Daly Analysis for WiMax under balanced and unbalanced traffic conditions in fixed priorities between stations

International Journal of Engineering Science Invention
ISSN (Online): 2319 – 6734, ISSN (Print): 2319 – 6726
www.ijesi.org ||Volume 4 Issue 10|| October 2015 || PP.43-52
www.ijesi.org 43 | Page
Daly Analysis for WiMax under balanced and unbalanced
traffic conditions in fixed priorities between stations
Ramdan A. Kalifa 1
, Nasar Aldian Ambark Shashoa2
1
(Electrical Engineering Department, The High Professional Institute for Comprehensive
Professions Mezda, Libya)
2
(Electrical and Electronics Engineering Department, Azzaytuna University, Libya)
ABSTRACT : Broadband Wireless Access (BWA) systems offer a solution for broadband access and high
data rate transmission of multimedia services with distinct Quality-of-Service (QoS) requirement through a
wireless medium. The IEEE802.16 standard does not specify how to efficiently schedule the traffic related to
different applications in order to meet their specific requirements. In this paper, we present a modified scheme
which is based on a Reservation Priority Access Control (RPAC) using TDD allocation approach, and the
behavior of this reservation scheme is analyzed and evaluated under balanced and unbalanced traffic
conditions in fixed priorities between stations.
KEYWORDS - Broadband Wireless Access, Quality-of-Service, Reservation Priority Access Control,
Downlink Channel
I. INTRODUCTION
The Broadband Wireless Access (BWA) is a technology proposed to offer wireless access to network stations
in a broadband metropolitan area environment. These networks are designed to operate at high data rates and to
deal with several applications, resulting in different types of traffic profiles and demands. Therefore, the system
is required to work with various types of real-time and non-real-time service classes, with different traffic
characteristics and quality of service (QoS) guarantees. In [1] , a new MAC scheme for BWA, incorporating a
scheduling mechanism based on message and/or station priorities, was proposed as an alternative protocol to the
IEEE 802.16 Standard [3]. However, only a fixed priority approach to service differentiation among terminals
was used.In this paper, we propose an extention to the MAC protocol described in [1], with a unbalanced traffic
conditions in fixed priorities between stations in the network. Moreover, we developed a simulation model to
compare the analytical results under balanced and unbalanced traffic conditions and the simulation results
exposed here.The remainder of this paper is organized as follows. Section II gives a short description of IEEE
802.16 standard.The proposed protocol is described in Section III. An analytical model, which allow to obtain
the average message waiting time for different priority classes, is provided in section IV. Section V presents
some numerical results and the paper is concluded with a few discussions concerning the proposal in Section VI.
II. IEEE 802.16 STANDARD
A- PHY and MAC Layers
In the basic architecture there are one Base Station (BS), and one or more Subscriber Stations (SSs).
Transmissions are assumed to take place through two independent channels: a Downlink Channel (DL) from the
BS to the SSs, and an Uplink Channel (UL) from the SSs to the BS. Hence, there is no contention associated
with the the DL channel, while the UL channel must be shared by the SSs through the use of some multiple
access control protocol. During the DL, only the BS transmits in broadcast to all the SSs. The BS determines the
number of slots to be allocated for each SS in the UL, and broadcast this information in an UL-MAP message at
the beginning of each frame. The stations transmit their data in predefined time slots as indicated in the UL-
MAP. A scheduling module for the UL is necessary to be kept in the BS in order to determine the transmission
opportunities using the bandwidth requests (BW-Request) sent by the SSs. Figure 1 illustrates the structure of
the MAC frame.

Daly Analysis for WiMax under balanced…
Fig.1. MAC frame structure
B- QoS Architecture
The IEEE 802.16 supports many traffic types (data, voice, video) with different QoS requirements. The standard
defines four types of data flows, each one associated with distinct applications and QoS requirements [2]:
1. Unsolicited Grant Service (UGS): for applications with constant bandwidth allocation requirements.
2. Real-Time Polling Service (rtPS): for applications with specific bandwith requirements and maximum
acceptable delay.
3. Non-Real-Time Polling Service (nrtPS): for applications with a minimum bandwidth allocation requirements,
that are intolerant to delay.
4. Best Effort Service (BE): for applications without bandwidth allocation requirements, that receive the
remaining bandwidth after the allocation to the three previous types of services. Fig. 2 shows the QoS
architecture present in 802.16. The UL packet scheduling (UPS) module controls all the packet transmissions in
the UL. As the protocol is connection-oriented, the application should establish a connection between the BS
and the associated service flow (UGS, rtPS, nrtPS or BE). The BS identifies the connections by assigning a
unique Connection ID (CID) to each one. The 802.16 defines the signaling process for the establishment of a
connection (Connection-Request and Connection-Response) between SS and BS, but does not specify the rules
for admission control.
Fig.2 QoS Architecture of IEEE802.16

In summary, the IEEE 802.16 specifies: the signaling mechanism for information exchange between the BS and
the SSs, as the connection configuration, BW-Request and UL-MAP; and the scheduling of UL for UGS traffic.
The standard does not define: the scheduling of UL for rtPS, nrtPS and BE services; admission control and
traffic policing.
III. PROPOSED PROTOCOL
In the paper, we use the TDD allocation scheme, and we will compare result for unbalanced traffic conditions
with the results of the balanced traffic condition which were achieved in [1]. the authors proposed a new MAC
protocol for IEEE802.16 that uses an access scheme called RPAC (Reservation-Priority Access Control),
described in [3]. This access scheme incorporate a traffic scheduling mechanism based on messages and/or
stations priorities, and a reservation period governed by a TDMA discipline with one time slot allocated per
station in the network. The difference between the protocols described in [1] and [3] is that, in the former, the
access scheme was reformulated and adapted to 802.16 MAC structure. The MAC frame structure for the
proposed protocol is illustrated in Figure 3. Differently from the 802.16 standard, the frame lengths are not fixed
in the proposed scheme. Basically, the length of each frame will depend upon the number of packets arriving to
the stations in the previous frame. Reservation periods are located at the end of the UL sub-frame (not at the
beginning as in 802.16). Those periods are used by each of the stations to inform the BS about the services for
which a bandwidth reservation is being requested, as well as the number of packets to be transmitted for each of
those services. After processing all the requests, the BS sends, in the DL sub-frame of the next frame, a UL-
MAP with transmission opportunities to all reserved stations.
Fig 3. MAC Frame Structure for the Presented Scheme [1]
For analysis of the protocol, the activity in the channel can be seenas a sequence of reservation, downlink, and
transmission intervals, where each pair of DL and UL constitutes a transmission cycle, as illustrated in Figure 4.
It is important to note that there is a difference between the MAC frame and the transmission cycle, despite both
having an equal size, because the reservation period is fixed. In fact, the n-th cycle is formed by the reservation
period of (j−1)-st frame, plus the downlink and transmission intervals of the j-th frame. With this definition of
the transmission cycles, the analytical approach described in [3] can be used for the analysis of the average
waiting-time of the messages, as will be seen in Section IV. Referring to the picture in Figure 4, we
define 𝐿 𝑛
𝑅
, 𝐿 𝑛
𝐷𝐿
𝑎𝑛𝑑 𝐿 𝑛
𝑇
respectively, as the lengths (or duration, given by the number of slots) of the
reservation, the downlink, and the transmission intervals in the n-th cycle. Thus, 𝐿 𝑛 = 𝐿 𝑛
𝑅
+ 𝐿 𝑛
𝐷𝐿
+ 𝐿 𝑛
𝑇
represents the total length of the n-th cycle. Following the TDMA protocol, each reservation period is composed
by M slots ( 𝐿 𝑛
𝑅
= M; n = 1, 2 ...), where M is the number of stations in the network. During this period, of
duration of Mτ seconds, each station is associated to a single slot in a fixed manner.
Fig 4. Consecutive Transmission Cycles.

We note that the length of a current (say, the n-th) cycle will depend on the number of messages that arrived
during the previous (say, the (n − 1)-st) cycle. This happens because request for the messages that arrived
during the (n−1)-st cycle will be transmitted in the reservation period of the n-th cycle. After this period, the BS
performs the centralized processing of the transmission opportunities and sends a UL-MAP in the DL sub-
channel, still in the n-th cycle. Afterwards, the stations transmit their messages in the transmission interval of
the same cycle, following the priorities established in the UL-MAP. Therefore, messages arriving during an
ongoing cycle get transmitted only in the subsequent cycle. We propose a medium access protocol with
priorities based on messages and/or stations, in accordance to the 802.16 protocol, that uses GPC or GPSS
admissions. We assume that, following the reservation phase and the DL sub-frame, the channel is allocated to
the stations following the sequence 1, 2, 3, ...,M. Thus, according to the priority rules used to determine the
order in which the messages should be transmitted during the transmission period, the following versions of the
protocol are considered:
• Version I, in which, for any p, q 2 {1, ..., P} such that p < q, all the class-p messages are transmitted before
any messages of class q, independent of which station it belongs to. For messages belonging to the same class
but in distinct stations, the order of the transmissions is according to the order in which the stations access the
channel (first station 1 and last station M). For messages in the same station with the same class of priorities, the
transmissions occur by order of arrival.
• Version II, in which, for any i, j 2 {1, ...,M} such that i < j, all the messages in station i are transmitted before
any message in station j, independently of it„s priority class. In any terminal, the messages are transmitted in
accordance with their priority classes and in order of arrival, in the case of belonging to the same class; that is, at
each station, the priority discipline HOL (Head-Of-the-Line) is applied with the highest priority assigned to
class 1 and the lowest assigned to class P. The behavior of the channel, according to Versions I and II of the
proposed protocol, is illustrated in Figures 5 and 6, respectively. Note that, in Version II, a higher priority
station transmits all it„s messages before one of lower priority. So, unlike Version I, it is possible that messages
of lower priority are transmitted before messages with higher priorities. We observe that a rigid admission
control carried by the BS and also by the SSs is necessary so that the heavy traffic of a specific class (or station)
does not overload (and hog) the channel, affecting the response time of the others.
Fig. 5. Version I of proposed protocol.
Fig. 6. Version II of proposed protocol.

IV. ANALYTIC MODEL
The technique used to obtain the average waiting-time for the messages is similar to the method used in [4]. The
analysis for Version I of the proposed scheme is introduced first, and then obtained results are adopted and used
to extend our analysis to cover Version II. As defined in [1], the system considered has one BS and M (M ¸1)
client stations (SSs), each of which has an infinite buffer-size and is already associated with the base station.
The transmission channel is assumed to be error-free, with a transmission rate equal to C bit/s. In general, the
messages generated at each station are composed of a random number of fixed units of data, called packets, each
of which containsµ−1
bits. The transmission time of each packet is made equal to a time slot (τ). Therefore τ =
1/( µC). Messages arriving at each station belong to one of the different P classes and we assume that class-1
messages have the highest priority and class-P messages the lowest. Moreover, the arrival of messages is
characterized by a Poisson point process, such that 𝜆𝑖
𝑝
(messages per slot) is the average arrival rate of class-p
messages to station i.
Fixed Priorities
1. Version I: As the protocol defined in [1] uses the reservation scheme of [3], the results from the later paper
were used to obtain the expression for the steady-state average waiting-time of class-p messages at station i, in
the former one:
 
   
  2
1
2
1
2
21
1 1
1
1











   

 


LE
LEb
DLEMW
p
j
M
g
i
j
p
j
j
g
p
i
p
ip
i 

(1)
Where 𝜌𝑖
𝑝
= 𝜆𝑖
𝑝
𝑏𝑖
𝑝
is the traffic in the terminal i due to the messages of class p, and
   DL
n
n
LEDLE

 lim . This expression is still given as a function of E [L] and E [𝐿2
], respectively,
the first and second steady-state moments of the cycle length:
 
 



1
DLEM
LE (2)
Where
1
1 1
  
 
P
p
M
i
p
i

and
(3)
2) Version II: The analysis for Version II follows in a direct manner noting that, the messages are transmitted in
the same order as in Version I, with the classes of the messages exchanged for the numbers of the stations and
vice-versa (see Section III). Therefore, the expression in the Version II is analogous to that in Version I,
changing only the M by the P and the i for the p:
(4)
with E[L] and E[L2] given by equations (2) and (3), respectively.

V. RESULTS
An evaluate the level of differentiation obtained with the described schemes, we considered two distinct
scenarios, where in each scenario there is a differentiated probability between four classes of traffic (P = 4), as
shown in Table I.
Table I, Traffic Scenarios Used
To simulate the presented models under balanced and unbalanced traffic conditions, we assume that:
1. Balanced case: there are 10 stations (M=10), each station shares with a traffic percentage (ρi = ρ/10), and
each station contributes with its traffic class according to Table I.
2. Unbalanced case: similarly M=10, but 72% of the traffic from station “5” and the remaining 28% of the
traffic from other stations (3% from each station), each station has traffic type as illustrated in Table I.
We assume that the number of packets in each message of class p in the station i is constant with average
bip = 5 and b2,ip = 25, for each p = 1, 2, 3, 4; and i = 1, ..., 10. So, the average waiting-time for the
class p is given by:
p
i
i
p
p
ip
WW 


10
1


(5)
Where
p
i
p
i
p
i
b 
and
p
i
p
M  
A. Average Waiting-time vs. Traffic Intensity
The performance of the model is measured for the average waiting-time per class. Fig. (7) and Fig. (8)
illustrate the average waiting-time in the queue for each class of priorities related to the offered traffic in the
channel using the TDD allocation schemes.
(a) Version I
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
50
100
150
200
250
300
Traffic Intensity
AverageWaiting-Time
class1 balan
class2 balan
class3 balan
class4balan
class1 unbalan
class2 unbalan
class3 unbalan
class4 unbalan
Traffic
classes
Scenario I Scenario II
Class 1 25% 15%
Class 2 25% 20%
Class 3 25% 30%
Class 4 25% 35%

(b) Version II
Fig7. Average Waiting-Time vs. Traffic intensity in Scenario1 under Balance and Unbalance
Case (a) Version I (b) Version II with M=10
(a) Version I
(b) Version II
Fig8 .Average Waiting-Time vs. Traffic Intensity in Scenario-2 under
Balance and Unbalance Case for (a) Version I (b) Vversion II with M=10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
100
200
300
400
500
600
700
Traffic Intensity
AverageWaiting-Time
class1balance
class2balance
class3balance
class4balance
class1unbalance
class2unbalance
class3unbalance
class4unbalance
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
50
100
150
200
250
300
Traffic Intensity
AverageWaiting-Time
class 1unbala
class2 unbala
class3 unbala
class4 unbala
class1 bala
class bala
class3 bala
class3 bala
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
100
200
300
400
500
600
700
800
Traffic Intensity
AverageWaiting-Time
class1 unbala
class2 unbala
class3 unbala
class4 unbala
class1bala
class2 bala
class3 bala
class3 bala

Both Figures that the average waiting-time of class-p message versus traffic intensity is completely equals for
both balanced and unbalanced traffic conditions. This occurs because in these models, we deal with over all
traffic intensity which equals to the sum of traffic intensities from all SSs. From fig (7) and fig (8) we note that
the average waiting-time increases with the increase in traffic intensity, especially at heavy traffic condition. It
can also notices that the differentiation between classes is clearly in Version I than Version II because in
Version I, the priorities between classes prevail over the priorities between stations, while in Version II Where
the priorities between the stations superpose. It observes that the waiting-time for the traffic of higher priority
(class 1). For Version I in both scenarios, we observe that the waiting-time for each class depends on traffic
intensity of this class and traffic intensities of all classes which have higher priority than this class
B. Average waiting- time versus station index
Fig. (9) and Fig. (10) Present the average waiting-time in the stations queue for traffic intensity (ρ= 0.9) for
balanced and unbalanced traffic conditions. In this way, the average waiting-time in the queue for the station i is
given by (6), where λi represents the message rate in the station-i.
p
i
p i
p
i
i WW 


4
1


(6)
(a) Version I
(b) Version II
Fig. 9. Average Waiting-Time vs. Station Index in Scenario-1 under
Balanced and Unbalanced Cases (a) Version I. (b) Version II
1 2 3 4 5 6 7 8 9 10
0
200
400
600
800
1000
1200
1400
1600
1800
Station Index
AverageWaiting-TimeperClass(Slots)
class1 bala
class2bala
class3bala
class4bala
class1 unbala
class2unbala
class3 unbala
class4 unbala
1 2 3 4 5 6 7 8 9 10
0
500
1000
1500
2000
2500
3000
3500
StationIndex
class1bala
class2bala
class3bala
class4bala
class1unbala
class2unbala
class3unbala
class4unbala

(a) Version I
(b) Version II
Fig. 10. Average Waiting-Time vs. Station Index in Scenario-2 under
Balanced and Unbalanced Cases (a) Version I. (b) Version II
Form Figures (9) and (10) it can be observe that for balanced and unbalance traffic conditions, the average
waiting-time increases uniformly with the increase in station index. But the increases in waiting-time with the
increases in station index are clearly in unbalance case than balance case.
VI. CONCLOSTION
In this paper, a modified MAC reservation scheme is proposed for BWA systems which incorporate traffic
scheduling functions with message-or station-based priorities Moreover, an analytical model is devised for the
mean message waiting-time for the different traffic scenarios, using TDD allocation scheme, and assuming
Poisson arrivals of messages and general distributions for message lengths. According to the results exposed in
this work , we can conclude that the maximum difference in waiting-time between stations for Version I is
smaller than that for Version II , for the balance and unbalance traffic conditions, the average waiting-time of
class-p versus traffic intensity is completely equals. Systematic increase in waiting-time is observed as the
station index increase. This obvious clearly in unbalance traffic case than balance traffic case.
1 2 3 4 5 6 7 8 9 10
0
500
1000
1500
2000
2500
Station Index
class1 unbala
class2 unbala
class3 unbala
class4 unbala
class1 bala
class2 bala
class3 bala
class4 bala
1 2 3 4 5 6 7 8 9 10
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
Station Index
class1 unbala
class2 unbala
class3 unbala
class4 unbala
class1 bala
class2 bala
class3 bala
class4 bala

REFERENCES
[1] L. F. M. de Moraes and P. D. Maciel Jr., “Analysis and Evaluation of a New MAC Protocol for Broadband Wireless Access,” in
IEEE WirelessCom 2005 - International Conference on Wireless Networks,Communications, and Mobile Computing, Maui,
Hawaii, June 2005.
[2] L. F. M. de Moraes and A. N. L. Valverde, “Waiting-Time Analysis of a Reservation Access-Control Scheme with message-
based priorities,” Third IFIP TC 6/wg 7.3 Inter. Conf. on Data Communication Systems and Their Performance, June 1987.
[3] IEEE 802.16, IEEE Standard for Local and Metropolitan Area Networks- Part 16: Air Interface for Fixed Broadband Wireless
Access Systems,IEEE Std. 802.16, Oct. 2004.
[4] L. F. M. de Moraes and P. D. Maciel Jr., “Proposta e Avaliac¸ ˜ao de um Novo Protocolo de Acesso ao Meio para Broadband
Wireless Acess,” in XXII Brazilian Symposium of Telecommunications - SBrT‟05, Campinas, SP - Brasil, Sept. 2005.
[5] I. Rubin and L. F. M. de Moraes, “Message Delay Analysis for Polling and Token Multiple-Access Schemes for Local
Communication Networks,” IEEE JSAC, vol. SAC-1, no. 5, Nov. 1983.
[6] L. F. M. de Moraes and A. N. L. Valverde, “Waiting-Time Analysis of a Reservation Access-Control Scheme with message-
based priorities,” Third IFIP TC 6/wg 7.3 Inter. Conf. on Data Communication Systems and Their Performance, June 1987.

Daly Analysis for WiMax under balanced and unbalanced traffic conditions in fixed priorities between stations

More Related Content

What's hot (17)

Viewers also liked (13)

Similar to Daly Analysis for WiMax under balanced and unbalanced traffic conditions in fixed priorities between stations (20)

Recently uploaded (20)

Daly Analysis for WiMax under balanced and unbalanced traffic conditions in fixed priorities between stations