International Journal of Scientific & Engineering Research, Volume 2, Issue 2, February-2011 1

ISSN 2229-5518

Comparative Study of Scheduling Algorithms in

WiMAX

Mohammed Sabri Arhaif

Abstract— In recent years, telecommunication operators are constantly seeking more efficient wireless broadband service, while tele- communication technology is continuously upgrading its access network technologies to cope with the high demands for high-speed in- ternet access and multimedia service by end-users. W iMAX seems to be the solution as it is able to provide easy deployment, high speed data rate and wide range coverage. Most importantly, W iMAX provides Quality of Service (QoS) that can support all kinds of real-time application in wireless networks that includes priority scheduling and queuing for bandwidth allocation that is based on traffic scheduling algorithms within wireless networks.

This paper aims to evaluate the implementation of the various types of scheduling algorithms of W iMAX wireless network technolo-

gy namely: Diffserv-Enabled (Diffserv), Round Robin (RR), Self-Clocked-Fair (SCF), Strict-Priority (SP), Weighted-Fair Queuing (W FQ) and Weighted-Round Robin (W RR). A detailed simulation study via the QualNet 5.0 simulator evaluation version was carried out with the aim to analyze and evaluate the performance of each scheduler to support the different QoS classes. The results of the simulation showed that effective scheduling algorithm can provide high service standards to support the QoS requirements to meet the different types of demands by the various end-users.

Index Terms— Scheduling Algorithms, W iMAX, QoS, QualNet

1 INTRODUCTION

—————————— • ——————————
he demand for high speed broadband wireless sys- tems, internet access and multimedia service has in- creased tremendously as these applications are used
in all sectors; trade and commerce, education and re- search, communications, and even leisure and entertain- ment. Consequently, the need for broadband wireless access (BWA) has grown significantly due to the increase in the number and types of users. Due to their mobility and need for data access at all times, an efficient broad- band connectivity is much sought after. Hence, WiMAX, (Worldwide Interoperability for Microwave Access) which is a trade name used to group a number of wireless technologies have emerged from IEEE (Institute of Elec- trical and Electronics Engineers) to meet the demands of the various end-users. It is deployed to serve all the end- users. Moreover WiMAX technology is based on a Stan- dard that is IEEE 802.16 which is (BWA) that offers mo- bile broadband connectivity.
WiMAX provides Quality of Service (QoS) that sup- ports five different categories of services namely: Unsolic- ited grant services (UGS), Real-time polling services (rtPS), Non- real-time polling service rate (nrtPS), Ex- tended real-time polling service (ertPS) and Best-Effort services (BE). As such, scheduling class services must ensure there is efficiency and fairness in meeting the vari- ous QoS requirements.
The scheduling class services in wireless networks in- cludes priority scheduling and queuing for bandwidth allocation based on traffic scheduling algorithms within wireless networks. Since the scheduling algorithm is still
an undefined territory, designing an efficient scheduling algorithm that can provide high throughput with mini- mum delay is indeed a challenging task for system devel- opers.
Although there are various studies on scheduling algo- rithms, there is a clear absence of a comprehensive per- formance study that provides a unified platform for com- paring such algorithms. Therefore, this research paper is aimed to investigate and compare several scheduling al- gorithms in terms of performance and abilities to support multiple classes of service. Besides that, the paper intends to identify significant scheduling algorithms for the Up- link and Downlink channels that use QualNet-5.0. Finally it aims to measure the important metrics of the schedul- ing algorithms.

1.1 WiMAX Architecture

The basic IEEE 802.16 architecture consists of one Base Station (BS) and one or more Subscriber Station (SS). BS acts as a central entity to transfer all the data from SSs through two basic operational modes: mesh and point-to- multipoint (PMP). Meanwhile, transmissions take place through two independent channels: Downlink Channel (from BS to SS) and Uplink Channel (from SS to BS). The Uplink Channel is shared among all SSs, while the Down- link Channel is used only by BS [1].

International Journal of Scientific & Engineering Research, Volume 2, Issue 2, February-2011 2

ISSN 2229-5518

In the mesh mode, subscriber stations (SS) can com- municate with each other as well as with the base station (BS). This means that traffic can be routed through other SSs. Also the traffic can occur directly among SSs. There- fore, within the mesh mode, uplink and downlink chan- nels are defined as traffic in the direction to and from the BS, respectively.
In the PMP mode, the SSs are only allowed to communicate through the BS. In this way, the provider can control the environment to ensure that the Quality of Service (QoS) meets the requirements of its customers. In the PMP mode, traffic only occurs between the Base Station (BS) and Subscriber Stations (SS).

1.2 WiMAX Quality of Services

WiMAX standard defines 5 service classes to support its wide range of applications as endorsed by IEEE 802.16.

1.2.1 Unsolicited grant services (UGS):

This class of service is designed to support fixed-sized data packets at a constant bit rate (CBR) such as E1/T1 lines that can sustain real-time data stream applications. This service provides guaranteed throughput, latency and jitter to the necessary levels as TDM services. UGS is used mainly to support Constant Bit Rate (CBR) services found in voice applications such as voice over IP [2,3,4,5,6].

1.2.2 Real-time Polling Services (rtPS):

This class of service is designed to support real-time ser- vice flow that generates variable-sized data packets on a periodic interval with a guaranteed minimum rate and guaranteed delay. The mandatory service flow parame- ters that define this service are inclusive of minimum re- served traffic rate, maximum sustained traffic rate, maxi- mum latency and request/transmission policy. rtPS is used extensively in MPEG video conferencing and streaming [2,3,4,5,6].

1.2.3 Non-real-time Polling Service (nrtPS):

This class of service is designed for non-real-time traffic with no delay guaranteed. The delay tolerant data stream consists of variable-sized data packets. The applications provided by this service are time-insensitive and a mini- mum amount of bandwidth. This service is especially suitable for critical data application such as in File Trans- fer Protocol (FTP) [2,3,4,5,6].

1.2.4 Extended real-time Polling Service (ertPS):

This class of service provides real-time applications which generate variable-sized data packets periodically that re- quire guaranteed data rate and delay with silence sup- pression. This service is only defined in IEEE 802.16e-
2005. During the silent periods, no traffic is sent and no bandwidth is allocated. However, there is a need to have a BS poll during the MS to determine the end of the silent periods. ertPS is featured in VoIP with silence suppres-
sion [2,3,4,5,6,7].

1.2.4 Extended real-time Polling Service (ertPS): This class of service provides real-time applications which generate variable-sized data packets periodically that require guaranteed data rate and delay with silence suppression. This service is only defined in IEEE 802.16e-

2005. During the silent periods, no traffic is sent and no
bandwidth is allocated. However, there is a need to have
a BS poll during the MS to determine the end of the silent
periods. ertPS is featured in VoIP with silence
suppression [2,3,4,5,6,7].

1.2.5 Best-Effort Services (BE):

This class of service provides support for data streams whereby no minimum service-level guarantee is required. The mandatory service flow parameters that define this service include maximum sustained traffic rate, traffic priority and request/transmission policy. BE supports data streams found in Hypertext Transport Protocol (HTTP) and electronic mail (e-mail) [2,3,4,5,6].

2 SHEDULING ALGORITHMS

The main focus of this research study is to examine the scheduling schemes in WiMAX network. In order to spe- cify high network performance, an efficient scheduling algorithm is essential as it manages and controls the pro- vision of an efficient level of QoS support.
Although many scheduling algorithms have been pro- posed in the literature for WiMAX network, the design of the algorithms are challenged by having to support dif- ferent levels of services, fairness and implementation complexity. Many researchers have compared their pro- posal schemes on different scheduling schemes, but there is no common, simple and standardized packet schedul- ing to make their comparisons with.
In this study, six carefully selected scheduling algo- rithms in WiMAX wireless network are investigated. These algorithms which are considered the most domi- nant and popular include Diffserv-Enabled (Diffserv), Round-Robin (RR), Self-Clocked-Fair (SCF), Strict-Priority (SP), Weighted-Fair Queuing (WFQ) and Weighted Round Robin (WRR). Furthermore, these common packet scheduling schemes provides QoS support for real time applications in IEEE 802.16 system.

2.1 Diffserv-Enabled: Diffserv is a simple, scalable and measurable mechanism for classifying and managing network traffic. Besides, it provides low-latency with guaranteed service to critical network traffic as well as to non-critical services. It relies on the principle of traffic classification by involving the 6-bit Differentiated Servic- es Code Point (DSCP) field in the header of IP packets to classify the packet and indicate the per-hop behavior (PHB). DSCP replaces the outdated IP precedence in clas- sifying and prioritizing types of traffic. Every router on the Diffserv network is configured to differentiate traffic

International Journal of Scientific & Engineering Research, Volume 2, Issue 2, February-2011 3

ISSN 2229-5518

based on class so that each traffic class can be managed differently, ensuring preferential treatment for higher- priority traffic on the network [8].

2.2 Round-Robin (RR): It is designed for a time-sharing system whereby the scheduler assigns time slots to each queue in equal portions without priority. It starts with the highest priority queue with packets, services a single packet, then visits the next lower priority queue with packets, and continues servicing every single packet from each queue. This is carried on until each queue with packets has been serviced once. Every queue is allocated with the same portion of system resources regardless of the channel condition, ultimately utilizing the same re- sources. However, the RR scheduler has the same band- width efficiency as a random scheduler, so it cannot guarantee different QoS requirements for each queue [9,

14].

2.3 Self-Clocked-Fair (SCF): It is an efficient queuing scheme which satisfies the quality of services (QoSs) in broadband implementation. The algorithm is based on the concept of virtual time that adopts the concept of an internally generated virtual time as the index of work in progress. It links virtual time to the work progress in the fluid-flow fair queuing (FFQ). As virtual time function is involved in determining the order of which packet should be served next, the virtual time that is produced depends very much on the progress of work in the actual packet- based queuing system. This scheme is efficient for the internal generation of virtual time as it involves negligible overhead. This is because virtual time is easily computed from the packet situated at the head of the queue. In addi- tion, the SCFQ algorithm can accomplish easier imple- mentation and it can maintain the fairness attribute in virtual time function. [10, 11,15].

2.4 Strict-Priority (SP): In Strict-Priority algorithm, the selection order is based on the priority of weight order. The packets are first categorized by the scheduler de- pending on the quality of service (QoS) classes and then allocated into different priority queues. The algorithm services the highest priority queue until it is empty, after which, it moves to the next highest priority queue. Thus, strict-priority algorithm may not be suitable in WiMAX network. This is because there is no compensation for inadequate bandwidth. Also this technique is only ap- propriate for low-bandwidth serial lines that currently uses static configuration which does not automatically adapt to changing network requirements. Finally, this process may result in bandwidth starvation for the low priority QoS classes whereby the packets may not even get forwarded and no guarantee is offered to one flow [6].

2.5 Weighted-Fair-Queuing (WFQ): This algorithm is employed for uplink traffic in WiMAX with different size packets. As it caters to different size packets, it emphasiz-

es on providing fair scheduling for the different flows by assigning finish times to the packets. The finish times are based on the size and weight of the packets. In general, the WFQ algorithm outperforms the WRR due to variable size packets. However, the weaknesses of WFQ algorithm are, the start time of a packet is not taken into considera- tion, and it can lower the scheduler system if many pack- ets occur in the priority region [12, 13].

2.6 Weighted Round Robin (WRR): It is a scheduling algorithm implemented for resource sharing in a comput- er or network. In fact, WRR is an extension of the Round Robin (RR) algorithm. In a network, WRR serves a num- ber of packets that are computed by normalizing weight of data divided by the average of packet size from non- empty connection queue. It begins by classifying packets into a variety of service classes followed by assigning a queue that is determined by the different percentage of bandwidth. Finally, it is serviced in round robin order. Since the bandwidth is assigned according to the weights, the algorithm will not provide good performance in the presence of variable size packets. However, WRR method makes certain that all service classes have access to at least some configured amount of network band width to avoid bandwidth starvation [2,6,7].

3 SIMULATION MODEL

The purpose of this simulation study is to investigate and evaluate different types of scheduling techniques in order to determine the one that is most efficient in WiMAX network. The simulations are performed using QualNet simulation. This simulation provides an intuitive model set up capability that includes core components such as animator, packet tracer analyzer, protocol designer and protocol stack.

FIG.1.SYSTEM MODEL IMPLEMENTATION BY QUALNET

The system parameter used in this simulation study con- sists of a single cell with a BS, and a number of MS that

International Journal of Scientific & Engineering Research, Volume 2, Issue 2, February-2011 4

ISSN 2229-5518

varies from 10 to 50 MS. Table 1 summaries the simula- tion parameters used in the experiments with 20MS.

TABLE 1.SIMULATION PARAMETERS

Parameter Value
BS range radius 1000
MS range radius 500
Frequency band 2.4
Channel bandwidth 20
Frame duration 20
FFT size 2048
Number of MS 10-50
Number of BS 1

0.18

0.16

0.14

0.12

0.1

0.08

0.06

0.04

0.02

Average Jitter vs. various number of MS

DI FFSERV RR

SCF

SP WFQ WRR

10 20 30 40 50

Number of MS

BS transmit power 20/5

EXPERIMENT 1: FIGURE 4.1: THE AMOUNT OF END-TO-END

DELAY

MS transmit power 15/1.5

Services types (QoS) BE, nrtPS, rtPS, ertPS, UGS

4.5

3.5

Average End-to-End Delay vs. various number of MS

DIFFSERV-ENABLED ROUND-ROBIN

SELF-CLOCK ED-FAIR STRIC T-PRIORI TY

WEIGHTED-FAIR

Simulation time 30

3 WEIGHTED-ROUND-ROBI N

2.5

4 SIMULATION RESULTS

Six experiments with varying simulation parameters were carried out and the findings show varying results.
The results of experiment 1 are shown in Figure 4.1 that SP, WRR, and WF are the best scheduling techniques in WiMAX network with respect to the end-to-end time de-

1.5

0.5

10 20 30 40 50

Num ber of MS

lay.
The results of experiment 2 are shown in Figure 4.3 in-
dicate that there is much difference between all algo-
rithms when the number of mobile stations (MS) is small
(10MS). This happens as MS produces the shortest
amount of time for packet latency. Another result ob-
tained is RR outperforms the other techniques when the
number of MS becomes more (20-50MS). The results also
indicate that SCF performs better than Diffserv, WRR, SP,
and WFQ when the number of mobile stations (MSs) is increased (30-40MS).
The result of experiment 3 are shown in Figure 4.5 clearly shows that RR technique has achieved the highest value of throughput for different numbers of MS (20-50) compared to the other five techniques. However, RR technique shows the same amount of throughputs as the

EXPERIMENT 2: FIGURE 4.3: THE PACKET LATENCY (JITTER)

Average Throughput vs. various number of MS

250

DI FFSERV-ENABLED

ROUND-ROBI N

SELF-CLOCKED-FAI R

200

STRI CT-PRI O RI TY

WEIGHTED-FAIR

WEI GHTED-ROUND-RO BI N

150

100

others when the number of MS ranges between 10-20. Furthermore, most of the algorithms have the same per- formance when the numbers of MS are fewer than 30.

10 20 30 40 50

Number of MS

EXPERIMENT 3: FIGURE 4.5: THE THROUGHPUT OF NETW ORK

International Journal of Scientific & Engineering Research, Volume 2, Issue 2, February-2011 5

ISSN 2229-5518

Meanwhile the results tabulated in Figure 4.2 shows that Diffserv has the lowest performance in producing the highest amount of end-to-end delay time. On the other hand, WF shows the best performance as the average end-to-end time delay has the lowest reading. Finally, it can be concluded that there is much difference in terms of the average end-to-end delay time among RR, SCF and WRR.
From the figure 4.4, it is noted that RR technique shows the most favourable results as the average jitter has low reading (0.124s), while Diffserv shows the most unfa- vourable result as the average jitter has higher reading (0.137s). The results also show that WF and WRR produce

0.138

0.136

0.134

0.132

0.13

0.128

0.126

0.124

0.122

0.12

0.118

Overall Average Jitter

The Scheduling algorithms

almost the same amount of average jitter (0.136s). How- ever, there is no big gap between the two algorithms, WF and WRR, in terms of overall average jitter and SP.
From the figure 4.6, it is noted that RR algorithm out-
performs the other five algorithms in terms of the overall
throughput 125Kbps. The results of the experiment shows
that WRR is a poor scheduling technique as it produces
the lowest amount of average throughput 100Kbps. Diff-
serv, SCF, and SP produce almost the same amount of
overall average for the throughput 110Kbps, while WF is
ranked after these algorithms as the average throughput
is 103Kbps.

FIGURE4.4: THE AVERAGE OF PACKET LATENCY (JITTER)

Overall Average of Throughput

140

120

100

2.5

Average End-to-End Delay

The Schedling algorithms

FIGURE 4.6: THE OVERALL AVERAGE OF THROUGHPUT

1.5

0.5

The Scheduling algorit hms

The results of experiment 4 are shown in Figure 4.7 that WF outperforms SP and WRR as it achieves the shortest amount of end-to-end delay time for all the classes of QoS. However, WF achieves the same amount of end-to-end delay time for the class BE and nrtPS. It is also noted that BE achieves the shortest amount of end-to- end delay time for the three algorithms, while UGS pro- duces the longest amount of end-to-end delay time for all

FIGURE 4.2: THE TOTAL AVERAGE OF END-TO-END DELAY TIME

the three algorithms.
The results of experiment 5 are shown in Figure 4.8 the observations of the average throughput for the best three scheduling algorithms; SP, WF, WRR with respect to the classes of the quality of services (QoSs). In Figure 4.7, it is clear that WF is the best algorithm as produces BE, nrtPS, rtPS, ertPS, and UGS class. However, the performance of SP is not favorable compared to the performance of WRR and WF. These three algorithms are selected due to the fact that they achieve the best performance with respect to different scenarios and various numbers of factors in the simulation experiments.
The results of experiment 6 are shown in Figure 4.9 re- veal that the classes of the QoSs positively influence the

International Journal of Scientific & Engineering Research, Volume 2, Issue 2, February-2011 6

ISSN 2229-5518

percentage of the fairness index for each scheduling tech- nique. It is also noted that RR scheduling technique with QoS achieves the best percentage of fairness index, while WRR with QoS shows the highest percentage of fairness index. However, RR technique with no QoS achieves higher percentage of fairness index, while WRR with no QoS shows the lowest percentage of fairness index. Final- ly, from Figure 4.8, it can be concluded that the QoS class services have a high impact on the percentage of the fair- ness index.

0.35

0.3

0.25

0.2

0.15

0.1

0.05

Fluctuation of Intra-class Fairness

wi th QoS

wi th no QoS

The Scheduling algorithms

The Best Scheduling Algorithms with respect to QoS Class Services

EXPERIMENT 6: 4.9: THE FLUCTUATION OF INTRA-CLASS FAIR- NESS

0.45

0.4

0.35

0.3

0.25

0.2

0.15

0.1

0.05

STRICT-PRI ORI TY W EIGHTED-FAIR WEIGHTED-ROUNDROBI N

BE nrtPS rtPS ertPS UGS

Classe s of QoS

6 CONCLUSIONS AND FUTURE WORK

In conclusion, the investigation of the behaviors of several wireless scheduling algorithms namely Diffserv, RR, SCF, SP, WFQ, WRR has shown the strengths of some of the scheduling algorithms that were under study. One of the best scheduling algorithms is WF, in terms of the amount of end-to-end delay. The other is RR, in terms of packet latency (Jitter). Finally WRR outperforms the rest by pro- ducing the highest rate of throughput of data packet in the network. As to the best scheduling algorithms in terms of the amount of delay time with respect to QoSs classes are WF, SP, and WRR respectively. Finally, it is

EXPERIMENT 4: FIGURE 4.7: THE AMOUNT OF DELAY TIME FOR

THE BEST SCHEDULING ALGORITHMS

Average Throughput of The Best Scheduling Algorithms

92000

90000

88000

86000

clear that there is not a single scheduling scheme that provides superior performance with respect to all the QoS requirements and characteristics of the IEEE 802.16 MAC layer. This is because issues such as the amount of infor- mation required by the grant scheduler at the BS and the allocation of time-slots and sub-channels to different SSs require attention.
As the scheduling in WiMAX wireless network is a challenging topic, future works should include further investigation on scheduling algorithms under different bandwidth request mechanisms and CAC schemes.

84000

82000

80000

78000

BE nrtPS rtPS ertPS UGS

Classes of QoSs

WRR

REFERENCES

[1] Jun Wang, Weijia Jia, and Liusheng Huang, "An Efficient Cen- tralized Scheduling Algorithm for IEEE 802.16 Multi-radio Mesh Networks", ICUIMC '2008 Proceedings of the 2nd inter-

EXPERIMENT 5: FIGURE 4.8: THE AVERAGE OF THROUGHPUT FOR THE BEST SCHEDULING ALGORITHMS

national conference on Ubiquitous information management

and communication, Suwon, Korea, January 31 - February 01,

2008.

[2] Jani Lakkakorpi, Alexander Sayenko and Jani Moilanen. (2008).

Comparison of Different Scheduling Algorithms for WiMAX Base Station Deficit Round-Robin vs. Proportional Fair vs. Weighted Deficit Round-Robin. Proceedings of The Wireless

International Journal of Scientific & Engineering Research, Volume 2, Issue 2, February-2011 7

ISSN 2229-5518

Communications Networking Conference WCNC 2008. 31

March- 3 April 2008, Las Vegas, Nevada, USA, pp.

[3] Chakchai So-In, Raj Jain, and Abdel-Karim Tamimi. (2009).

Scheduling in IEEE 802.16e Mobile WiMAX Networks: Key Is- sues and a Survey. IEEE Journal on selected areas in communi- cations, vol. 27(2). February. pp. 156- 171.

[4] Mikael Gidlund and Gang Wang. (2009). Uplink Scheduling Algorithms for QoS Support in Broadband Wireless Access Networks. Journal of communications. Vol. 4(2). pp. 133-142.

[5] Ashish Jain and Anil K. Verma. (2008). Comparative Study of Scheduling Algorithms for WiMAX. Proceedings of the Nation- al Conference on Mobile and Pervasive Computing, Compc 08,

7-8 August. Chennai, India. pp. 10-13.

[6] Ahmed H. Rashwan, Hesham M. ElBadawy and Hazem H. Ali. (2009). Comparative Assessments for Different WiMAX Sche- duling Algorithms. Proceedings of the World Congress on En- gineering and Computer Science WCECS 2009. Vol. I. 20-22 Oc- tober, 2009. San Francisco, USA. pp. 362-366.

[7] Najah Abu Ali, Pratik Dhrona and Hossam Hassanein. (2009).

A performance study of uplink scheduling algorithms in point- to-multipoint WiMAX networks. Computer communications. vol. 32. pp. 511–521.

[8] Yuxiao Jia, Miguel Lopez Guerrero, Ognian Kabranov, Dimi- trios Makrakis and Luis Orozco Barbosa. (2003). Design and testbed implementation of adaptive MPLS- Diffserv enabled virtual private networks. Proceedings of the IEEE Canadian Conference Electrical and Computer Engineering, CCECE 2003.

4-7 May. Montreal, Canada. pp 965-968.

[9] Jason Nieh, Chris Vaill and Hua Zhong. (2001). Virtual-Time Round-Robin:An O(1) Proportional Share Scheduler. Proceed- ings of the USENIX Annual Technical Conference. 25-30 June. Boston, Massachusetts, USA. pp. 245-259.

[10] S. Jamaloddin Golestani. (1994). A Self-Clocked Fair Queueing Scheme for Broadband Applications. Proceedings of the IEEE Conference on Computer Communications, 13th Annual Joint Conference of the IEEE Computer and Communications Socie- ties, INFOCOM '94, 12-16 June. Toronto, Ontario, Canada. pp.

636-646.

[11] Hassan Halabian, Hossein Saidi and Reyhaneh Changiz. (2008).

LVT-SCFQ: A Modified Self Clocked Fair Queueing Algorithm for Broadband Networks. Proceedings of the 3rd International Conference on Broadband Communications, Information Tech- nology & Biomedical Applications BroadCom 2008. 23- 26 No- vember. Pretoria, Gauteng, South Africa. pp. 175-180.

[12] Richard Kautz, Raymond Keh, Kee Chaing Chua and Alberto Leon-Garcia. (2000). A Distributed Fair Queuing (DFQ) Archi- tecture for Wireless ATM Local Access Networks. International journal of wireless information networks. vol. 7(4). pp. 221-229.

[13] Jean-Philippe Georges, Thierry Divoux and Eric Rondeau.

(2005). Strict Priority versus Weighted Fair Queuing in Switch- ed Ethernet networks for time critical applications. Proceedings of the 19th IEEE International Parallel and Distributed Processing Symposium (IPDPS'05). 4-8 April. Denver, Colora- do, USA. pp. 141- 148.

[14] Jason Nieh, Chris Vaill and Hua Zhong. (2001). Virtual-Time Round-Robin:An O(1) Proportional Share Scheduler. Proceed- ings of the USENIX Annual Technical Conference. 25-30 June. Boston, Massachusetts, USA. pp. 245-259.

[15] Hassan Halabian, Hossein Saidi and Reyhaneh Changiz. (2008).

ABOUT THE AUTHOR

Mohammed Sabri Arhaif

College of Information Technology

Master in IT

Universiti Tenaga Nasional (UNITEN) Selangor – Malaysia

H\P: +60173282194

E-mail: albayaams@yahoo.com

ST20645@student.uniten.edu.my

Was born on June 14, 1982 in Karbala’, Iraq. He obtained first degree in Computer Science and earned his Bachelor of Science/Computer from AL-MAMON University College (Baghdad-Iraq) in 2006.

In 2008, he furthered his studies in the College of Graduate Studies at Universiti Tenaga Nasional (UNITEN). He pursued a Masters degree in Information Technology, specializing in Network in WiMAX. He completed a research on a Comparative Study of Sche- duling Algorithms in WiMAX.