Quality of Service Performance of the Best-M OFDMA Feedback Strategy Hani H. Alyazidi Florida Institute of Technology Florida, USA halyazid2006@my.fit.edu Ivica Kostanic Florida Institute of Technology Florida, USA kostanic@fit.edu Abstract— Quality of Service (QoS) is an important perfor- mance indicator for any technique intended for its inclusion in commercial systems, as the provided service must fulfil the customers minimum demands. Several indicators for QoS are available in the literature, where the Symbol Error Rate (SER) outstands as one of the most important ones due to its large relations to correct symbols detection as well as to the achieved throughput. The Best-M feedback scheme has been proposed as an interesting technique for its use in Orthogonal Frequency Division Multiple Access (OFDMA) systems, mainly because of its data rate and feedback performance. This paper will study its QoS performance providing the SER closed form expressions to indicate the effect of the number of users, the operating SNR and/or the available number of feedback bits on the offered QoS. The performance of the proposed scheme is also presented via computer simulations where interesting conclusions are obtained. I. I NTRODUCTION Orthogonal Frequency Division Multiple Access (OFDMA) [1] is the main multiple access technology in current and future broadband wireless systems, mainly thanks to its capability to deal with the frequency selective wireless channel through its conversion into a set of flat fading subchannels (subC) that are easily equalized. Within OFDMA each user will be allocated a bundle of subCs to achieve its demanded rate and throughput requirements [1]. In order to enhance the system performance and to obtain the promised figures in the 3GPP Long Term Evolution (LTE) cellular systems [2], the transmitter must adapt its processing to the channel characteristics before transmitting, so that the Channel State Information at the Transmitter side (CSIT) [3] is required. Feedback is the tool to obtain these characteristics at the transmitter side, but its benefits are questioned by the system resources that are employed to carry the feedback information to the transmitter side. Therefore, a tradeoff has always been envisaged between the feedback advantages and its consumed resources [3]. Partial CSIT [3] seems to be a suitable option to stand on a feasible point within the tradeoff, nevertheless, the feedback over each one of the subCs is definitely non feasible and some solution is required. Based on the large amount of required feedback, some proposals targeted random users and subCs selection [4], while others developed a single bit feedback to indicate an ON-OFF state over all the subCs [5]. In [6], each user is allowed to feedback over one subC, the one showing the highest Signal- to-Noise-Ratio (SNR), thus largely decreasing the feedback load, but then the achieved data rate is far from maximum achievable one. In recent standardization activities of the 3GPP for the LTE technology [2] [7], the option of Best-M scheme [8] is recommended, where each user is asked to feedback the M subCs showing largest SNR values. The heterogeneity of applications in the current and future telecommunication system necessitates characterization of the QoS, in terms of several metrics based on the required de- sign. These metrics can be interpreted by minimum required rate and maximum tolerable SER for the reception quality, among others. This paper tackles the aforementioned QoS for the Best-M scheme, where the serviced user has strict QoS demands in terms of its reception SER. In order to characterize all the involved parameters in the Best-M scheme, this paper obtains the closed form expressions for its SER and data rate. Therefore, the Best-M scheme can be easily optimized by any commercial operator and/or man- ufacturer to account for each operating scenario conditions. Furthermore, with both the SER and data rate expressions, the system throughput [9] is also obtained. Thanks to the mathematical formulated expressions, interesting results are obtained about the impact of the number of available users, the average SNR and the allowed feedback load. The remainder of this paper is organized as follows: while section II deals with the system model, together with two subsections related to the user opportunistic scheduling and the Adaptive Coding and Modulation (ACM) used in LTE. Section III presents the Best-M technique together with its QoS perfor- mance in 3GPP LTE systems through the data rate, SER and throughput metrics, and in closed form expressions with all the involved parameters. Section IV shows the numerical results and simulations to check for the mathematically formulated results. The paper finally draws the conclusions in section V. II. SYSTEM MODEL A wireless broadband downlink communication system is considered in this paper, where the total bandwidth is equally partitioned into S subcarriers or subchannels (subC) through the OFDMA technology. A total of N single antenna users are available in the system and asking for service from the 978-1-4673-5307-6/13/$31.00 ©2013 IEEE 978-1-4673-5307-6/13/$31.00 ©2013 IEEE The 3rd International Conference on Communications and Information Technology (ICCIT-2013): Wireless Communications and Signal Processing, Beirut The 3rd International Conference on Communications and Information Technology (ICCIT-2013): Wireless Communications and Signal Processing, Beirut 396