Capacity and efficiency of IEEE 802.11n in wireless mesh operation

Since its publication in 1997, the standard 802.11 of the Institute of Electronics and Electrical Engineering (IEEE) rapidly became the de facto standard for Wireless Local Area Network (WLAN). Driven by its success, several amendments have been developed to extend and improve its functionality. Two of the most recent and substantial amendments are n (published 11/2009) and s (estimated publication 06/2011). The focus of them is considerably different: n applies latest advances in transmission technology and reduces the protocol overhead. Thereby, the net transmission rate between two terminals exceeds 100 Mb/s. In contrast, s introduces the ability to formWireless Mesh Networks (WMNs), and thus to provide cost-efficient wireless Internet access for multiple terminals in areas bigger than typical for WLAN. This thesis evaluates how the improvements introduced by amendment n perform in the complex scenarios enabled by amendment s. A key factor of these scenarios is the dynamic sharing of the available radio resources among mutually interfering terminals, performed by the Medium Access Control (MAC). To assess its efficiency three steps are carried out. First, realistic WMN scenarios are defined, based on a comprehensive system model that includes the characteristics of the wireless channel and the capabilities of amendment n. Then, the upper bound capacity of these networks is calculated, assuming an optimal sharing of radio resources. Finally, the capacity achieved by the standard 802.11 MAC in the same scenarios is evaluated by simulation. The efficiency of the radio resource management is the ratio of this capacity divided by the upper bound capacity. In simple scenarios, the 802.11 MAC reaches an efficiency of 0.8. However, in WMNs, only 0.4 is achieved. This degradation is caused by the inability of the MAC to handle complex interference present in WMNs. A reduction of interference is possible by Physical Layer (PHY)-layer improvements, namely dynamic beamforming, leading to an efficiency in the WMN between 0.5 and 0.75, depending on the scenario complexity. As an alternative to beamforming, the thesis proposes a MAC enhancement, denoted Interference Avoidance (IA). Its principle is simple: Instead of starting a transmission if the channel is idle at the transmitter, it is delayed to a time point when the channel is expected to be interference-free at the receiver. Evaluation shows that IA enables an efficiency of 0.6 to 0.8 in the WMNs under consideration. If additionally beamforming is applied, an efficiency of up to 0.9 becomes possible.