No abstract

Campus and enterprise wireless networks are increasingly characterized by ubiquitous coverage and rising traffic demands. Lfficiently assigning channels to access points (APs) in these networks can significantly affect the performance and capacity of the WLANs. Ihe state-of-the-art approaches assign channels statically, without considering prevailing traffic demands. In this paper, we show that the quality of a channel assignment can be improved significantly by incorporating observed traffic demands at APs and clients into the assignment process. We refer to this as traffic-aware channel assignment. We conduct extensive trace-driven and synthetic simulations and identify deployment scenarios where tratfic-awareness is likely to be of great help, and scenarios w~here the benefit is minimal. WXe address key practical issues in using traffic-awareness, including measuring an interference graph, handling non-binary interference, collecting tratfic demands, and predicting future demands based on historical information. We present an implementation of our assignment scheme for a 25-node WLAN testbed. Our testbed experiments show that traffic-aware assignment offers superior network pertormance under a wide range ot real network configurations. On the whole, our approach is simple yet effective. It can be incorporated into existing WLANs with little modification to existing wireless nodes and infrastructure. I. INTRODUCTIONEnterprises and university campuses are deploying WLANs at a remarkable rate and effectively managing such networks has become increasingly important. The broadcast nature of wireless communication makes the task of supporting good end-user experience very difficult. Emerging trends such as rapidly growing densities and increasing traffuc volumes only exacerbate this problem (see [13] for a detailed analysis).

Multihop wireless mesh networks are becoming a new attractive communication paradigm. Many cities and public places have deployed or are planning to deploy mesh networks to provide Internet access to residents and local businesses. Routing protocol design is critical to the performance and reliability of wireless mesh networks. Traditional routing protocols send traffic along pre-determined paths and have been shown ineffective in coping with unreliable and unpredictable wireless medium. In this paper, we develop a Simple Opportunistic Adaptive Routing protocol (SOAR) for wireless mesh networks. SOAR maximizes the progress each packet makes by using priority-based timers to ensure that the most preferred node forwards the packet with little coordination overhead. Moreover, SOAR minimizes resource consumption and duplicate transmissions by judiciously selecting forwarding nodes to prevent routes from diverging. To further protect against packet losses, SOAR uses local recovery to retransmit a packet when an ACK is not received within a specified time. SOAR uses a combination of selective ACKs, piggyback ACKs, and ACK compression to protect against ACK loss while minimizing ACK overhead. We evaluate SOAR using NS-2 simulations. Our preliminary results show that SOAR is promising to achieve high efficiency and effectively support multiple simultaneous flows.

The importance of channel assignment in wireless LANs has been long recognized and has attracted significant research attention. Despite much recent progress, several important challenges still remain: First, what is the right "performance metric" to optimize for? Most existing approaches to channel assignment try to minimize the number of mutually-interfering APs. However it is not clear if this metric accurately reflects client performance. Second, to what extent is the quality of channel assignment improved by incorporating the observed traffic demands at APs and clients? Recent work in IP traffic engineering has shed light on the tremendous effectiveness of using traffic demands in network engineering decisions. However, to date, no approach to channel assignment has taken traffic demands into account.In this paper, we conduct extensive simulations over publiclyavailable wireless traffic traces, to study the relative efficacy of different optimization metrics in tracking client performance. We show that being "traffic-aware" could substantially improve the overall quality of a channel assignment, irrespective of the metric employed. We also examine the effect of incomplete traffic information on the quality of a channel assignment. Finally, we develop and evaluate a preliminary set of practical traffic-aware assignment algorithms that predict future demands based on historical information and use the predicted demands for assigning channels.

No abstract