0018-9545 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/TVT.2015.2425960, IEEE Transactions on Vehicular Technology
> VT-2014-01473 <
3
AIFS and CW differentiation under the saturated condition. In
[21], Kong et al. analyzed the throughput performance of
differentiated service traffic for the 802.11e EDCA mechanism
and proposed a recursive method capable of calculating the
mean access delay. In [22], Tao and Panwar proposed an
analytical model to calculate the maximum sustainable
throughput and service delay distribution for each priority class
of the 802.11e EDCA mechanism under saturated load. In [23],
Pan and Wu considered AIFS events among each backoff
procedure of the 802.11e EDCA mechaism and evaluated the
saturation throughput under heterogeneous traffic scenarios. In
[24], Xiang et al. used the signal transfer function of the
generalized state transition diagram to analyze the performance
of the 802.11e EDCA mechanism from a non-saturated channel
to a saturated medium by using an M/G/1/K queuing model. In
[25], Inan et al. proposed an analytical model which
incorporates an accurate CW, AIFS and transmitting
opportunity differentiation of the 802.11e EDCA at any traffic
load and captured the effect of the buffer size at the MAC layer.
In [26], Yan and Pan developed a model to calculate the
sustained probabilistic properties for each prioritized AC of the
802.11e EDCA mechanism under the saturated condition.
However, a common limitation of these models is that they are
based on 3-D Markov Chain. However, 3-D Markov chain
based analytical models introduce unnecessary calculation
complexity and the complexity of the solution for the scenario
with four ACs is prohibitive.
III. PERFORMANCE MODELING FOR THE IEEE 802.11P EDCA
MECHANISM
In this section, we first give a brief overview of the IEEE
802.11p EDCA mechanism and then develop two Markov
models to describe the backoff procedure and the contention
period of an AC queue, respectively. Based on the two Markov
models, we further derive performance models for analyzing
the access performance of the 802.11p EDCA mechanism.
A. Overview of the 802.11p EDCA Mechanism
The EDCA mechanism is a channel access mechanism
designed for supporting different QoS requirements at the
MAC layer. It defines four ACs: Voice-VO, Video-VI, Best
effort-BE, and Back ground-BK. As shown in Figure 1, each
AC has a queue independently contending for transmission
with its own parameters, including
min
CW
,
max
CW
, AIFSN
and the retransmission limit value. Let
AIFS
T denote the
duration of the arbitration inter-frame space (AIFS), i.e., the
idle period after a busy period. Then, we have
SIFSsAIFS
TTAIFSNT +×=
, (1)
where
s
T is the duration of a slot and
SIFS
T is the duration of
the short inter-frame space (SIFS).
In the EDCA mechanism, if a channel is sensed idle when a
packet arrives at an AC queue in a station, and keeps idle for
AIFS
T , the packet will be transmitted. Otherwise, if the channel
is sensed busy, the station will continue to sense the channel. If
the channel becomes idle again and keeps idle for
AIFS
T , it will
initiate a backoff procedure. In this case, the station will first
start up a backoff counter with the initial value set to one
randomly selected from
],0[ W
, where
min
WW = . During the
backoff period, if the channel is sensed busy, the backoff
counter will be frozen at the current value. After the channel
becomes idle and keeps idle for
AIFS
T , the counter will be
resumed. Then if the channel is sensed idle in a slot, the backoff
counter will be decremented by 1. When the backoff counter
becomes zero, the packet will be transmitted. If the station does
not receive an Acknowledgment (ACK) packet in a given time,
the packet will be retransmitted. At each retransmission, the
value of W will first be doubled and then a new backoff
procedure is initiated. After the value of W reaches
max
CW
, it
will keep this. If the retransmission limit is reached, the value of
W will be reset to
min
CW
and a new backoff counter will be
initiated. If the transmission is successful, the value of W is
reset to
min
CW
and then a backoff procedure is initiated. If the
AC queue becomes empty and its backoff counter becomes
zero, the AC queue will wait until it receives a packet. The
access process of the EDCA mechanism for an AC queue is
illustrated in Figure 2. In the case of multiple AC queues, an
internal collision will occur when more than one AC queue in a
station has packets to transmit at the same time. Once an
internal collision occurs, the station will grant the transmission
to the AC queue with the highest priority. Meanwhile, it will
allow the AC queues with lower priorities to backoff and then
transmit.
Backoff Backoff Backoff Backoff
Internal collision handing
Transmission Contention
AC0
queue
AC1
queue
AC2
queue
AC3
queue
Voice
data
Video
data
Best effert
data
-
Back ground
data
Figure 1. Illustration of the EDCA mechanism
Figure 2. Illustration of the EDCA process of an AC queue