Modulation aware connection admission control and uplink scheduling algorithm for wimax networks

International Journal of Wireless & Mobile Networks (IJWMN) Vol. 7, No. 1, February 2015
DOI : 10.5121/ijwmn.2015.7105 75
MODULATION AWARE CONNECTION ADMISSION
CONTROL AND UPLINK SCHEDULING ALGORITHM FOR
WIMAX NETWORKS
Shwetha D1
, Mohan Kumar N M2
and Devaraju J T2
1
Department of Electronics, Tumkur University, Tumkur, India
2
Department of Electronic Science, Bangalore University, Bangalore, India
ABSTRACT
Mobile WiMAX standard defines air interface for mobile broadband wireless access (BWA) systems and
aimed to provide high data rate with seamless mobility while maintaining the quality of service (QoS).
However, QoS of real-time applications depends on the link quality, which necessitate the channel-aware
connection admission control (CAC) and scheduling schemes. In this paper Modulation aware CAC
scheme and base station (BS) uplink scheduling algorithm are proposed to improve the system capacity and
QoS performance respectively. The proposed CAC scheme aims to increase the admission of higher priority
real time services by degrading the bandwidth of admitted lower priority connections on the basis of their
channel status without deteriorating their QoS performance much. The performances of proposed
algorithms are evaluated through simulation by considering the metrics like throughput, delay and number
of connections admitted.
KEYWORDS
Mobile WiMAX, Broadband Wireless Access, Connection Admission Control, channel aware and Uplink
scheduling
1.INTRODUCTION
Quality of service (QoS) guarantee is one of the important requirements in today’s
communication era which is due to the stringent quality requirements of multimedia services. In
communication systems QoS is provided through connection admission control (CAC), traffic
policing and scheduling. These traffic control mechanisms enable a balance between the
utilization of the network resources and the QoS provisioning. This tradeoff between utilization
and QoS is of fundamental importance in WiMAX networks [1], which aggregate different types
of traffic in a limited resources architecture.
The scheduling mechanism aims at guaranteeing the bandwidth required by the subscriber
stations (SS) as well as enabling the efficient wireless link usage. In a WiMAX network, the
downlink scheduling requires a single scheduler at the base station (BS), whereas the uplink
scheduling needs two components, one of them at the BS and the second one at the SS. The BS
scheduler allocates bandwidth for the SSs and the SS scheduler determines which packets will be
sent in the received transmission opportunities. In WiMAX, scheduling will be made on the basis
of scheduling classes (service types) defined. WiMAX defines five scheduling classes, unsolicited
grant service (UGS), extended real time polling service (ertPS), real time polling service (rtPS),
non real time polling service (nrtPS) and best effort (BE) services. The 802.16e standard defines

76
uplink scheduling only for UGS services and for all other services scheduling is kept as an open
issue.
In WiMAX network, before an SS initiates a new connection, it must make a request to the BS.
The task of CAC is to accept or reject the arriving requests for a connection in order to maximize
the bandwidth utilization, by accepting as many connections as possible, while maintaining the
QoS level of admitted connections. If the CAC admits too few connections, it results in wastage
of system resources. On the other hand, if CAC admits too many connections, admitted
connections contend for resources and scheduler will not be able to provide QoS. The CAC only
guarantees the bandwidth resources for an admitted connection, the actual allocation of
bandwidth resources for that admitted connection will be made by the BS uplink scheduler based
on the availability of bandwidth.
Awareness of channel status in scheduling and CAC improves the system performance, as there is
a possibility of deterioration of QoS of connections as the channel quality varies [2]. Hence in this
paper, modulation aware BS scheduling and CAC schemes are proposed to improve the network
utilization and QoS. The rest of this paper is organized as follows. Section 2 outlines the related
work in the literature. In Sections 3 and 4, proposed CAC and scheduling algorithms are
presented respectively. Section 5 discusses the simulation results followed by conclusion in
Section 6.
2. RELATED WORK
As the CAC and scheduling are the major issues in providing QoS for WiMAX network, many
researchers have proposed the algorithms for CAC and scheduling. Authors of [3-4] proposed
CAC algorithms which provide the highest priority for UGS flows and maximizes the bandwidth
utilization, in [3] bandwidth borrowing and degradation method is employed and in [4] handoff
connections are given higher priority. In order to avoid the QoS degradation authors of [5] have
proposed a statistical CAC mechanism which considers the traffic variability and overflow.
Authors of [6-8] have proposed CAC algorithms which are based on bandwidth estimation of
connections, in which [8] considers the shadow cluster concept to estimate the resources. Authors
of [9-10] proposed bandwidth degradation scheme for CAC, in which [9] employs interference
ratio (IR) as the criteria of degradation and [10] uses the concept of adaptive bandwidth
degradation. Scheduling and CAC schemes are proposed in [11-12], among which [11] is
concerned with real-time video applications and [12] considers bandwidth polling overhead
incurred by services being requested and fairness in resource allocation. Y. Ge et.al [13] proposed
a CAC scheme for adaptive multimedia services. Authors of [14-16] have proposed CAC scheme
along with packet scheduling algorithm. In [14] dual partition (DP) of the bandwidth for CAC
and priority earliest deadline for packet scheduling is proposed.
3. PROPOSED MODULATION AWARE CONNECTION ADMISSION CONTROL
MECHANISM
In this paper modulation aware connection admission control (CAC) algorithm and BS uplink
scheduling algorithms are proposed to improve the system capacity in terms of number of
connections admitted along with enhancement in QoS.
In WiMAX standard, adaptive modulation and coding (AMC) technique is used to enhance the
system performance in varying channel condition. To aid the AMC process the SS feedbacks the
channel state information (CSI) to the BS. After receiving the CSI, the BS decides the appropriate

77
modulation and coding scheme (MCS) for that SS, i.e., based on CSI it selects one of the
modulation schemes among 64-QAM, 16QAM and QPSK modulation. The SS uses this MCS to
transmit its data burst [17-18]. This work makes an attempt to utilize channel information to
design the proposed CAC and scheduling algorithms.
The basic CAC algorithm works as follow. When a new connection request arrives, the basic
algorithm considers maximum sustained rate (MSR) of that connection as an admission criterion
for UGS/ertPS and average rate as an admission criterion for rtPS/nrtPS. If the available
bandwidth ( ) is greater than the admission criterion bandwidth ( ), then that
connection will be admitted by guaranteeing , otherwise will be rejected. The is
updated after admitting a connection as per equation 1. The BE connections are admitted directly
since they do not require QoS guarantee.
The basic algorithm admits the rtPS connections by guaranteeing average rate even though the
network is less congested. This makes the utilization of bandwidth less efficient and leads to the
wastage of resources. Also if low priority connections are already admitted and if there is no
enough bandwidth to admit the high priority UGS/ertPS connection request then the basic
algorithm rejects the high priority UGS/ertPS connection.
In order to overcome the above mentioned drawbacks of basic algorithm, the proposed CAC
algorithm tries to improve the utilization of bandwidth efficiently by providing higher bandwidth
than the to the rtPS connection when the network is less congested. Also an attempt has
been made to improve the admission of higher priority connections when there is no sufficient
bandwidth to admit the request, by degrading the lower priority connections without affecting
their QoS performance based on the modulation scheme adopted by SSs.
When a new connection request belonging to UGS/ertPS seeks for admission, the proposed
algorithm also considers MSR as an admission criterion as that of basic algorithm. When the
network is less congested, the criterion for admission considered for rtPS is 80% of MSR and for
nrtPS is average rate. When the network is congested, the admission criterion for rtPS is average
rate and for nrtPS is minimum reserved rate (MRR). If the is greater than the , that
connection will be admitted by guaranteeing and is updated as per the equation 1.
If the connection request is for a real time service and is less than the , then the
proposed algorithm tries to admit that connection by degrading the lower priority connections.
The additional bandwidth required ) to admit the higher priority connection request is
calculated using equation 3.
In order to avail the to admit the real time services, initially nrtPS connections are
considered for degradation on the basis of modulation scheme used by the SSs, i.e., connections
of SSs with higher order modulation are considered first and then connections of SSs with lower
order modulation schemes.
One of the nrtPS connections of SSs which uses 64QAM modulation scheme is selected for
degradation from the list of connected SSs. Since 64QAM provides the higher data rate and is

78
capable to suffice the QoS requirements of nrtPS connection even with the MRR bandwidth
guarantee.
The maximum possible bandwidth ( ) that can be degraded from the nrtPS connection
considered is calculated using the equation 4,
Figure 1. Flow chart of proposed CAC mechanism
If is greater than the , then bandwidth of nrtPS connection is degraded as per the
equation 5 and the requested connection is admitted. The is updated as per the equation
6.

79
If is less than the , then bandwidth of nrtPS connection considered for
degradation is degraded to MRR. The and are updated as per the equation 7 and
8 respectively.
Further to avail the to admit the real time services, other nrtPS connections of SSs which
uses 64QAM modulation scheme are considered for degradation one by one and still the available
bandwidth is not sufficient to admit the real time connection request, then nrtPS connections of
SSs which uses 16QAM and QPSK modulation scheme are considered for degradation based on
the order of modulation one by one.
Further if the bandwidth available is not sufficient to admit the requested real time connection and
the request is for UGS/ertPS, to avail the the rtPS connections are considered for
degradation. The guaranteed bandwidth of rtPS connections is degraded till average rate
following the same degradation procedure as that of the nrtPS connection. If the available
bandwidth after degrading nrtPS and rtPS connections is not sufficient to admit the higher priority
connection request that connection request will be rejected. The flow chart of proposed CAC
mechanism is shown in figure 1.
4. PROPOSED MODULATION AWARE UPLINK SCHEDULING MECHANISM
In this paper along with CAC algorithm a modulation aware BS uplink scheduling algorithm is
proposed to improve the QoS performances of ertPS, rtPS and BE services. The proposed
scheduling algorithm tries to improve the performance of ertPS and rtPS services by making an
effort to allocate the requested bandwidth when requested bandwidth is more than the guaranteed
bandwidth. Also the proposed algorithm attempts to improve the performance of BE services by
allocating the requested bandwidth to the BE services in a Round-robin way by considering the
modulation schemes used by the SSs.
In the basic scheduling algorithm, the bandwidth is allocated to the admitted connections on the
basis of service type priority of the connections. That is, bandwidth is allocated to all the UGS
connections first, then to ertPS connections, later to the requested rtPS, nrtPS and BE
connections. Weighted fair queue (WFQ) scheduling algorithm is used to schedule the
connections of same service type. For UGS connections the BS allocates the MSR and for non-
UGS connections BS allocates the bandwidth based on the requested bandwidth ( ). If the
of a connection is less than the guaranteed bandwidth ( ) of that connection, then
BS allocates requested bandwidth for that connection, otherwise that connection is allocated with
the guaranteed bandwidth. The available bandwidth ( ) is updated after allocating the
bandwidth to each connection as per the equation 9.
The proposed algorithm follows same procedure as that of the basic algorithm for allocating the
bandwidth to the connections of all service types, except for ertPS and rtPS connections when
requested bandwidth is more than the guaranteed bandwidth.
When requested bandwidth of ertPS/rtPS connections is more than the guaranteed bandwidth, the
proposed algorithm degrades the nrtPS connections for which requested bandwidth is less than

80
the guaranteed bandwidth. To allocate the requested bandwidth for ertPS/rtPS connections,
additional bandwidth required ( ) is calculated using the equation 10.
Figure 2. Flow chart of proposed scheduling algorithm
In order to avail , one of the nrtPS connections of SSs which use 64QAM modulation
scheme is considered from the list of connected SSs for degradation first. If the requested
bandwidth by that nrtPS connection is less than the guaranteed bandwidth for that connection,
then the bandwidth is taken out in that frame.

81
The additional bandwidth obtained and the are updated as per the equation
11 and 12 respectively.
Further if the > 0, then to avail , other nrtPS connections of SSs which uses 64QAM
modulation scheme are considered for degradation and further if > 0, then nrtPS
connections of SSs which uses 16QAM and QPSK modulation scheme are considered based on
the order of modulation one by one. After degrading all the nrtPS connections, if still > 0,
no further degradations will be carried out. The additional bandwidth ( ) obtained by
the degradation process is allotted to the considered ertPS/rtPS connection along with the
guaranteed bandwidth. Meanwhile, after each degradation process if , then ertPS/rtPS
connection is allotted with the requested bandwidth.
Further in the proposed algorithm, the BE connections are served in a round robin manner based
on the modulation schemes used. In the first frame, one of the BE connections of SSs which uses
64QAM modulation scheme is considered based on the WFQ algorithm for bandwidth allocation.
If > of the considered connection, then requested bandwidth is allotted to that
connection, otherwise available bandwidth is allotted.
If the bandwidth is available after allocation to considered BE connection, then other BE
connections of SSs which uses 64QAM modulation scheme are considered for bandwidth
allocation one by one in the similar procedure. If the bandwidth is still available, then BE
connections of SSs which uses 16QAM and QPSK modulation schemes are considered for
bandwidth allocation in the similar procedure till available bandwidth exhaust.
In the second frame, similar procedure is followed to allocate bandwidth to the BE connections.
But the BE connections of SSs considered for bandwidth allocation is in the order 16QAM,
QPSK and 64QAM. In the third frame, similar procedure is followed to allocate the bandwidth by
considering the BE connections of SSs in the order of QPSK, 64QAM and 16QAM modulation
schemes.
Further, in the subsequent frames similar procedure is followed to allocate the bandwidth to the
BE connections by considering the modulation scheme used by SSs in Round Robin procedure as
explained. In this way all the BE connections are served by giving equal priority. The figure 2
gives the flow charts of proposed scheduling algorithm.
5. SIMULATION AND RESULTS
The performance of proposed channel aware CAC and BS scheduling algorithm is evaluated
through simulation studies using QualNet 5.0.2 simulator [19]. The performance of proposed
algorithm is compared with the basic algorithm present in simulator. A single WiMAX cell is
considered in the simulation area of 2Km X 2Km working at a frequency 2.4 GHz. The path loss
model selected is two-ray with constant shadowing model of shadowing mean 4dB. The traffic
generated for various service types along with QoS parameters (MSR and MRR) considered are
given in table 1.

82
Table 1. Traffic considered for simulation.
Service Type Type of Traffic Data Rate MSR MRR
UGS CBR 64000 64000 64000
ertPS VBR 64000 64000 16000
rtPS VBR 256000 256000 64000
nrtPS VBR 256000 256000 64000
BE FTP 1000000 0
5.1. Scenario 1
This scenario is designed to study the performance of implemented work for different channel
conditions (64QAM, 16QAM and QPSK region). In this scenario, six SSs are placed in each
region, hence total number of SSs considered is eighteen. The SSs have to transmit five traffics
each mapped to UGS, ertPS, rtPS, nrtPS and BE respectively. Hence the total number of
connections can be 90, among which each service type has 18 connections. The snapshot of
scenario is shown in figure 3. The simulation studies have been carried out by considering the
generation of these traffics in four different schemes for both basic and proposed algorithms. The
performances of basic and proposed algorithms are compared by considering the metrics like
number of connections admitted, throughput and delay.
Figure 3. Snapshot of scenario 1
5.1.1 Scheme 1
Simulation study has been carried out by considering the generation of traffic on the priority basis
from highest priority to least priority, i.e., from UGS to BE. The simulator has been programmed
to generate considered traffics at various simulation times as given in the table 2.
Table 2. Traffic generated and Number of connections admitted for scheme 1.
Simulation
time (second)
Number of connections
generated
Service
type
No. of Connections admitted
Basic Proposed
1 18 UGS 18 18
30 18 ertPS 18 18
60 18 rtPS 18 18
90 18 nrtPS 17 17
120 18 BE 0 0

83
The table 2 and Figure 4 depict the number of connections of different service types admitted by
basic and proposed algorithms for scheme 1. It is observed from the table 3 and figure 4 that the
number of connections admitted by basic and proposed algorithms is same for scheme 1, since in
this scheme, as the higher priority connections are seeking for admission prior to the low priority
connections, there is no chance for degradation of low priority admitted connections. It is also
observed from the table 3 and figure 4 that the number of BE connections served is zero since BE
connections have least priority and all the available bandwidth may be utilized to higher priority
connections.
UGS ertPS rtPS nrtPS BE
0
2
4
6
8
10
12
14
16
18
NumberofConnections
Admitted
Service Types
Basic
Proposed
Figure 4. Number of connections admitted - scheme 1
As both basic and proposed algorithms perform in similar manner for scheme 1, the throughput
performance is same for all the service types. Hence the throughput performance is not
considered here for scheme 1.
5.1.2 Scheme 2 and Scheme 3
Scheme 2: In this scheme, the simulation study has been repeated by generating all the 90 traffics
(UGS, ertPS, rtPS, nrtPS and BE connections of 18 each) simultaneously at the beginning of
simulation.
Scheme 3: In this scheme the simulation study has been repeated by considering the random
generation of traffic. Connections from each service types of one each are generated at every
interval of 5 seconds.
Table 3. Number of connections admitted - Scheme 2 and Scheme 3.
Service
Types
Number of Connections admitted
Scheme 2 Scheme 3
Basic Proposed Basic Proposed
UGS 13 15 14 16
ertPS 14 15 15 15
rtPS 13 15 14 15
nrtPS 13 14 14 15
BE 5 7 9 8

84
0
2
4
6
8
10
12
14
16
(a)
NumberofConnectionsadmitted
Service Types
Basic
Proposed
0
2
4
6
8
10
12
14
16
18
(b)
NumberofConnectionsadmitted
Service Types
Basic
Proposed
Figure 5. Number of connections admitted for (a) scheme 2 (b) scheme 3
The table 3 and Figure 5(a-b) depict the number of connections admitted by basic and proposed
algorithms for scheme 2 and scheme 3. It is observed from table 3 and figure 5(a-b) that the
number of connections admitted in proposed algorithm is more compared to the basic algorithm.
0 30 60 90 120 150
0.0
0.1
0.2
0.3
0.4
0.5
0.6 (a)
UGSThroughput(Mbps)
Simulation time (Sec)
Basic
Proposed
0 30 60 90 120 150
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
(b)
ertPSThroughput(Mbps)
Basic
Proposed
0 30 60 90 120 150
0.0
0.5
1.0
1.5
2.0
2.5
rtPSThroughput(Mbps)
Basic
Proposed
(c)
0 30 60 90 120 150
0.0
0.5
1.0
1.5
2.0
2.5
(d)
nrtPSThroughput(Mbps)
Basic
Proposed
0 30 60 90 120 150
0
2
4
6
8
10
12
14
16
18
(e)
BEThroughput(Kbps)
Basic
Proposed
Figure 6. Throughput performance of Scheme -2 for (a) UGS (b) ertPS (c) rtPS (d) nrtPS (e) BE
connections
0 30 60 90 120 150
0.0
0.1
0.2
0.3
0.4
0.5
0.6
(a)
UGSThroughput(Mbps)
Basic
Proposed
0 30 60 90 120 150
0.0
0.1
0.2
0.3
0.4
0.5
0.6
(b)
Basic
Proposed
0 30 60 90 120 150
0.0
0.5
1.0
1.5
2.0
2.5
(c)
Basic
Proposed

85
0 30 60 90 120 150
0.0
0.5
1.0
1.5
2.0
(d)
Basic
Proposed
0 30 60 90 120 150
0
50
100
150
200
250 (e)
BEThroughput(Kbps)
Basic
Proposed
Figure 7. Throughput performance of Scheme - 3 for (a) UGS (b) ertPS (c) rtPS (d) nrtPS (e) BE
connections
Figures 6(a-e) and 7(a-e) give the throughput performance of connections of different service
types with respect to simulation time for scheme 2 and scheme 3 respectively. It is observed from
the figures 6(a-d) and 7(a-d) that the throughput performance of UGS, ertPS, rtPS and nrtPS
connections for proposed algorithm is better compared to basic algorithm since the number of
connections admitted in proposed algorithm is more compared to basic algorithm. It is observed
from the figures 6(e) and 7(e) that the throughput performance of BE connections for proposed
algorithm is marginally better compared to basic algorithm.
5.1.3 Scheme 4
In this scheme the simulation study has been repeated by considering the generation of traffic in
reverse order compared to scheme 1, i.e., the connections are generated from low priority (BE) to
high priority (UGS). The traffics generated at various simulation times are given in table 4.
Table 4. Traffic generated and number of connections admitted for scheme 4.
Simulation
time (second)
Number of connections
generated
Service type No. of Connections admitted
Basic Proposed
1 18 BE 0 14
30 18 nrtPS 0 15
60 18 rtPS 18 17
90 18 ertPS 18 16
120 18 UGS 13 13
0
2
4
6
8
10
12
14
16
18
BEnrtPSrtPSertPSUGS
NumberofConnectionsAdmitted
Service Types
Basic
Proposed
Figure 8. Number of connections admitted - scheme 4
The table 4 and Figure 8 depict the number of connections of different service types admitted by
basic and proposed algorithms for scheme 4. It is observed from table 4 and figure 8 that the

86
number of connections admitted in proposed algorithm is more compared to the basic algorithm.
The table 4 and figure 8 depict that for scheme 4 the UGS and ertPS connections are not admitted
in basic algorithm, whereas the proposed algorithm successfully admits 14 UGS and 15 ertPS
connections. The proposed algorithm performs better for scheme 4 since in this scheme the
chance to degrade the low priority connections is more than for any other schemes.
Figures 9(a-e) give the throughput performance of connections of different service types with
respect to simulation time for scheme 4. It is observed from figure 9(a) that UGS throughput is
zero till 120th
second for both basic and proposed algorithm, as the UGS connections are started at
120th
second. After 120th
second the throughput of UGS connections for proposed algorithm
increases but the throughput of basic algorithm remains zero. Similarly it is observed from figure
9(b) that ertPS throughput is zero till 90th
second, and after 90th
second the throughput of ertPS
connections for proposed algorithm increases but the throughput of basic algorithm remains zero.
As in this scheme the nrtPS and rtPS connections of higher datarate are started before UGS and
ertPS, in basic algorithm almost all of the resources of admission control are utilized for
admitting the nrtPS and rtPS connections, so the higher priority UGS and ertPS connections are
not admitted. Whereas in proposed algorithm, UGS and ertPS connections are admitted since the
BS degrades the lower priority connections in order to admit higher priority connections.
It is observed from figure 9(c) and 9(d) that the throughput of rtPS and nrtPS connections for
proposed algorithm is less than the basic algorithm, because higher priority connections are get
admitted by degrading lower priority connections in proposed algorithm. It is observed from
figure 9(e) that throughput performance of BE connections for proposed algorithm is marginally
better compared to basic algorithm.
0 30 60 90 120 150
0.0
0.1
0.2
0.3
0.4
0.5
(a)
UGSThroughput(Mbps)
Basic
Proposed
0 30 60 90 120 150
0.0
0.1
0.2
0.3
0.4
0.5
0.6 (b)
Basic
Proposed
0 30 60 90 120 150
0.0
0.5
1.0
1.5
2.0
2.5
3.0
(c)
Basic
Proposed
0 30 60 90 120 150
0.0
0.5
1.0
1.5
2.0
2.5
3.0
(d)
Simulation time (Second)
Basic
Proposed
0 30 60 90 120 150
0
200
400
600
800
1000
(e)
BEThroughput(Kbps)
Simulation Time (Sec)
Basic
Proposed
Figure 9. Throughput performance of Scheme - 3 for (a) UGS (b) ertPS (c) rtPS (d) nrtPS (e) BE
connections
5.2. Scenario 2
In this scenario, design of scenario 1 is retained and the performance of basic and proposed
algorithm are studied by considering number of connections admitted, throughput and delay as

87
performance metrics for varying number of connections. Initially 15 connections are considered
for study among which 5 connections are considered in each channel conditions (i.e., 5
connections each at 64QAM, 16QAM and QPSK region). Five connections of each region are
mapped to UGS, ertPS, rtPS, nrtPS and BE service types. The simulation study has been carried
out by generating all the traffics simultaneously at the beginning of simulation. The simulation
studies are repeated by increasing the number of connections up to 90 connections insteps of 15.
Table 5 and figure 10 give the number of connections admitted for varying number of
connections. It can be observed that for less number of connections (upto 30 connections) the
number of connections admitted in both proposed and basic algorithm are same, since for less
number of connections the bandwidth requirement may be less than the available bandwidth,
hence degradation is not necessary. As the number of connections increases, the proposed
algorithm outperforms basic algorithm. From table 4 and figure 10 it is evident that the number of
connections admitted in each service type is more for proposed algorithm compared to basic
algorithm. It is also noted that the BE connections served is almost same in both proposed and
basic algorithm, as the BE connections do not go through admission control.
Table 5. Number of connections admitted for varying number of connections
(B-Basic algorithm, P – Proposed algorithm)
Total number
of Connections
generated
Total Number
of Connections
admitted
Total Number of Connections admitted in each
service type
B P B P B P B P B P B P
15 15 15 3 3 3 3 3 3 3 3 3 3
30 29 29 6 6 6 6 6 6 6 6 5 5
45 36 42 7 9 7 9 8 9 7 9 6 6
60 43 57 9 12 8 12 9 12 8 12 9 9
75 64 70 13 15 14 15 14 15 13 15 10 10
90 70 83 14 18 14 18 15 18 13 18 14 12
15 30 45 60 75 90
0
10
20
30
40
50
60
70
80
90
TotalNumberofConnectionsadmitted
Number of Connections
Basic
Proposed
Figure 10. Number of connections admitted for varying number of connections

88
15 30 45 60 75 90
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Troughput-UGS(Mbps)
Basic
Proposed
15 30 45 60 75 90
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Troughput-ertPS(Mbps)
Basic
Proposed
15 30 45 60 75 90
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Troughput-rtPS(Mbps)
Basic
Proposed
15 30 45 60 75 90
0.5
1.0
1.5
2.0
2.5
3.0
3.5
Troughput-nrtPS(Mbps)
Basic
Proposed
15 30 45 60 75 90
0.4
0.6
0.8
1.0
1.2
1.4
1.6
Troughput-BE(Mbps)
Basic
Proposed
Figure 11. Throughput performance of (a) UGS (b) ertPS (c) rtPS (d) rtPS (e) BE connections
Figures 11 (a-e) give the throughput performances of UGS, ertPS, rtPS, nrtPS and BE
connections with respect to varying number of connections respectively. It is observed that for
less number of connections the throughput performance is same for both proposed and basic
algorithm. As the number of connections increases, the throughput performances of proposed
algorithm is better compared to basic algorithm, since the number of connections admitted and
served are more in proposed algorithm. It is observed from figure 11(e) that for 90 connections
the throughput of BE for basic algorithm is more because the number of BE connections served is
more in basic algorithm.
15 30 45 60 75 90
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
AverageDelayofrealtimeservices(Second)
Basic
Proposed
Figure 12. Average delay performance of real time services for varying number of connections
Figure 12 gives the delay performance of real time services for varying number of connections. It
is observed that the delay performances are almost same for proposed and basic algorithm.
5.3. Scenario 3
This scenario is designed to study the performance of the proposed BS uplink scheduling
algorithm for different service types. It is observed from the results of scenario 2 (table 5) that all
the generated connections get admitted for the number of connections less than 30. Hence in this
scenario 20 connections are considered for study. The simulation time considered is 100s.

89
Initially five connections (one connection from each service type) are considered for simulation
and as the simulation progresses the connections of different service types are generated at the
interval of three seconds. At 3rd
second one UGS connection is generated, at 6th
second one ertPS
connection is generated and so on. In this way 20 connections (4 connections of each service
type) are generated till 45th
second. The performance metrics considered for simulation study are
throughput and delay. As the proposed scheduling algorithm works in similar way to basic
algorithm for UGS and nrtPS cases, the performances of those connections are not considered.
0 20 40 60 80 100
0
20
40
60
80
100
120
140
160
180
ertPSThroughput(Kbps)
Simulation Time (Second)
Basic
Proposed
0 20 40 60 80 100
0
100
200
300
400
500
600
700
rtPSThroughput(Kbps)
Basic
Proposed
0 20 40 60 80 100
0
100
200
300
400
500
BEThroughput(Kbps)
Basic
Proposed
Figure 13. Throughput performance of (a) ertPS (b) rtPS (c) BE connections
Figures 13 (a-c) show the throughput performance of ertPS, rtPS and BE connections respectively
with respect to simulation time. It is observed from the figures that throughput increases for both
proposed and basic algorithm initially since for every 3 seconds a new connection is get added up.
After 50th
second the throughput becomes almost constant since no new connections get added
after 50th
second. Figures also depict that throughput performance of proposed algorithm is better
compared to basic algorithm since in proposed algorithm the scheduler degrades the bandwidth
guaranteed for nrtPS connections if real time services require more bandwidth, also in proposed
algorithm the scheduler serves the BE services in a round robin manner which gives equal priority
to all connections of different MCS regions.
0 20 40 60 80 100
0
1
2
3
4
ertPSDelay(Second)
Basic
Proposed
0 20 40 60 80 100
0.0
0.2
0.4
0.6
0.8
1.0
rtPSDelay(Second)
Basic
Proposed
Figure 14. Delay performance of (a) ertPS connections (b) rtPS connections
Figures 14 (a-b) depict the delay performance of ertPS and rtPS connections respectively with
respect to simulation time. It is observed from the figures that delay performance of proposed
algorithm is better compared to basic algorithm since the proposed algorithm allocates more
bandwidth to ertPS and rtPS connections.
6. CONCLUSIONS
In this paper modulation aware CAC scheme and BS uplink scheduling algorithms are proposed.
A connection bandwidth degradation method for lower priority services is proposed to increase

90
the system capacity to admit more number of higher priority connections without affecting the
QoS performance of lower priority connections much. Also a channel aware BS uplink
scheduling algorithm is proposed to improve the QoS performances of ertPS, rtPS and BE
services. The proposed algorithms are implemented using QualNet simulation tool. It is observed
from the simulation results that the proposed algorithms outperform the basic algorithms.
REFERENCES
[1] IEEE802.16e: IEEE Standard for Local and metropolitan area networks Part 16: Air Interface for
Fixed and Mobile Broadband Wireless Access Systems, 2005.
[2] Byeong Gi Lee and Sunghyun Choi, “Broadband Wireless Access and Local Networks Mobile
WiMAX and WiFi”, Artech House, 2008.
[3] Haitang Wang, Wei Li and Dharma P. Agrawal, “Dynamic admission control and QoS for 802.16
wireless MAN”, Wireless Telecommunications Symposium, 60, 2005.
[4] Kalikivayi Suresh, Iti Saha Misra and Kalpana saha (Roy), “Bandwidth and Delay Guaranteed Call
Admission Control Scheme for QOS Provisioning in IEEE 802.16e Mobile WiMAX”, Proceedings of
IEEE GLOBECOM, 1245, 2008.
[5] Ke Yu, Xuan Wang, Songlin Sun, Lin Zhang and Xiaofei Wu, “A Statistical Connection Admission
Control Mechanism for Multiservice IEEE 802.16 Network”, In Proceedings of VTC Spring, 2009.
[6] R Murali Prasad, “An Efficient Connection Admission Control Mechanism for IEEE 802.16
Networks”, Global Journal of Computer Science and Technology, Vol. 10, Issue 8, 97, 2010.
[7] Sarat Chandra and Anirudha Sahoo, “An Efficient Call Admission Control for IEEE 802.16
Networks”, In Proceedings of the 15th IEEE Workshop on Local & Metropolitan Area Networks,
188, 2007.
[8] David A. Levine, Ian F. Akyildiz and Mahmoud Naghshineh, “A Resource Estimation and Call
Admission Algorithm for Wireless Multimedia Networks Using the Shadow Cluster Concept”,
IEEE/ACM Transactions on Networking, Vol. 5, No. 1, 1, 1997.
[9] Lang Xie and Jie Xiang, “A Novel Bandwidth Degradation Scheme for Admission Control in IEEE
802.16e Networks”, 4th International Conference on Wireless Communications, Networking and
Mobile Computing (WiCOM), 2008.
[10] Romesh Laishram and Iti Saha Misra, “A Bandwidth Efficient Adaptive Call Admission Control
Scheme for QoS Provisioning in IEEE 802.16e Mobile Networks”, International Journal of Wireless
Information Networks, Volume 18 (2), 108, 2011.
[11] Ou Yang and Jianhua Lu, “A New Scheduling and CAC Scheme for Real-Time Video Application in
Fixed Wireless Networks”, IEEE CCNC proceedings, 2006.
[12] Samuel K. Falowo and Neco Ventura, “Admission Control and Packet Scheduling for IEEE 802.16
Standards”, in proceedings of SATNAC, 2010.
[13] Y. Ge and G.S. Kuo, “An Efficient Admission Control Scheme for Adaptive Multimedia Services in
IEEE 802.16e Networks”, IEEE 64th Vehicular Technology Conference (VTC-Fall), 1, 2006.
[14] D. S. Shu’aibu and S. K. Syed Yusof, “Link aware call admission and packet scheduling for best
effort and UGS traffics in mobile WiMAX”, International Journal of the Physical Sciences Vol. 6(7),
pp. 1694-1701, 2011
[15] Wan Kim and Hwangjun Song, “QoS-aware joint working packet scheduling and call admission
control for video streaming service over WiMAX network”, Springer - Wireless Networks,1083,
2011.
[16] Shafaq B. Chaudhry and Ratan K. Guha “Adaptive Connection Admission Control and Packet
Scheduling for QoS Provisioning in Mobile WiMAX”, IEEE International Conference on Signal
Processing and Communications (ICSPC), 2007.
[17] Loutfi Nuaymi, “WiMAX: Technology for Broadband Wireless Access”, John Wiley & Sons Ltd,
2007.
[18] Kwang-Cheng Chen and J. Roberto B. de Marca, “Mobile WiMAX”, IEEE PRESS, IEEE
Communications Society and John Wiley & Sons Ltd, 2008.
[19] http://www.scalablenetworks.comqualnetdocumentation Documentation of QualNet simulation tool

Modulation aware connection admission control and uplink scheduling algorithm for wimax networks

More Related Content

What's hot (20)

Viewers also liked (17)

Similar to Modulation aware connection admission control and uplink scheduling algorithm for wimax networks (20)

Recently uploaded (20)

Modulation aware connection admission control and uplink scheduling algorithm for wimax networks