UWB versus 802.11 - a Network Perspective Gyouhwan Kim Arjunan Rajeswaran Rohit Negi gyouhwan@cmu.edu, arjunan@cmu.edu, negi@ece.cmu.edu Department of Electrical and Computer Engineering Carnegie Mellon University Abstract— Debate has been raging on the relative merits of Ultra Wide Band (UWB) and 802.11 as the technology of choice to achieve high speed wireless networking. The comparisons have focused on the single-link rate versus range issues. However, in real world applications, these radios will operate in a networking environment with significant interference effects. The interference handling capabilities of these two radios are drastically different due to their dramatically opposite power-bandwidth trade-offs. In this paper, a networking comparison, of UWB and 802.11, is performed through the application of a formal optimization theoretic framework. Simulations are conducted for network topologies typical of WPAN and WLAN scenarios. It is demon- strated that the link range over which UWB outperforms 802.11 is larger than is to be expected from a single-link comparison. Thus, a network-level comparison of different physical layers is shown to be essential in choosing the appropriate wireless technologies. A few possible future variations of 802.11 and UWB are also investigated in this flexible framework. Keywords: Wireless networks, Ultra Wide Band, 802.11, WLAN, WPAN, cross layer design, scheduling, MAC. I. I NTRODUCTION The vision of ubiquitous high-speed wireless networking has received recent impetus through two significant develop- ments. The first is the well documented WLAN (Wireless Local Area Network) revolution, that began with the IEEE 802.11b standard and has met with phenomenal success in var- ious successive versions. Today a ‘Wi-Fi’ card, a generic term for any of the 802.11a/b/g or upcoming 802.11n standards, is a standard component of most personal computers. The 802.11 based wireless standards have been designed for such WLAN applications, where an Access Point (AP) services the needs of multiple users (Wi-Fi cards) connected to it. Thus, IEEE 802.11 based Wi-Fi products are currently providing wireless internet access within the local area, such as home, office, and campus, at high data rates (e.g., 54Mbps with 802.11a [1]). The operational bands are unlicensed, one of the reasons for the mass proliferation of these devices. The second development is the FCC’s (Federal Communi- cations Commission) regulation in 2002 [2], that make the commercial use of UWB radios legal. This regulation, which allows unlicensed operation over a large bandwidth, is also a turning point in the legacy policy of separate licensed bands. UWB devices operating over such a large bandwidth can potentially achieve high data rates greater than 100 Mbps, over This work was supported in part by the National Science Foundation under awards CNS-0347455 and CNS-0520153, and by Samsung Electronics. short link distance in the range of 1 - 10m [3]. The immediate application of UWB is thought to be in the WPAN (Wireless Personal Area Network) space, such as for home networks or cable replacement around a personal computer. Thus the vast market space of consumer electronics is expected to be opened to the high-speed wireless revolution. The recently disbanded IEEE 802.15.3a task group was an attempt at standardization of UWB, while products in the form of silicon implementation of the radio are already available. Thus, each of these technologies seems to be developed to service its own distinctive networking application; UWB for WPAN and 802.11 for WLAN. While UWB as a WPAN technology is characterized by low power, high rate and short range (link distance), 802.11 as a WLAN technology trades off some rate and power for a larger range. However a distinction between applications of choice amongst these technologies is unclear, since a user could utilize either access technology for a chosen application. The argument in favor of UWB is substantiated by low power consumption and higher rate, while the counter from the 802.11 perspective is the higher range and imminent improved data rates, in future versions such as 802.11n [4]. To clarify this question, comparisons between the UWB and 802.11 PHY (physical layer) have been considered in [5],[6]. These papers compare UWB and 802.11 from the viewpoint of data rate versus range trade-off. Probable future enhancements and some practical issues have also been noted [6],[7],[8], in the context of a PHY-level comparison of the two technologies. Whereas issues, such as cost and regulatory policy, have to be considered in determining which technology should be chosen in a real world scenario, in this paper, we focus on a capacity comparison of UWB and 802.11. However, even from a capacity perspective, the purely PHY-level comparisons in previously published research, are insufficient to determine the true performance of these systems. The reason is that in WPAN and WLAN applications, there exist multiple links in sufficient proximity, so as to cause significant interference to each other. In the home 802.11 WLAN case, each AP may be associated with multiple computers and there are multiple AP’s in neighboring houses. In the UWB WPAN scenario, each individual user’s wireless links (e.g. keyboard to mobile phone) will interfere with each other. With the proliferation of these unlicensed devices, interference between devices (possibly belonging to different users) will become an increasingly important factor in determining the performance of a PHY technology.