In this section, we use simulations to evaluate the performances of the 2L-DMA and the proposed algorithm.
Table 2. System simulation parameters
Table 2 shows that the parameters used in the simulation, which are from the suggestions in the WiMAX forum. Resource requests are generated randomly with the
27
constraint that sum of all MUST data within 12 30 slots. The channel bandwidth is 10 MHz, frame duration is 5 ms, and run 20000 frames.
Figure 18. Average overhead
Fig. 18 illustrates the impact of the number of MSs on the average overhead per DL subframe. The average overhead for the proposed algorithm is smaller than that for 2L-DMA. Where 2L-DMA has more overhead because of it allocate one burst for a single MS. However, the requests could be grouped into a burst with the same MCS in our proposed algorithm. In addition, one overhead occupies 5 slots. Overhead would influence on system throughput and channel utilization, when the number of overhead increases.
28
Figure 19. Average over-allocated slots.
Figure 20. Average unused slots
29
Fig. 19 illustrates the impact of the number of MSs on the average over-allocated slots and Fig. 20 illustrates the impact of the number of MSs on the average unused slots. In Fig. 19, it is observed that the proposed algorithm has fewer over-allocated slots than 2L-DMA when the number of MSs under 24. The proposed algorithm has better performance because we group requests to allocate into DL subframe. Both the proposed algorithm and 2L-DMA might have more requests cannot be allocated in Phase 1 when the number of MSs increases. In the proposed algorithm, we assume that a request consists of several packets, and we return WISH part data packet by packet to allocate MUST data might produce over-allocated slots. However, 2L-DMA returns WISH part data row by row and selects the row with maximum over-allocated slots to remove data. Therefore, 2L-DMA can reduce more over-allocated slots when the number of MSs increases.
In Fig. 20, it is observed that 2L-DMA has fewer unused slots than the proposed algorithm. The reason is that 2L-DMA didn’t consider about packet-level problem. In phase 2, both 2L-DMA and the proposed algorithm have remaining requests, which
30
Figure 21. Average MUST part ratio
Fig. 21 illustrates the impact of the number of MSs on the average MUST part ratio. The MUST part ratio can be calculated by using the equation (1). We observe that the proposed algorithm can allocate more MUST data than 2L-DMA. In a common case, there might have some large bursts with low MCS. In our proposed algorithm Step 3 of Phase 2, we could allocate some packets of MUST data to fit the strip. However, 2L-DMA has to allocate whole MUST data to DL subframe. If there is no strip can allocate whole MUST data, the request will be returned to the scheduler.
31
Figure 22. Average channel utilization
Figure 23. Average throughput
32
Fig. 22 illustrates the impact of the number of MSs on the channel utilization.
The channel utilization can be calculated using the equation (2). Fig. 23 illustrates the impact of the number of MSs on the average throughput. We have lower channel utilization than 2L-DMA because we produce more unused slots. Because we allocate bursts according to MCS but not the size of request that the proposed algorithm achieve higher throughput than 2L-DMA.
33
Chapter 6
CONCLUSION
This thesis presents an efficient packet-level mapping algorithm for the allocation of downlink resources for WiMAX systems that operate in PUSC mode.
The proposed algorithm meets the rectangle shape allocation constrain, reduces map overheads, achieves high throughput by considering mapping the requests with high MSC first, and considering the bursts allocation in packet-level. In order to reduce over-allocated slots and map overhead, the idea that we group requests to map is from OBBP [3]. In addition, we select the strip with maximum unused slots rectangle space to allocate requests that could improve channel utilization and system complexity as MaxRectangle [4].
The basic idea of the proposed algorithm is to serve MUST data as much as possible by returning less low MCS WISH data. The performance of the proposed algorithm is compared with 2L-DMA. Furthermore, the algorithm is enhanced to improve the ratio of allocated urgent data, map overhead and throughput. Simulation results show that the proposed algorithm outperforms 2L-DMA.
34
REFERENCES
[1] Erta, C. Cicconetti, and L. Lenzini, “A Downlink Data Region Allocation Algorithm for IEEE 802.16e OFDMA,” IEEE ICICS’07, Dec. 2007.
[2] C. So-In, R. Jain, and A. Al-Tamimi, “eOCSA: An Algorithm for Burst Mapping with Strict QoS Requirements in IEEE 802.16e Mobile WiMAX Networks,”
Proceedings of the 2nd International Conference on Computer and Automation Engineering (ICCAE 2010), Feb. 2010.
[3] Eshanta, M. Ismail, K. Jumari,“An efficient burst packing algorithm for OFDMA systems,”Proceedings of the 2nd International Conference on Computer Technology and Development (ICCTD), Nov. 2010.
[4] T. H. Lee, C. H. Liu, J. Yau, Y. W. Kuo, “Maximum Rectangle-Based Down-Link Burst Allocation Algorithm for WiMAX Systems,” TENCON, Nov.
2011.
[5] T. H. Lee, C. H. Liu, A. S. G. Campbell, “A Data Mapping Algorithm for Two-Level Requests inWiMAX Systems,”Vehicular Technology Conference (VTC 2012-Spring), 2012 IEEE 75th.
[6] H. C. Chen, K. P. Shih, S. S. Wang, and C. T. Chiang,“An Efficient Downlink Bandwidth Allocation Scheme for Improving Subchannel Utilization in IEEE 802.16e WiMAX Networks,” Vehicular Technology Conference (VTC 2010-Spring), May. 2010.
[7] IEEE 802.16-2009, “IEEE Standard for Local and Metropolitan Area NetworksPart 16: Air Interface for BroadbandWireless Access Systems,”May 2009.
[8] M. Ergen, “Mobile Broadband : Including WiMAX and LTE,” Springer Science+Business Media, 2009
[9] X. Zhu, J. Huo, C. Xu and W. Ding,“QoS-guaranteed scheduling and resource allocation algorithm for IEEE 802.16 OFDMA system,” in Proc. IEEE ICC’08, May 2008.
[10] Shrivastava, S. ; Vannithamby, R, “Performance analysis of persistent scheduling for VoIP in WiMAX networks,” Wireless and Microwave technology Conference,2009.
[11] Satrya, G.B. ; Agung, I.W.P. ; Cahyani, N.D.W,”Performance analysis of packet scheduling with QoS in IEEE 802.16e networks”Telecommunication Systems,
35
Services, and Applications (TSSA), 2012 7th International Conference on Telecommuncation Systems, Services, and Applications (TSSA).
[12] Xin Jin ; Jihua Zhou ; Jinlong Hu ; Jinglin Shi ; Yi Sun ;Dutkiewicz, E, “An Efficient Downlink Data Mapping Algorithm for IEEE802.16e OFDMA Systems,” Global Telecommunications Conference, 2008. IEEE GLOBECOM 2008. IEEE.
[13] Ahmed M Husein Shabanni; Prof.M.T.Beg and Ammar Abdul-Hamed Khader,”Survey of Down Link Data Allocation Algorithms in IEEE 802.16 WiMAX” International Journa.