSS physical layer in 802.11b, we have m' =5.
As the chain is regular, for each k∈[0,Wi-1] we have:
bi , k
m −1

W i − k (1 − p ) ⋅ ∑ b j , 0 + bm , 0
j =0
=
⋅
Wi
 p ⋅ bi −1, 0
,

,
i=0
0<i≤m
(4)
Using (2), (4) can be simplified as:
bi , k =
Wi − k
⋅ bi , 0
Wi
,
(5)
0≤i≤m
Equations (2) and (5) express all bi,k values as a function of
b0,0 and of collision probability p. If the normalization
condition is imposed, we have:
1=
W i −1
m
∑ ∑b
k =0
i=0
i ,k
=
m
∑
i=0
W i −1
bi,0 ∑
k =0
1-p
Wi − k
=
Wi
m
∑b
i=0
i ,0
⋅
Wi + 1
2
0,0
0,1
1
0,2
1
zzz
1
zzz
1-p
i,0
i+1,0
zzz
1
m,0
i+1,1
1
zzz
1
i,1
zzz
1
1
zzz
m,1
1
i,2
i+1,2
zzz
1
0,W0-1
1
1
zzz
1
m,2
1
zzz
1
zzz
zzz
p/Wi+1
zzz
zzz
zzz
zzz
p/Wm
zzz
i,Wi-2
zzz
1
1
i,Wi-1
i+1,Wi+1-2
1
zzz
1
m,Wm-2
i+1,Wi+1-1
zzz
1
m,Wm-1
Fig. 4 Markov chain model
By means of (3) and after some algebra, finally b0,0 is given
by (6).
A station transmits when the backoff counter reaches the
value of zero and the transmission probability τ that a station
transmits a frame in a randomly chosen slot time can be
evaluated as:
m
m
i=o
i=o
τ = ∑ bi , 0 = ∑ p i ⋅ b0,0 = b0,0 ⋅
1 − p m +1
(1 − p )
(7)
and b0,0 values can be acquired from (6).
The transmission probability τ depends on the collision
probability p which is still unknown and it will be derived
next. The probability p that a transmitted frame encounters a
collision, is the probability that at least one of the n-1
remaining stations transmit in the same time slot. If we
assume that all stations see the system in the steady state and
transmit with probability τ, the collision probability p is given
by:
(8)
p = 1 − (1 − τ ) n −1
Equations (7) and (8) form a nonlinear system with two
unknowns τ and p. This nonlinear system can be solved
utilizing numerical methods and has a unique solution.
The previous analysis follows closely the performance
analysis of IEEE 802.11 in [17].
2 ⋅ (1 − 2 p) ⋅ (1 − p)

W ⋅ (1 − (2 p) m +1 ) ⋅ (1 − p) + (1 − 2 p) ⋅ (1 − p m +1 )

b0,0 = 

2 ⋅ (1 − 2 p) ⋅ (1 − p)

m′ +1
m +1
m′
m′ +1
m − m′
W ⋅ (1 − (2 p) ) ⋅ (1 − p) + (1 − 2 p) ⋅ (1 − p ) + W ⋅ 2 ⋅ p ⋅ (1 − 2 p) ⋅ (1 − p )
'
0,W0-2
p/W1
1-p
Let b(t) be the stochastic process representing the backoff
time counter and s(t) be the stochastic process representing the
backoff stage for a given station at slot time t. The
bidimensional process {b(t), s(t)} can be modelled with a
discrete-time Markov chain depicted in Fig. 4. If
b i , k = lim t → ∞ P {s ( t ) = i , b ( t ) = k } , where i∈[0,m] k∈[0,Wi-1]
is the stationary distribution of the Markov chain, then we can
calculate the probability bi,k. We have the following relations :
,
(1)
0<i≤m
b = p ⋅b
i ,0
(1-p)/W0
1
, m ≤ m′
(6)
, m > m′
B. Throughput analysis
The maximum load that the system can carry in stable
conditions is defined as the saturation throughput and is the
limit that the system throughput reaches as the offered load
increases [16]. Ptr is defined as the probability that at least
one transmission occurs in a given slot time. Since n stations
contend to access the medium and each station transmits with
probability τ, Ptr is given by:
(9)
Ptr = 1 − (1 − τ ) n
The probability Ps that an occurring transmission is
successful is given by the probability that a station is
transmitting and the remaining n-1 stations remain silent,
conditioned on the fact that at least one station transmits:
n ⋅ τ ⋅ (1 − τ ) n −1
(10)
Ps =
1 − (1 − τ ) n
Thus, throughput can be expressed by dividing the time
needed to transmit payload information transmitted in a slot
time with the average length of a slot time:
S =
Ptr ⋅ PS ⋅ E [ P ]
(1 − Ptr ) ⋅ σ + Ptr ⋅ PS ⋅ T S + (1 − PS ) ⋅ Ptr ⋅ T C
(11)
where Ts is the average time that the medium is sensed busy
due a successful transmission, Tc is the average time that the
medium is sensed busy by each station when a collision
occurs and σ is the duration of an empty slot.
The values of Ts and Tc depend on the channel access
mechanism. Assuming that all stations use the same channel
access mechanism, Ts and Tc are defined as follows:
The above values of Tc represent the period of time during
which the channel is sensed busy after a collision by the noncolliding stations. For the two or more colliding stations,
there is an extra delay because the colliding stations have to
wait for time equal to ACKtimeout or CTStimeout before sensing
the medium again. This additional delay is neglected and for
these colliding stations Tc has a bigger value than the value
considered here.
IV. THROUGHPUT ANALYSIS RESULTS
This paper uses all the parameters for Direct Spread
Sequence Spectrum (DSSS) physical layer used in 802.11b.
Fig. 5 plots throughput versus frame size for both the basic
access and RTS/CTS cases for three different network sizes
(n=5, 25 and 50) for C=1Mbps. The throughput increases as
the frame size increases. However, we see that if the number
of the active stations is relatively small, n=5, the throughput
of the basic access is higher than the throughput of the
RTS/CTS if the frame size is up to 7000 bits. That means that
the RTS/CTS mechanism should be employed when the
packet size exceeds a specific threshold and in this case the
threshold is equal to about 7000 bits. This threshold decreases
to about 1900 and 1000 when the network is composed by 25
and 50 stations respectively. Additionally, the figure shows
that the RTS/CTS mechanism does not result in significant
throughput improvement for small networks However, for
large networks, the RTS/CTS mechanism is extremely
beneficial for the performance compared to the basic access
mechanism.
0.9
If H=MAChdr+PHYhdr is the frame header, the average time
delays Ts and Tc for the basic access mechanism are:
0.8
T Sbas = DIFS
+ H + E [ P ] + SIFS
T Cbas = DIFS
+ H + E [ PC ] + σ
+ ACK
+ σ
where E[Pc] is the average length of the longest frame
payload involved in a collision. Since it is assumed for
simplicity that the size of all frames is the same and fixed,
therefore, E[P] = E[Pc] = P. Furthermore, there is an extra
delay equal to a slot time σ because the next slot is empty
after a transmission.
If the RTS/CTS access mechanism is employed, we have:
+ RTS + SIFS
+ H + E [ P ] + SIFS
T CRTS = DIFS
+ RTS + σ
0.7
0.6
0.5
0.4
0.3
1000
2000
3000
4000
5000
◆
RTS, n=5
‘ Basic, n=5
S
RTS, n=25
U Basic, n=25
■
RTS, n=50
□ Basic, n=50
6000
7000
8000
9000
10000
Frame size (bits)
RTS/CTS access mechanism:
T S RTS = DIFS
Throughput Efficiency
Basic access mechanism:
+ CTS
+ ACK
+σ
+ SIFS
Fig. 5 Throughput versus frame size for C=1Mbps
Since the IEEE 802.11b standard specifies various data
rates, it is interesting to study how the throughput is affected
by the medium data rate. Fig. 6, plots throughput versus
number of stations for three different data rates (C=1, 5.5 and
11 Mbps) for both medium access mechanisms for frame size
(l=8224 bits).
0.9
Substituting (13) and (15) into (12), the average frame delay
can be easily calculated.
0.85
VI. CONCLUSIONS AND FUTURE WORK
0.75
This figure illustrates that the throughput performance weakly
depends on the number of the stations for all data rates when
RTS/CTS is employed. On the other hand, if basic access is
used, the throughput decreases as the number of the
stations increases because more collisions take place.
Moreover, the throughput efficiency is reduced when the data
rate increases. The situation is explained by considering that
the time spent for frame transmission is decreased as the data
rate increases but the time overhead spent on DIFS, SIFS and
the backoff delay remains the same.
This paper presents an analytical model using a Markov chain
to evaluate the system throughput performance of IEEE 802.11,
the most widely accepted standard for Wireless LANs. This
model can be used for both access mechanisms of Distributed
Coordination Function, in the absence of hidden stations and
transmission errors. The effect of the length of transmitted
frames, the number of contenting stations in the network and
the data rate on the throughput of the system is examined.
Furthermore, a frame delay analysis is introduced under traffic
conditions that correspond to the maximum load that the
network can support in stable conditions.
We have concluded that the throughput performance strongly
depends on the number of stations of the wireless network
when the basic access method is used. Moreover, if the network
is composed of a small number of stations and for small length
frames, the basic access method achieves a better throughput
performance than the RTS/CTS case. On the contrary, when
the RTS/CTS mechanism is employed, the throughput is only
slightly influenced regardless of the data rate, the frame size or
the number of stations of the network.
Future work can include the throughput and delay analysis
including the effect of hidden stations. Another possible area of
research is a consideration of an erroneous wireless channel.
V. DELAY ANALYSIS
REFERENCES
Throughput Efficiency
0.8
0.7
0.65
0.6
◆ RTS, C=1Mbps
‘ Basic, C=1Mbps
S RTS, C=5.5Mbps U Basic, C=5.5Mbps
0.55
■ RTS, C=11Mbps
□ Basic, C=11Mbps
0.5
5
10
15
20
25
30
35
40
45
50
Number of stations
Fig. 6
Throughput versus number of stations
Since no hidden nodes are considered, collisions take place
because two or more contending stations choose the same
backoff slot to transmit. The time needed for a frame
transmission is considered to start when a frame becomes head
of the station’s queue and is finalized when an positive
acknowledgement is received.
Assuming that the frame drop probability is very low and
can be neglected, the average frame delay E[D] is given by:
E[D] = E[X] . E[length of a slot time]
(12)
where E[X] is the average number of slot times required for
successfully transmitting a new frame and E[length of a slot
time] is the average length of a slot time.
According to [16], E[length of a slot time] is equal to:
E[length of a slot time] = (1− Ptr ) ⋅σ + Ptr ⋅ Ps ⋅ Ts + Ptr ⋅ (1− Ps ) ⋅ Tc (13)
Moreover, E[X] is equal to:
m −1
E[ X ] = ∑ ( p i ⋅
i =0
Wi + 1
p m Wm + 1
)+
⋅
2
1− p
2
(14)
After some algebra, (14) reduces to:
E[ X ] =
(1 − 2 p ) ⋅ (W + 1) + pW ⋅ (1 − ( 2 p ) m )
2 ⋅ (1 − 2 p ) ⋅ (1 − p )
(15)
[1] A. De Simone and S. Nanda, “Wireless data: Systems,
standards, services,” J. Wireless Networks, vol. 1, no.
3, pp. 241–254, Feb. 1996.
[2] IEEE standard for Wireless LAN Medium Access
Control (MAC) and Physical Layer (PHY)
specifications, ISO/IEC 8802-11:1999(E), Aug. 1999.
[3] HiperLAN- High Performance Radio Local Area
Network Draft ver. 1.1, ETSI, 1995.
[4] L. Kleinrock and F. Tobagi, “Packet switching in radio
channels, Part II—The hidden terminal problem in
carrier sense multiple access and the busy tone
solution,” IEEE Trans. On Communications, vol.23,
pp. 1417–1433,1975.
[5] K.Huang, K.Chen. “Interference Analysis of
Nonpersistent CSMA with Hidden Terminals in
Multicell Wireless Data Networks”, Proc. PIMRC,
Toronto, pp.907-911, 1995.
[6] L. Kleinrock, F.A. Tobagi, “Packet Switching in Radio
Channels: Part I”, IEEE Trans. Comm., Vol.23, pp.
1400-1416, 1975.
[7] V. Bharghavan , A. Demers , S.Shenker , L. Zhang,
“MACAW: a media access protocol for wireless
LAN's”, ACM SIGCOMM Computer Communication,
v.24, p.212-225, 1994.
[8] P. Karn, “MACA - A New Channel Access Method
for Packet Radio”, Proceedings of the 9th ARRL
Computer Networking Conference, Canada, 1990.
[9] G.Bianchi.
L.Fratta,
M.Oliveri.
“Performance
Evaluation and Enhancement of the CSMA/CA MAC
Protocol for 802.11 Wireless LANs”, Proc. PIMRC,
Taipei, pp.392-396, 1996.
[10] B. P. Crow, I. Widjaja, J. G. Kim, and P. T. Sakai.
“IEEE 802.11 Wireless Local Area Networks”, IEEE
Communication Magazine, 1997.
[11] T. Ho, K. Chen. “Performance Analysis of IEEE
802.11 CSMA/CA Medium Access Control Protocol”,
Proc. PIMRC’1996. Taipei, Taiwan, pp.407-411,
1996.
[12] J. Weinmiller, H. Woesner, JP Ebert, A. Wolisz,
"Analyzing and Tuning the Distributed Coordination
Function in the IEEE 802.11 DFWMAC Draft
Standard", Pro. of MASCOT‘96, San Jose, California,
1996.
[13] H.S. Chhaya, S. Gupta, “Performance modeling of
asynchronous data transfer methods of IEEE 802.11
MAC protocol”, Wireless Networks, pp.217-234,
1997.
[14] J. Weinmiller, H. Woesner, J-P. Ebert, A. Wolisz,
“Analysing the RTS/CTS Mechanism in the
DFWMAC Media Access Protocol for Wireless
LANs, IFIP TC6 Workshop Personal Wireless
Communications”, Czech Republic, 1995.
[15] F.Cali, M.Conti, E.Gregori. “Dynamic Tuning of the
IEEE 802.11 Protocol to Achieve a Theoretical
Throughput
Limit”,
IEEE/ACM
Trans.
On
Networking, V8, N6, 2000
[16] G.Bianchi. “Performance Analysis of the IEEE 802.11
Distributed Coordination Function”, IEEE Journal on
Selected Area in Comm. V18, N3, 2000.
[17] H. Wu, Y. Peng, K. Long, S. Cheng, J. Ma,
“Performance of Reliable Transport Protocol over
IEEE 802.11 Wireless LAN: Analysis And
Enhancement”, IEEE INFOCOM'2002.