Embedding Trust in the Media Access Control Protocol for Wireless Networks

Abstract

IEEE 802.11 is one of the most common medium access control (MAC) protocols used in wireless networks. The carrier sense multiple access with collision avoidance (CSMA/CA) mechanisms in 802.11 have been designed under the assumption that all nodes in the network are cooperative and trustworthy. However, the potential for non-cooperative nodes exists, nodes that may purposefully misbehave in order to, for example, obtain extra bandwidth, conserve their resources, or disrupt network performance. This issue is further compounded when receivers such as Wi-Fi hotspots, normally trusted by other module nodes, also misbehave. Such issues, their detection, and mitigation have, we believe, not been sufficiently addressed in the literature. This research proposes a novel trust-incorporated MAC protocol (TMAC) which detects and mitigates complex node misbehavior for distributed network environments. TMAC introduces three main features into the original IEEE 802.11 protocol. First, each node assesses a trust level for their neighbors, establishing a verifiable backoff value generation mechanism with an incorporated trust model involving senders, receivers, and common neighbors. Second, TMAC uses a collaborative penalty scheme to penalize nodes that deviate from the IEEE 802.11 protocol. This feature removes the assumption of a trusted receiver. Third, a TMAC diagnosis mechanism is carried out for each distributed node periodically, to reassess neighbor status and to reclassify each based on their trust value. Simulation results in ns2 showed that TMAC is effective in diagnosing and starving selfish or misbehaving nodes in distributed wireless networks, improving the performance of trustworthy well-behaving nodes. The significant feature of TMAC is its ability to detect sender, receiver, and colluding node misbehavior at the MAC layer with a high level of accuracy, without the need to trust any of the communicating parties.

Keywords: CSMA/CA; IEEE 802.11; MAC; network security; trust; wireless networks.

Figure 1

IEEE 802.11 MAC protocol: DCF-based channel access mechanism.

Figure 2

The TMAC protocol monitoring and verification process with the CSMA/CA DCF.

Figure 3

TMAC with collaborative penalty scheme.

Figure 4

TMAC CTS frame with the embedded ${\mathit{TV}}_{request}$ message.

Figure 5

TMAC diagnosis mechanism.

Figure 6

Distributed dynamic network topology with random CBR traffic in ns2.

Figure 7

Diagnosis accuracy of a good receiver when it monitors a misbehaving sender.

Figure 8

Diagnosis accuracy of a common neighbor when it monitors senders and receivers.

Figure 9

Empirical CDF of trust value of good/misbehaving senders monitored by a good receiver.

Figure 10

Trust value of good/misbehaving senders monitored by a good receiver.

Figure 11

Empirical CDF of trust value of a misbehaving sender colluding with a receiver, observed by a neighbor.

Figure 12

Empirical CDF of trust value of a misbehaving receiver, observed by a neighbor.

Figure 13

Trust value of a misbehaving sender colluding with a receiver, observed by a neighbor.

Figure 14

Trust value of a misbehaving receiver, observed by a neighbor.

Figure 15

Classification of a misbehaving node based on a neighbor’s observations.

Figure 16

Classification of a good node based on a neighbor’s observations.

Figure 17

Throughput gained under TMAC prevention scheme by good/bad nodes against the misbehavior percentage, with a 95% confidence interval.

Figure 18

Throughput gained without the TMAC prevention scheme by good/bad nodes against the misbehavior percentage.

Figure 19

TMAC prevention scheme penalty allocation against the misbehavior percentage.

Figure 20

Comparison of the throughput gained under the TMAC prevention scheme and IEEE standard protocol [1] by good/bad nodes against the misbehavior percentage.

Figure 21

Comparison of the throughput gained under the TMAC prevention scheme and existing work [8] by good/bad nodes against the misbehavior percentage.