Efficient Resource Allocation in Hybrid Wireless Networks Benjamin Bengfort and Weiyi Zhang Department of Computer Science North Dakota State University Fargo, North Dakota, 58102, USA {benjamin.bengfort, weiyi.zhang}@ndsu.edu Xiaojiang Du Dept. of Computer and Information Sciences Temple University Philadelphia, Pennsylvania, 19122, USA xjdu@temple.edu Abstract— In this paper, we study an emerging type of wireless network - Hybrid Wireless Networks (HWNs). A HWN consists of an infrastructure wireless network (e.g., a cellular network) and several ad hoc nodes (such as a Mobile ad hoc network). Forming a HWN is a very cost-effective way to improve wireless coverage and the available bandwidth to users. Specifically, in this work we investigate the issue of bandwidth allocation in multi-hop HWNs. We propose three efficient bandwidth allocation schemes for HWNs: top-down, bottom-up, and auction-based allocation schemes. In order to evaluate the bandwidth allocation schemes, we develop a simulated HWN environment. Our simulation results show that the proposed schemes achieve good performance: the schemes can achieve maximum revenue/utility in many cases, while also providing fairness. We also show that each of the schemes has merit in different application scenarios. Keywords- Hybrid wireless networks; resource allocation; fairness; profit maximizing allocation; bandwidth I. INTRODUCTION The combination of IEEE 802.11 Wi-Fi and 3G/4G cellular networks could significantly increase the coverage and/or bandwidth of mobile users while keeping additional costs low. Bandwidth requirements continue to increase due to new, bandwidth-hungry “apps” and the increasing computing resource requirements of today's mobile devices. These applications, including web browsing, VoIP, and streaming media, require larger and larger amounts of bandwidth and backbone infrastructure resource support, and therefore create new challenges for today's wireless networks. Previously, it was easy for network providers to allocate data channels for a few laptops in predictable high usage areas. These users had a distinct preference to hard line Internet connections, or local wireless networks, which generally have a surplus of bandwidth; so it was easy to predict the locations where the most intense mobile wireless usage would be – areas like airports and train stations where no (free) Wireless LANs (WLANs) exist. Coffee shops and libraries already had free WLANs (usually 802.11 Wi-Fi) connectivity; therefore urban areas that contained these types of places were not high priority for dedicated data channels. Now, modern wireless networks provide “last mile” broadband connectivity, and must provide multiple entry points into the wide area network, where there are no clear geographic constraints like those that laptop users faced. Static wireless networks are necessarily constrained in their ability to support the requirements of its users because of a myriad of factors, including the mobile nature of those users, limitations in resources across segments of the network and routing inefficiencies between different devices [1]. Prioritizing an 802.11 connection over a 3G connection lowers the burden on traditional cellular networks somewhat, but not completely [2]. The emergence of a hybrid 802.11 and 802.16 network (e.g., Sprint 4G) to replace wired cable only means that bandwidth allocation issues are still a priority. Other services like Clear 802.16 4G for laptops (which also have 802.11 radios), show a trend that in the near future. The emergence of Hybrid Wireless Networks (HWNs) is necessary to accommodate the increasing wireless broadband demand, and not only provide cost reduction for better coverage but also to increase resource availability [3]. HWNs may achieve better performance by dynamic network construction rather than stand alone static infrastructure or ad hoc networking only. A HWN is a combination of an infrastructure network (such as a WiMAX, WLAN, or 3G Cellular networks) and ad hoc components (like Mobile ad hoc networks). HWNs could expand infrastructure coverage either horizontally (expanding coverage) [4] or vertically (expanding bandwidth) [5] at a low cost. HWNs provide mobile users the benefits of several types of networks- bandwidth, coverage, and mobility in a seamless fashion. Most existing papers on HWNs did not study the resource allocation issue, but rather only discussed how hybridization could improve bandwidth availability [6]. A number of papers proposed resource allocation schemes for a non-hybrid context, for example, agent-based approaches as in [7]; micro-economic or game theoretic approaches as in [8], [9], and [10]; or via some min-max algorithms for certain network metric as in [11], [12], and [13]. In this work, we investigated the resource allocation issue specifically for HWNs. In this paper, we consider a hybrid network made up of base stations (BS) that connects to relay stations (RS). A mobile device may connect to either a BS or a RS (then to a BS). The BS has a fixed amount of bandwidth that it can provide to each RS, and each RS has a set amount of demands based on the number and requirements of each of its mobile nodes. When user bandwidth demand exceeds the available bandwidth, the BS has to allocate its resources in an efficient, fair, and profit maximizing manner. Although the resource allocation problem has been well studied in traditional cellular IEEE WCNC 2011 - Network 978-1-61284-253-0/11/$26.00 ©2011 IEEE 1631