The effects of the channel quality on the total throughput are investigated. In this
simulation, the mean SNR of the subchannels received by each SS is varied from 10db to 20db, and the standard deviation is kept as 5db. As shown in Figure 11(a), since the mean SNR increases and the bursts have high chance to adopt better MCSs, the total throughputs of BCO, BSO, and Raster increase. From Figure 11 (b), as the channel quality gets better, the improvement ratios of BCO and BSO become smaller.
The reason has two points. First, Raster does not consider subchannel diversity, so the raise of throughput is simply caused by the increase of channel quality. However, BCO and BSO consider subchannel diversity and thus achieve satisfactory throughput even the mean SNR is not high. Therefore, when the mean SNR is increased, the increasing slopes on throughput in BCO and BSO are less than that in Raster. Second, when the mean SNR increases, the larger throughput achieved by Raster causes the less improvement ratios obtained by BCO and BSO.
10 12 14 16 18 20
Figure 11. Effects of Mean SNR 5.5. Variation of Channel Quality
The effects of variation of the channel quality on the total throughput are investigated. In this simulation, the mean SNR of the subchannels received by the SS is kept as 15db, and its standard deviation is varied from 0db to 10db. Figure 12(a) shows that all algorithms have the same throughput when the standard deviation is 0.
In this case, since all subchannels have the same channel quality, the total throughput of all algorithms, no matter with considering subchannel diversity or not, are the same.
Also internal fragmentation does not happen when all subchannels have the same channel quality and the load is heavy. Under the heavy load, CQQ will allocate the number of slots less than the requirement. Also since all suchannels have the same channel quality, the burst always occupies its all allocated slots.
Figure 12(a) also shows that the total throughput of Raster decreases dramatically as the standard deviation of the SNR increases, while that of BCO and BSO changes slightly. This leads that the improvement ratios of BCO and BSO raise
enormously, as shown in Figure 12(b). As the standard deviation of the SNR increases, both numbers of subchannels with good and poor SNR become larger. Also note that one burst only adopts a MCS based on the worst SNR of all assigned subchannels.
Since Raster does not construct bursts with considering the channel quality, the bursts will be constructed in the subchannels with worse SNRs as the standard deviation of the SNR increases.
Meanwhile, the throughput of BSO raises slightly and the throughput of BCO drops slightly. When the variation of SNR increases, the value of the highest SNR of the subchannels will raise and BSO can adopt a better MCS to satisfy the requested bandwidth by fewer slots. Contrarily, when the variation of SNR increases, BCO may construct a burst with a lower MCS or a higher MCS. However, constructing a burst with a lower MCS is more likely happen since one burst only adopts a MCS based on the worst SNR of all assigned subchannels. Thus the throughput of BCO is decreased slightly. Figure 12. Effects of Variation of Channel Quality 6. Conclusions
The characteristics of the IEEE 802.16 wireless communication make the burst construction be a challenge. We presented two algorithms, BCO and BSO, to maximize the throughput at constructing the burst for each connection. Our proposed algorithms not only comply with the uplink burst structure specified in the standard [14], but also consider the external fragmentation, internal fragmentation, and subchannel diversity.
From the simulation results, BCO and BSO can provide higher throughput than Raster. BSO can achieve the best throughput at the expense of a higher time complexity. Under the heavy load of 900Kbpps, BSO and BCO can achieve twice and 1.5 times, respectively, throughput obtained by Raster. Also the improvement ratios achieved by BCO and BCO raise when the requested bandwidth increases, the
number of burst decreases, and channel quality becomes better or more diverse.
References
[1] S.-Y. Wu, “Optimal infrastructure design and expansion of broadband wireless access networks,” European Journal of Operational Research, vol. 178, no. 1, April 2007, pp. 322-329.
[2] Y. Ben-Shimol, I. Kitroser, and Y. Dinitz, “Two-dimensional mapping for wireless OFDMA systems,” IEEE Transactions on Broadcasting, vol. 52, no. 3, Sep. 2006, pp. 388-396.
[3] M. Einhaus, B. Wolz, and B. Walke, “The influence of subchannel diversity on the performance of OFDMA systems based on IEEE 802.16,” in Proc. IEEE International Conference on Circuits and Systems for Communications, May 2008, pp. 20-24.
[4] S.-T. Sheu, C.-C. Yang, and H.-S. Chang, “A dynamic frequency allocation scheme for IEEE 802.16 OFDMA-based WMANs using Hungary algorithm,” in Proc. Conference on Emerging Direction in Embedded and Ubiquitous Computing, Dec. 2007, pp. 205-214.
[5] I. Toufik and R. Knopp, “Channel allocation algorithms for multi-carrier systems,” in Proc. IEEE 60th Vehicular Technology Conference, vol. 2, Sept.
2004, pp. 1129-1133.
[6] Y. Chen, S. H. Shon, S.-J. Yoo, and J. M. Kim, “Dynamic frequency selection in OFDMA,” in Proc. The 8th International Conference on Advanced Communication Technology, vol. 1, Feb. 2006, pp. 574-578.
[7] S. Najeh, H. Besbes, and A. Bouallegue, “Greedy algorithm for dynamic resource allocation in downlink of OFDMA system,” in Proc. The 2nd International Symposium on Wireless Communication Systems, Sept. 2005, pp.
475-479.
[8] D. Kivanc, G. Li, and H. Liu, “Computationally efficient bandwidth allocation and power control for OFDMA,” IEEE Transactions on Wireless Communications, vol. 2, no. 6, Nov. 2003, pp. 1150-1158.
[9] M. Ergen, S. Coleri, and P. Varaiya, “QoS aware adaptive resource allocation techniques for fair scheduling in OFDMA based broadband wireless access systems,” IEEE Transactions on Broadcasting, vol. 49, no. 4, Dec. 2003, pp.
362-370.
[10] T. Ohseki, M. Morita, and T. Inoue, “Burst construction and packet mapping scheme for OFDMA downlinks in IEEE 802.16 systems,” in Proc. IEEE Global Telecommunications Conference, Nov. 2007, pp. 4307-4311.
[11] C. So-In, R. Jain, and A.-K. Al Tamimi, “eOCSA: an algorithm for burst mapping with strict QoS requirements in IEEE 802.16e mobile WiMAX
networks,” in Proc. IFIP Wireless Days Conference, Dec. 2009, pp. 1-5.
[12] A. Erta, C. Cicconetti, and L. Lenzini, “A downlink data region allocation algorithm for IEEE 802.16e OFDMA,” in Proc. The 6th International Conference on Information, Communications & Signal Processing, Dec. 2007, pp. 1-5.
[13] T. Wang, H. Feng, and Bo Hu, “Two-dimensional resource allocation for OFDMA system,” in Proc. IEEE International Conference on Communications Workshops, May 2008, pp. 1-5.
[14] IEEE 802.16 Working Group, “DRAFT standard for local and metropolitan area networks—part 16: air interface for broadband wireless access systems,” IEEE P802.16Rev2/D1, Oct. 2007.
[15] Y.-C. Lai and Y.-H. Chen, “A channel quality and QoS aware bandwidth allocation algorithm for IEEE 802.16 base stations,” in Proc. The 22nd International Conference on Advanced Information Networking and Applications, Mar. 2008, pp. 472-479.