Interference-aware and secure communications for wireless LANs

Abstract

Wireless local area network (WLAN) is a key technology driving development of the Internet of Things (IoT) by enabling smart devices such as small battery-operated sensors and smartphones to connect directly to each other, or to other networks to support ubiquitous networking. In order to provide seamless and high quality service, WLANs can adopt dynamic channel access technologies such as dynamic bandwidth or channel hopping schemes in order to avoid interference for better link quality. In addition, battery-driven circuit design for power saving WLAN has become more and more important because mobile devices and applications are required to support high throughput in long service coverage and, at the same time, exhibit long battery lifetime. However, in dense networks, the dynamic channel access leads to a higher probability of co-channel interference and adjacent channel interference. The efficiency of IEEE 802.11-based WLANs using multi-channel and wide dynamic ranges is thus severely degraded by interferences in dense networks, and the energy efficiency is severely degraded by such interference. Therefore, in this thesis, interference-aware and secure communications mechanisms for high quality and high energy efficiency are proposed. The proposed interference-aware secure communications method includes the following three key features. Firstly, WLANs can adopt channel hopping technologies in order to avoid unintentional interferences such as radars or microwaves, which function as proactive jamming signals. Even though channel hopping technologies are effective against proactive types of jamming, it has been reported that reactive jammers could attack the targets through scanning busy channels. However, the reactive jamming is only effective against channel hopping WLAN devices in non-dense networks and that it is not effective in dense networks. Therefore, in this research, a new jamming attack called “persistent jamming” is exposed, which is a modified reactive jamming that is effective in dense networks. The persistent jamming attack can track a device that switches channels using the following two features, and it can attack the specific target or a target group of devices. The first feature is that the proposed attack can use the partial association ID (PAID), which is included for power saving in the IEEE 802.11ac/af/ah frame headers, to track and jam the targets. The second feature is that it is possible to attack persistently based on device fingerprints in IEEE 802.11a/b/g/n legacy devices. The evaluation results demonstrate that the persistent jamming can improve the attack efficiency by approximately 80% in dense networks compared with the reactive jamming scheme, and it can also shut down the communications link of the target nodes using 20 dBm of jamming power and a 125 ms response time. In order to defend the persistent jamming attack, three defense mechanisms for anti-tracking and anti-jamming are proposed

a digital fingerprints predistortion, dynamic ID allocation, and dual channel friendly jamming. The experimental results demonstrate that the proposed defense mechanisms are feasible and effective to significantly decrease the device tracking success ratio of the persistent jamming attack. Secondly, an interference-aware self-optimizing (IASO) WLAN design is proposed. The proposed IASO incorporates a multi-channel multi-level carrier sense and adaptive initial gain control scheme. This scheme controls carrier sensing thresholds in each band for multi-level sensors, as well as initial gains for amplifiers. The proposed scheme reduces false carrier sensing and avoids saturation of amplifiers while simultaneously improving the dynamic range of the receiver. The prototype evaluation results demonstrate that the IASO can improve the dynamic range of the receiver by approximately 45 dB and 30 dB for a low data rate and a high data rate mode, respectively, compared with the conventional receiver designs. Furthermore, network emulation results demonstrate that the IASO-WLAN can improve the average throughput, latency, and energy efficiency by approximately 32% (24%), 41% (43%), and 13% (17%), respectively, compared with the conventional receiver designs (and channel hopping techniques) in dynamically varying interfered channel conditions. Thirdly, IEEE 802.11-based WLAN devices adopt power saving mechanisms to reduce power consumption because energy consumption is an important criterion when evaluating portable devices or sensors due to its impact on battery life. In this research, a new battery-draining attack called a Wi-Fi Vampire Attack (WiVa) is exposed, that might be used to evade the power saving features of WLAN devices. The WiVa utilizes a protocol weakness in the 802.11 power saving mechanism. The evaluation results demonstrate that the WiVa can make the target node consume approximately 14 times more than the power consumed by a normal node. In order to prevent from draining power, an energy oriented power saving or security-based power saving mechanism are proposed. And in order to mitigate unintentional interference effect, an interference-aware power saving (IAPS) mechanism is proposed. The IAPS incorporates a signal quality and interference measurement process for determining link quality based on the incoming signal from a receiver. In the IAPS, the signal field decoding process is used to determine the required link quality for the received signal, and a physical layer power saving control based on the estimated link quality and the required link quality is implemented. The proposed IAPS technique achieves 29% and 26% average energy efficiency improvement over the conventional scheme and the channel hopping scheme, respectively.