Priority-Based Resource Allocation Scheme

Priority-Based Resource Allocation scheme for Scalable Video Multicast in IEEE Various researchers have dedicated their effortstowards improving the resource utilization of IEEE802.16e mobile WiMAX networks. The authors of [13],[14] proposed an uplink and downlink resource allocationschemes for IEEE 802.16e networks respectively. In [13]Huang et al., argued that the scheduling and resourceallocation in uplink is complicated as compared todownlink and presented an optimization-basedformulation technique. Whereas in [14] Rashid et al.,proposed a queue and channel aware scheme as aresource allocation framework to provide quality ofservice support in the downlink of an IEEE 802.16emobile WiMAX system. J. She et al. [15] concentrated onbandwidth allocation problem in mobile WiMAX (IEEE802.16e) networks. They have presented a two-levelmodulation schemes, in first level they used QPSK andBPSK (lower data modulation code) in order to transmitthe base-layer video data whereas higher datamodulations, such as 16-QAM and 64-QAM are used insecond level to transmit the enhancement layer videocontents.However, these methods cannot produce comparableperformance for WiMAX relay networks, because 1) Theresources should be allocated for both the BS and RSs inthe network, and 2) Interference problems may occur dueto RSs during the data multicast.III. PROPOSED MODELThe proposed model is concentrated on the resourceallocation problems in two-hop WiMAX relay networksas similar to the previous approaches [16-18], [20-26]. Asshown in Fig. 2, the proposed model consists of one BS,one RS and four SSs. In Fig. 2, a solid line represents thechannel quality (CQ) of the link between the BS and RSwhile dotted line represents the channel quality betweenRS and SS. DR represents the data rate required by SSs.This model consists of four modulation and codingschemes (BPSK, QPSK, 16-QAM and64-QAM). Assumethat the lower channel quality links should use highlyreliable modulation schemes. BPSK offers the highreliability among these four schemes (suitable for linkswith bad channel quality) and 64-QAM is suitable goodchannel quality links which provides the fastesttransmission rate. Assume that every SSs has its owndevice capabilities and this SSs may request the samevideo with different quality. The H.264/SVC standarddefines many video quality levels for their individualdata-rate requirements. Here the proposed systemconsiders six video quality levels 1, 1b, 1.1, 1.2, 1.3, and2 [21], and this model uses the maximum bit rate as a datarate, 64, 128, 192, 384, 768, and 2,048 kbit/s respectively.These data rates are specified for user convenience onlybut the proposed resource allocation scheme can alsooperate under any other data rates. Depending on thedevice capability the user can select a quality level. Forexample, when a user requests a video under video qualitylevel 1, the BS should guarantee a 64 kbit/s data rate tothat user, and its DR equals 64 kbit/s. In order to providedifferent data rates, the H.264/SVC splits a video streaminto one base layer and multiple enhancement layers. Forinstance, a user with the requirement of 64 kbit/s can besatisfied by receiving the base layer, whereas a user withM.R. Thansekhar and N. Balaji (Eds.): ICIET’14CQ 2SS1DR 128Kbit/sCQ 4CQ 6BSCQ 4SS2RS1CQ 4CQ 2DR 64Kbit/sSS3DR 192Kbit/sSS4DR 128Kbit/sSS5DR 64Kbit/sFig. 2. Example of proposed modelthe requirement of 128 kbit/s can be satisfied by receivingthe base layer and one enhancement layer. Resourceallocation decisions are made for each frame in IEEE802.16j relay networks. This study concentrated on thedownlink multicast problems. The downlink process canbe divided into two zones, named as an access zone and arelay zone. The BS transmits the video data to its RSs andSSs in access zone while in relay zone, the RSs furtherrelay the video data to their served SSs. The BS shouldmake a scheduling decision at the beginning of everyframe to determine the data transmissions by using anappropriate resource allocation scheme. Shannon-Hartleytheorem is used to calculate the bandwidth consumptionfor every user. This theorem states that,B C/(log 2 L)(1)Here C signifies the channel capacity of the link (inbit/s), B represents bandwidth (in Hertz), and L representsthe number of signal elements for a modulation scheme.For example, L’s value for BPSK, QPSK, 16-QAM, and64-QAM are 2, 4, 16, and 64, respectively. If SSx,y linksto the BS with QPSK has the required data-rate DRx,y is128 kbit/s, then the bandwidth required to serve the SSx,ywould be 128/ (log 2 4 ) 64k Hertz. Here the selectionof an appropriate modulation scheme is based on thechannel quality and we divide the channel quality intofour subclasses as 1, 2, 4, and 6, the correspondingmodulation schemes are BPSK, QPSK, 16-QAM, and 64QAM. The minimum bandwidth requirement B iscalculated as,𝐵𝑟𝑒𝑞 𝐷𝑅𝑥,𝑦 /𝐶𝑄𝑥 𝐷𝑅𝑥 ,𝑦 /𝐶𝑄𝑥 ,𝑦(2)Where 𝐷𝑅𝑥 ,𝑦 /𝐶𝑄𝑥 signifies the required bandwidth forfirst hop (from BS to RS) and 𝐷𝑅𝑥,𝑦 /𝐶𝑄𝑥,𝑦 signifies therequired bandwidth for second hop transmission (from RSto SS). For instance, the minimum bandwidth requirementfor SS4 in Fig. 2 is, 𝐵𝑟𝑒𝑞 𝐷𝑅4 /𝐶𝑄3 𝐷𝑅4 /𝐶𝑄3,2 128/6 128/2 85.33 𝑘𝐻𝑧. When a BS or RSmulticasts a video data, group of SSs (within the BS/RS’scoverage range) will receive the data concurrently.935

Priority-Based Resource Allocation scheme for Scalable Video Multicast in IEEE TABLE I.SERVING PRIORITY FOR NETWORK 564128/2 64kHz64/4 16kHz192/6 192/4 80kHz128/6 128/2 85.33kHz64/6 64/4 26.67kHzWeightedvalue128/4 264/16 4192/80 2.4Priority412128/85.33 1.5564/26.67 2.43TABLE II. MULTICAST TABLES(A) FOR SS2MCSBPSK(1)QPSK(2)16QAM(4)64QAM(6)(B) FOR SS1DR forRS(kbit/s)DR forBS(kbit/s)DR forRS(kbit/s)MCSDR forBS(kbit/s)00BPSK(1)00000 to 12800 to640640 to1920000QPSK(2)16QAM(4)64QAM(6)If the BS/RS multicasts the video data using the modulation scheme corresponding to its CQ, then the videocan only be received by the SSs who having the channelquality higher than or equal to the corresponding CQ.Consider the Fig. 2, If the base station multicasts thevideo by use of 64-QAM (with CQ 6) then the RS1 onlycan receive the video. On the other hand, the BS uses 16QAM (with CQ 4) then the RS1 and MS2 can receive thesame video data simultaneously.A. Proposed Resource Allocation MethodHere we have considered an IEEE 802.16 relaynetwork with a limited amount of resource and proposesan algorithm to increase the network throughput andnumber of satisfied users. Due to bandwidth limitationsall the SSs cannot be satisfied at the same time. It isimportant to determine that, which set of SSs should availthe bandwidth first and the serving priority for the sameSS should be calculated in order to improve the networkperformance. We developed an algorithm called GreedyWeighted Algorithm (GWA) which decides the set of SSsand serving priority to efficiently use the limitedbandwidth. The weighted value W for each SS is definedas the performance gain by using the bandwidth. ProposedGWA algorithm sequentially examines the SSs forbandwidth allocation based on the weighted value W indecreasing order. If one or more SSs have the sameweighted value W, then the priority to MS will be inrandom order.equivalent weighted values for all the SSs and the servingpriority has calculated and listed in Table I. Initially allthe DR fields in the multicast table is set to be zero. SS 2has the highest priority and it is analyzed first. In order tosatisfy SS2 the DR field of 16-QAM in multicast table (forBS alone) is modified from 0 to 64 Kbit/s (Table II. A).The next prioritized user is SS3 with req-uired data rate of192 Kbit/s. In previous step BS is allocated to 64 Kbit/sbut it is not enough to satisfy SS3, because it required highdata rate. So the DR field of 16-QAM is modified from 64to 192 Kbit/s in BS table. In this case the DR field for RSshould be updated because SS3 has two-hopcommunication (BS to RS1 & RS1 to SS3) thus DR for RSis modified from 0 to 192 Kbit/s in 16-QAM modulationscheme. GWA further examines SS5 and it requires only64 Kbit/s with CQ 4.Last bandwidth assignment is well enough to satisfythe SS5 because the BS and RS both have alreadyassigned more than 64 Kbit/s in 16-QAM field. Next SS1is examined, it requires 128 Kbit/s with poor channelquality (CQ 2). Even though BS has equal bandwidth(128 Kbit/s DR1) this data rate using 64-QAM cannot bereceived by SS1 due to its poor channel quality. In thisscenario, the DR field of QPSK in multicast table shouldbe updated as 128 Kbit/s with respect to the requirementof SS1 (Table II. B). Similar procedure is repeated for theremaining SSs until all SSs requests have been processed.If the available bandwidth is insufficient to satisfy any SSmeans the proposed GWA simply skips that SS andproceeds to serve the next SS.2) Increasing Number of Satisfied Users underLimited ResourceSatisfying a user is a major concern for serviceproviders and this part concentrated on increasing thenumber of satisfied users under the limited resource.TABLE I. SERVING PRIORITY FOR SATISFIED USERSUserUserxBandwidthSS11128/2 64Weighted value(Userx/B)1/64 0.015SS2164/4 161/16 0.0621SS32192/6 192/4 802/80 0.0253SS41128/6 128/2 85.331/85.33 0.0115SS5164/6 64/4 26.671/26.67 0.0372Priority4TABLE II. ALLOCATION STATUS FOR ALL MSSPriorityUserResidual bandwidth123SS2SS580-(64/4) 6464-(64/4) 4848-(128/6 128/4) -5.33 0AllocationstatusYesYesNo(48 128/6)-(128/4) 5.33Yes(5.33 128/4)-(128/2) 0No4SS3SS15SS41) Throughput Maximization Under Limited ResourceTABLE III. FINAL MULTICASTING TABLE FOR SATISFIED USERSTo increase the network throughput here we areconsidering GWA algorithm, by using the weighted valueW as throughput gain per bandwidth DR/B, where DR isthe data rate required by user and B is the amount ofbandwidth to serve the particular SS for both the first hopand the second hop. Consider the model in Fig. 2, itsM.R. Thansekhar and N. Balaji (Eds.): ICIET’14MCSBPSK(1)QPSK(2)16-QAM(4)64-QAM(6)DR for BS (kbit/s)0128640DR for RS (kbit/s)001920936

2000Throughput dwidthconsumption ratioPriority-Based Resource Allocation scheme for Scalable Video Multicast in IEEE 21.510.5054030(b)2010015253545No of SSsFig . 3 (a) Network throughput for different number of MSs and (b)Number of satisfied users for different number of MSsIn this case we define that, a user is a satisfied user whenthe data-rate requirements are fully achieved. This sectionalso using a weighted value but it is different from theprevious section. Here we define the weighted value Was, 𝑊𝑥 𝑢𝑠𝑒𝑟𝑥 /𝐵𝑥 . Where 𝑢𝑠𝑒𝑟𝑥 represents the userscount that how many users getting satisfiedsimultaneously. For example in Fig. 2, 𝑢𝑠𝑒𝑟3 2,because while satisfying SS3, SS5 also got satisfied due toits equal channel condition. Whereas, 𝑢𝑠𝑒𝑟2 1 becausewhile serving SS2, its neighbors SS1 and RS1 are havingdifferent channel quality and thus cannot be satisfiedconcurrently. Table III lists the userx, B, W and theserving priority for all the SSs.Assume that the limited bandwidth 𝐵𝑙𝑖𝑚𝑖𝑡 80 𝑘 𝐻𝑒𝑟𝑡𝑧.The proposed algorithm prioritized all the SSs based onthe W as shown in Table IV. As per the priority order SS2is examined first, the initial bandwidth (80 k Hertz) issufficient to support the DR 64 kbit/s from the BS to SS3 (80 – (64/4) 64 kHz 0 ). Hence, the DR fields of16-QAM for BS is first modified as 64 kbit/s. Next SS5 istaken and to satisfy this user we need 64 kbit/s for RS. Weadditionally need only 64/4 16𝑘𝐻𝑧 bandwidth tosupport SS5 (64 (64/4) 48 𝑘𝐻𝑧 0). While SS3 isexamined it requires 192 kbit/s, but already the BS andRS has assigned 64 kbit/s. Thus we need extra 128kbit/sto support the SS3 with channel quality of 6 for first-hoptransmission and 128 kbit/s with CQ of 4 (i.e., (48 ((128/6) (128/4)) 5.33 0). So we cannotsatisfy the SS3 with this limited bandwidth (Blimit 80KHz). The proposed GWA algorithm consecutivelyM.R. Thansekhar and N. Balaji (Eds.): ICIET’14Satisfied user-toBandwidth consumptionratioNumber of Satisfied Users505253545No of SSsNo of SSs(b)150.050.040.030.020.0101 2 3 4 5 6 7 8 9 10Bandwidth (MHz)Fig. 4 (a) Network throughput-to-bandwidth consumption ratio andUsers-to-bandwidth consumption ratioexamines SS1 and SS4 in similar manner. While servingSS1 the previous bandwidth allocation for BS isunavailable for SS1 because, it has poor channel quality.To avoid bandwidth wastage the proposed GWA reclaimsthe extra bandwidth and select the suitable bandwidth forall the users. Thus bandwidth for SS1 is reclaimed (128kbit/s with CQ 6) and multicast table is as (𝑖. 𝑒. , (48 (128/6))). After all the process, the multicast tables aredetermined and final multicast table for BS and RS isshown in Table V.IV. RESULTS AND PERFORMANCE ANALYSISThis part presents the simulation results of the proposedGWA algorithm. We used NS2 (Network Simulator 2) forsimulation analysis and the proposed WiMAX relaynetwork consists of five RSs which all are controlled bysingle BS. Initially the locations of the RSs and MSs arerandomly selected. Based on these locations the channelquality of both access links (BS to MSs or RSs to SSs) andrelay links (BS to RSs) are set as randomized manner (1, 2,4, or 6) and corresponding modulation schemes assignedto these links. The uplink and downlink bandwidthselected for simulation is 5 Mb. The channel quality ispurely based on the distance between the links fromBS/RS to RS/SS. If the distance is less, then the channelquality will be high and vice versa. The entire simulationwork is carried out by using Destination SequencedDistance Vector Protocol (DSDV) ad-hoc routing protocol.Fig. 3(a) shows the graph for network throughput fordifferent users and Fig.3 (b) plot the graph for number ofsatisfied users among selected users.937

Priority-Based Resource Allocation scheme for Scalable Video Multicast in IEEE TABLE III.THE PARAMETERS USED FOR SIMULATIONNumber of BSs1Number of RSs5Number of MSs45Packet size1000 bytesInterval0.1 msMobility modelRandom way point modelMobility rate20mModulation schemesBPSK,QPSK,16-QAM,64-QAMData rates64 kbit/s, 128 kbit/s, 192 kbit/sThis is calculated by using PDR (Packet Deliver Ratio)and number of users joined to BS. Figs. 4a and 4b showsthe graph between throughput-to-bandwidth ratio fordifferent users and number of users-to-bandwidth ratio as afunction of bandwidth respectively. While increasing theamount of bandwidth, the number of user-to-bandwidthratio decreases and the number of satisfied users increases.The parameters are listed in Table VI.V. CONCLUSIONIn this paper we concentrated on the resource allocationproblem in WiMAX relay networks. For this purpose, wehave formulated a network with single hop and two hopcommunications by use of relay stations. We haveconstructed a table consulting mechanism for an efficientresource allocation in WiMAX relay networks. By usingweighted value and priority the proposed GWA algorithmconstructs multicast tables for BS and RSs. Throughsimulation results we can conclude that the proposedGWA algorithm achieves increased throughput andmaximum number of satisfied users while increasing theamount of users.REFERENCES[1][2][3][4][5][6][7][8][9]Bo Rong, YiQian, K. Lu, H. H Chen, and M. Guizani, ―CallAdmission Control Optimization in WiMAX Networks‖ IEEETransactions on Vehicular Technology, Vol. 57, No. 4, July 2008T. Hu, Y. P. Chen, and W. Banzhaf, ―A Genetic AlgorithmicApproach to Planning IEEE 802.16 Networks‖, provided byMemorial University of newfoundland, October 2008J. Rakesh, W. Vishal, and U. Dalal, ―A Survey of Mobile WiMAXIEEE 802.16m Standard‖, International Journal of ComputerScience and Information Security, Vol. 8, No. 1, April 2010.Chakchai So-In, Raj Jain and A. K Tamimi, ―IEEEScheduling inIEEE 802.16e Mobile WiMAX Networks: Key Issues and aSurvey‖ IEEE Journal on Selected Areas in Communications, Vol.27, No. 2, February 2009Deepti, K. Deepika, and S. P. Ahuja, ―A Survey of Rogue BaseStation Attacks in WIMAX/IEEE802.16‖, International Journal ofAdvanced Research in Computer Science and SoftwareEngineering, Volume 2, Issue 1, January 2012.J. Lee, B. Chul Yeo, Ji-Su Kim, M. Jang and J. K. Choi, ―EnergyEfficient Scalable Video Coding Based Cooperative MulticastScheme with Selective Layer Forwarding‖ IEEE CommunicationsLetters, Vol. 17, No. 6, June 2013M. W Park, and G. H. Park, ―Realistic Multi-view Scalable VideoCoding Scheme‖ IEEE Transactions on Consumer Electronics,Vol. 58, No. 2, May 2012H. Lu, W. Liao, and F. Y. Sung Lin, ―Relay Station PlacementStrategy in IEEE 802.16j WiMAX Networks‖, IEEE Trans.Communications, Vol. 59, No. 1, January 2011.C. H. Lin, J. F. Lee, and J. H. Wan, ―A Utility-Based Mechanismfor Broadcast Recipient Maximization in WiMAX MultilevelRelay Networks‖, IEEE Trans. Vehicular Technology, Vol. 61, No.5, June 2012.M.R. Thansekhar and N. Balaji (Eds.): ICIET’14[10] S. E. Hong, and M. Kim, ―Design and Analysis of a WirelessSwitched Digital Video Scheme for Mobile TV Services overWiMAX Networks‖ IEEE Transactions on Broadcasting, Vol. 59,No. 2, June 2013[11] E. Haghani, Shyam Parekh, D. Calin, E. Kim, and N. Ansari,―Quality-Driven Cross-Layer Solution for MPEG Video StreamingOver WiMAX Networks‖ IEEE Transactions on Multimedia, Vol.11, No. 6, October 2009[12] J. Huuskoa, J. Vehkaperaa, P. Amonb, C. Lamy-Bergot, G.Panzad, J. Peltolaa, M.G. Martini, ―Cross-layer architecture forscalable video transmission in wireless network‖ ImageCommunication 22 (2007) pp. 317–330[13] J. Huang, Vijay G. Subramanian, Rajeev Agrawal, and R. Berry,―Joint Scheduling and Resource Allocation in Uplink OFDMSystems for Broadband Wireless Access Networks‖ IEEE JournalOn Selected Areas In Communications, Vol. 27, No. 2, February2009[14] M. M. Rashid, and V.K.Bhargava, ―A Model-Based DownlinkResource Allocation Framework for IEEE 802.16e MobileWiMAX Systems‖ IEEE Transactions on Vehicular Technology,Vol. 59, No. 8, October 2010[15] J. She, F. Hou, P.-H. Ho, and L.-L. Xie, ―IPTV over WiMAX: KeySuccess Factors, Challenges, and Solutions [Advances in MobileMultimedia],‖ IEEE Comm. Magazine, vol. 45, no. 8, pp. 87-93,Aug. 2007.[16] M.A.Brahmia, A.Abouaissa, and P.Lorenz ―Improving IPTVForwarding Mechanism in IEEE 802.16j MMR Networks Basedon Aggregation‖ ETRI Journal, Volume 35, Number 2, April 2013,pp 234-244[17] M. W. Park, and G. H. Park, ―Realistic Multi-view Scalable VideoCoding Scheme‖, IEEE Trans. Consumer Electronics, Vol. 58, No.2, May 2012[18] Y. Guo, Y. Chen,Y. KuiWang, H. Li, M. M. Hannuksela, and M.Gabbouj, ―Error Resilient Coding and Error Concealment inScalable Video Coding‖, IEEE Trans. Circuits And Systems forVideo Technology, Vol. 19, No. 6, June 2009.[19] J. P. Sheu, C. C. Kao, S. R. Yang, and L. F. Chang,‖ A ResourceAllocation Scheme for Scalable Video Multicast in WiMAX RelayNetworks,‖ IEEE Trans. Mobile Computing, VOL. 12, NO. 1, pp90-104, January 2013[20] W. H. Kuo, W. Liao, and T. Liu, ―Adaptive Resource Allocationfor Layer-Encoded IPTV Multicasting in IEEE 802.16 WiMAXWireless Networks‖, IEEE Trans. Multimedia, Vol. 13, No. 1,February 2011[21] A.K. Al Tamimi, C. So-In, and R. Jain, ―Modeling and ResourceAllocation for Mobile Video over WiMAX Broadband WirelessNetworks‖, IEEE Journal on Selected Areas in Communications,Vol. 28, No. 3, April 2010[22] J. Cho and Z.-J. Haas, ―On the Throughput Enhancement of theDownstream Channel in Cellular Radio Networks throughMultihop Relaying,‖ IEEE J. Selected Areas in Comm., vol. 22,no. 7, pp. 1206-1219, Sept. 2004.[23] Y. Choi, J. W. Kim, J. H. Kim, J. S. Jeong, and C. G. Kang,―Scalable Transmission Control: SVC-Based Dynamic ResourceAllocation for Enhanced Multicast and Broadcast Service‖, IEEEComm. Letters, Vol. 16, No. 9, pp. 1436-1438. September 2012[24] P. Li, H. Zhang, B. Zhao, and S. Rangarajan, ―Scalable VideoMulticast With Adaptive Modulation and Coding in BroadbandWireless Data Systems‖, IEEE/ACM Trans. Networking, Vol. 20,No. 1, pp. 57-68, February 2012[25] X. Liu, X. Guan, Q. Liu, H. Wang, ―Multiple-layer scalable videocoding technology based on MB-level data partition for erCommunications 35 (2012) 1846-1854[26] J. She, F. Hou, P.-H. Ho, and L.-L. Xie, ―IPTV over WiMAX: KeySuccess Factors, Challenges, and Solutions [Advances in MobileMultimedia],‖ IEEE Comm. Magazine, vol. 45, no. 8, pp. 87-93,Aug. 2007938

towards improving the resource utilization of IEEE 802.16e mobile WiMAX networks. The authors of [13], [14] proposed an uplink and downlink resource allocation schemes for IEEE 802.16e networks respectively. In [13] Huang et al., argued that the scheduling and resource allocation in uplink is complicated as compared to